A data object compression encoding method and system

Abstract

The application discloses a data object compression encoding method and system, and relates to the technical field of data processing. The method acquires a plurality of service data objects, generates a compression expression structure according to data organization relationships and attribute expression characteristics among the service data objects, and establishes a structure mapping relationship between the compression expression structure and the service data objects; the service data objects are compressed and encoded based on the compression expression structure, compression expression data is obtained, and structure constraint information is determined; the compression expression data is parsed based on the structure mapping relationship and the structure constraint information, attribute information and object association relationships are determined, and the service data objects are reconstructed according to the object association relationships. The application can reduce repeated expression of data, maintain the association relationship between the compression data and the original service objects, and improve data processing efficiency and recovery accuracy.

Description

Technical Field

[0001] This application relates to the field of big data processing technology, and in particular to a method and system for compressing and reconstructing business data objects. Background Technology

[0002] With the development of information-based services, business processes typically require the generation, transmission, storage, and retrieval of large amounts of business data. To facilitate business processing, relevant business data is usually organized into business data objects according to certain data organization methods. These business data objects describe the corresponding business content through multiple attribute information. For example, in application scenarios such as intelligent services, business interface interaction, data management, and information processing, business data objects can be used to carry different types of business information and serve as the basis for data interaction between different business processing stages.

[0003] In existing technologies, to reduce the resource consumption of business data objects during transmission and storage, data compression encoding methods are typically used to process business data. These methods include reducing repetitive character expressions, adjusting data arrangement, or using preset encoding rules to reduce data length. However, business data objects usually contain not only specific data content but also the organizational relationships between attributes, the correspondence between attributes and objects, and the associations between different objects. When compressing business data objects, if only the data content is compressed without fully considering the aforementioned structural relationships, the compressed data may lack necessary structural information. This can make it difficult to accurately determine the business objects corresponding to attribute information during subsequent recovery, or to maintain consistency between the recovered business data objects and the uncompressed ones.

[0004] Therefore, a technical solution is needed that can compress business data objects based on their data organization characteristics. This solution can reduce the redundancy in the expression of business data objects while maintaining the structural association information during the compression process, so that the compressed business data can still be accurately parsed and recovered based on the corresponding relationships. Summary of the Invention

[0005] This application aims to solve the problem that in the existing business data object compression process, the loss of object structure information due to the reduction of data expression content leads to an incorrect correspondence between the restored business data object and the object before compression.

[0006] To achieve the above objectives, this application provides a data object compression encoding method, comprising: Multiple business data objects are acquired, and a corresponding compressed expression structure is generated based on the data organization relationship and attribute expression characteristics among the multiple business data objects. A structural mapping relationship is established between the compressed expression structure and the multiple business data objects. The multiple business data objects are compressed and encoded based on the compressed expression structure to obtain compressed expression data. Structural constraint information is determined according to the arrangement and association relationships between data elements in the compressed expression structure to maintain the association relationship between the compressed expression data and the multiple business data objects.

[0007] Optionally, the method further includes: parsing the compressed expression data based on the structural mapping relationship and the structural constraint information, determining the attribute information and object association relationship corresponding to the compressed expression element, and combining and reconstructing multiple attribute information according to the object association relationship to obtain the restored business data object.

[0008] Optionally, generating the corresponding compressed expression structure includes: The attribute information in the multiple business data objects is traversed to obtain the attribute name, attribute type and attribute value expression relationship of the corresponding attributes in different business data objects, and the common attribute information used for shared expression is determined based on the repeated occurrence relationship of attribute information among multiple business data objects. Based on the attribute position, attribute reference relationship and attribute arrangement order of the public attribute information in different business data objects, establish the attribute mapping relationship between the public attribute information and the compressed expression identifier; Based on the attribute mapping relationship, common attribute information in multiple business data objects is converted into corresponding compressed expression identifiers, and the correspondence between the compressed expression identifiers and the original attribute positions is stored in the compressed expression structure, so that when the compressed expression data is restored, the original attribute information corresponding to the compressed expression identifier can be determined according to the attribute mapping relationship.

[0009] Optionally, generating the corresponding compressed expression structure includes: The attribute information in the multiple business data objects is classified and processed to determine the set of attribute values ​​with the same attribute type in the multiple business data objects. Based on the arrangement relationship of business data objects corresponding to different attribute values ​​in the attribute value set, the original attribute values ​​organized according to business data objects are converted into a sequence of attribute values ​​organized according to attribute type; An index relationship is established between the attribute values ​​in the attribute value sequence and the corresponding business data objects, so that after multiple business data objects undergo attribute dimension transformation, the original correspondence between the business data objects can still be restored based on the index relationship.

[0010] Optionally, obtaining compressed expression data includes: Obtain the identification attribute values ​​from the multiple business data objects, analyze the numerical continuity relationship between the multiple identification attribute values, and determine the set of identification attribute values ​​that meet the condition of continuous change. Range expression information is generated based on the start and end identifier attribute values ​​in the set of identifier attribute values, and an association is established between the range expression information and the corresponding business data object; The expression method of replacing multiple independent identifier attribute values ​​based on the range expression information is used to reduce the duplication of identifier attribute values ​​in the compressed expression data, and the corresponding identifier attribute values ​​are restored according to the range expression information during the recovery process.

[0011] Optionally, obtaining compressed expression data includes: Obtain the status attribute information from the plurality of business data objects, and determine the arrangement position of the status attribute information according to the arrangement order among the plurality of business data objects; Multiple state attribute information is converted into consecutive encoded bits according to the arrangement position, and a state encoding sequence is formed according to the positional correspondence between the consecutive encoded bits and the business data object; When restoring the business data object, the state attribute information corresponding to each encoding bit in the state encoding sequence is determined according to the position correspondence, so as to restore the state information corresponding to different business data objects.

[0012] Optionally, obtaining compressed expression data includes: Obtain text attribute information from the multiple business data objects, analyze the repeated text content and text boundary relationships between different text attribute information, and determine the text fragments that can be shared and expressed. A text segment index is generated based on the reference relationships between the text segments, and multiple text attribute information is serialized and expressed according to the text segment index; The serialized expression result is associated with the corresponding business data object to reduce the repetitive expression of text attribute information while recovering the complete text attribute information in the corresponding business data object based on the text fragment index.

[0013] Optionally, the structural constraint information includes element boundary constraint information, and the generation of structural constraint information includes: Based on the arrangement relationship between multiple compressed expression elements in the compressed expression structure, the boundary positions between adjacent compressed expression elements are determined; Element boundary constraint information is generated based on the boundary position, and the element boundary constraint information is associated with the compressed representation data; When parsing the compressed expression data, independent compressed expression elements are determined based on the element boundary constraint information to avoid boundary mismatches between the corresponding attribute information of different business data objects during the recovery process.

[0014] Optionally, the structural constraint information includes attribute location constraint information, and the generation of structural constraint information includes: Based on the arrangement relationship between multiple attribute information in the compressed expression structure, the positional relationship of different attribute information in the corresponding attribute sequence is determined; Attribute position constraint information is generated based on the positional relationship, and the attribute position constraint information is associated with the compressed expression data; During the recovery process, the correspondence between different attribute information is determined based on the attribute location constraint information, so that the multiple attribute information encoded separately can be mapped to the correct business data object.

[0015] Optionally, the structural constraint information includes object association constraint information, and the generation of structural constraint information includes: Based on the combination relationship between different attribute information in the multiple business data objects, determine the attribution relationship between attribute information and business data objects; Based on the attribution relationship, object association constraint information is generated, and the object association constraint information is associated with the compressed representation data; During object reconstruction, attribute information stored in different compressed expression structures is recombined according to the object association constraint information to restore multiple business data objects corresponding to those before compression.

[0016] Optionally, the compressed expression structure includes encoding rule information, and generating the corresponding compressed expression structure includes: Based on the attribute types, attribute arrangement relationships, and attribute value expression characteristics in the multiple business data objects, determine the encoding expression method corresponding to different attribute information; Generate a corresponding compressed expression structure based on the encoding expression method, and associate the encoding expression method with the structure mapping relationship so that different types of attribute information are compressed and expressed in the corresponding way; During the recovery process, the parsing method for the corresponding attribute information is determined based on the encoding expression method.

[0017] Optionally, parsing the compressed representation data based on the structural mapping relationship and the structural constraint information includes: The attribute information corresponding to the compressed expression elements in the compressed expression data is determined based on the structural mapping relationship. The positional relationships between different attribute information and their respective business data objects are determined based on the structural constraint information. Based on the location relationship and the business data object to which it belongs, multiple attribute information is combined to recover the recovery business data object corresponding to the multiple business data objects.

[0018] Optionally, parsing the compressed representation data based on the structural mapping relationship and the structural constraint information includes: The correspondence between the compressed expression elements in the compressed expression data and the multiple business data objects is determined based on the structural mapping relationship; Determine the attribute recovery order, attribute association, and object composition relationship in the correspondence based on the structural constraint information; According to the attribute recovery order, attribute association, and object combination relationship, multiple compressed expression elements in the compressed expression data are parsed and combined to maintain the structural consistency between the compressed expression data and the recovered business data object.

[0019] To achieve the above objectives, this application also provides a data object compression encoding system, including a business data generation server, a compression expression processing server, a compressed data management server, and a business data recovery server; The business data generation server is used to acquire multiple business data objects and send the multiple business data objects to the compressed expression processing server; The compressed expression processing server is used to generate a corresponding compressed expression structure based on the data organization relationship and attribute expression characteristics among the multiple business data objects, establish a structural mapping relationship between the compressed expression structure and the multiple business data objects, and compress and encode the multiple business data objects based on the compressed expression structure to obtain compressed expression data, while generating structural constraint information to maintain the association relationship between the compressed expression data and the multiple business data objects. The compressed data management server is used to store the compressed expression data and the corresponding structural mapping relationship and structural constraint information; The business data recovery server is used to acquire the compressed expression data, the structural mapping relationship, and the structural constraint information, parse the compressed expression data, determine the attribute information and object association relationship corresponding to the compressed elements, and combine and reconstruct multiple attribute information according to the object association relationship to obtain the recovered business data object.

[0020] Compared with the prior art, this application has at least the following beneficial effects: First, this application does not simply encode the data content in the business data object, but generates a compressed expression structure based on the data organization relationship and attribute expression characteristics between multiple business data objects, so that repetitive attribute information, related attribute information and object structure information can be organized and expressed according to the corresponding relationship, thereby reducing the repetitive expression content in the business data object.

[0021] Secondly, while generating compressed expression data, this application determines structural constraint information based on the arrangement and association relationships between data elements in the compressed expression structure, so that the compressed data not only retains the data content, but also retains the interpretation relationship between compressed elements and business data objects, avoiding the loss of object ownership relationships due to the compression process.

[0022] Furthermore, this application parses compressed representation data based on structural mapping relationships and structural constraint information. After determining the attribute information corresponding to the compressed elements, it recombines multiple attribute information according to object association relationships, so that the restored business data objects maintain the same data organization relationship as before compression. Attached Figure Description

[0023] Figure 1 A flowchart of a data object compression encoding method provided in an embodiment of this application; Figure 2 This is a schematic diagram illustrating an application scenario of the data object compression encoding method provided in the embodiments of this application; Figure 3 This is a schematic diagram of the structure of the data object compression encoding system provided in the embodiments of this application. Detailed Implementation

[0024] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Other implementation methods obtained by those skilled in the art based on the embodiments of this application without creative effort are all within the protection scope of this application.

[0025] This embodiment provides a data object compression encoding method, which is applied to business processing scenarios that require compression, storage, transmission, and retrieval of multiple business data objects. See also... Figure 1 The data object compression encoding method includes the following steps: S1. Obtain multiple business data objects, and generate corresponding compressed expression structures based on the data organization relationships and attribute expression characteristics among the multiple business data objects, and establish a structural mapping relationship between the compressed expression structures and the multiple business data objects.

[0026] In this embodiment, a business data object represents structured business information generated during business processing. It typically consists of multiple interrelated attribute information items, each describing different dimensions of the business object. For example, in applications such as intelligent teaching, business management, and content generation, a business data object may include a type attribute describing the object's category, a status attribute describing the object's state, a number attribute representing the object's identifier, a text attribute recording the object's content, and a reference attribute expressing the relationships between objects. These attribute information items are not simply data sets, but rather organized according to certain business rules to form a complete business object that can be recognized and processed by the business system.

[0027] In actual business processing, business data objects are typically generated or transmitted in batches. These multiple business data objects are not entirely independent; rather, they may be related based on the same business template, processing flow, or data source. For example, during a course content generation process, multiple course objects, chapter objects, text objects, and character objects may be generated simultaneously. These different objects may share the same course type, content category, status expression method, or attribute structure. Similarly, multiple record objects generated in the same business task processing may all contain attributes such as object number, creation time, processing status, and business category, differing only in specific business content, processing result, or object identifier. Therefore, multiple business data objects typically exhibit data characteristics such as duplicate attribute information, related attribute information, and hierarchical relationships.

[0028] Among these, duplicate attribute information refers to attribute content in multiple business data objects that has the same business meaning and the same expression method. For example, when multiple business data objects belong to the same business type, their type attributes may be completely identical; when multiple objects are generated from the same business template, their partial descriptive information, status definitions, or configuration parameters may also be consistent. Related attribute information refers to the business correspondence between different attributes, such as a one-to-one correspondence between object number and object status, an association between text content and text annotation, and an inclusion relationship between parent and child objects. Object hierarchy relationships are used to describe the organizational structure between multiple business data objects. For example, a business object can contain multiple sub-business objects, and sub-business objects further contain corresponding attribute information; objects at different levels together form a complete business data structure.

[0029] Because multiple business data objects share the aforementioned common characteristics, if each business data object is saved or transmitted separately in the traditional way, the same attribute information needs to be repeatedly expressed in multiple objects, and the relationships and hierarchical relationships also need to be repeatedly described through the complete object structure. This not only increases the length of the business data expression but also reduces the resource utilization efficiency in the data processing process. Especially in scenarios that require processing a large number of structurally similar business objects simultaneously, such as when a large language model generates multiple structured business objects, a business platform processes multiple records in batches, or multiple clients synchronize the same type of business information, a large number of duplicate field names, duplicate attribute values, and duplicate object structures will occupy a lot of data space. Therefore, it is necessary to analyze the data organization relationships and attribute expression characteristics between multiple business data objects, and, while preserving the relationships between business objects, to compress and express multiple business data objects in a way that is more suitable for the overall data characteristics.

[0030] Existing data compression methods typically encode the data content itself, reducing data size by minimizing repetitive characters or adjusting data representation. However, for business data objects, data content and data organization relationships usually coexist. Attribute information in business data objects not only represents specific business content but also determines the object to which that attribute belongs, the attribute's position, and the relationships between different attributes. If only the data content of a business data object is compressed without preserving these structural relationships, problems can easily arise during recovery, such as attribute information not mapping to the correct business object, attribute mismatches between multiple objects, or changes to the restored object structure.

[0031] Therefore, after acquiring multiple business data objects, this implementation first analyzes the data organization relationships and attribute expression characteristics among these objects to determine the compressible representations and structural associations that need to be maintained. Specifically, for the acquired business data objects, they are first divided according to their business object types, and the attribute fields of each object are traversed to obtain the field name, field type, field value, and position of each field within the object structure. For example, for multiple course objects, the course type field, subject field, status field, text field, and object collection field can be obtained for each course object; for multiple character objects, the character content field, number field, status field, and annotation field can be obtained. By analyzing the attribute composition of multiple business data objects, a correspondence between the business data objects and their attribute information can be established, thus providing a data foundation for subsequently determining the repetition relationships, association relationships, and expression methods among different attribute information.

[0032] After parsing the attribute information, the attribute expression characteristics of multiple business data objects are further analyzed to determine the distribution of different attribute information across these objects. Specifically, a horizontal comparison can be made of attribute information with the same field name, statistically analyzing the frequency of occurrence, consistency of values, and range of variation of the same attribute field in different business data objects. Simultaneously, the order of attribute fields in different business data objects is recorded to determine the positional relationship of each attribute within the business object structure. Furthermore, by considering the common occurrence of different attribute fields, it can be analyzed whether multiple attribute information is used to describe the same business object. Through this process, the data organization patterns between business data objects can be determined from multiple dimensions, rather than relying solely on a single field value to determine whether compression is necessary.

[0033] The repetition of attribute information is used to determine whether there are shared, reproducible common attribute information across multiple business data objects. Specifically, attribute information with identical field names and values ​​across multiple business data objects can be identified as candidate common attribute information. For example, when multiple chapter objects belong to the same course, the course type, subject, and content type fields may all be identical, indicating a high degree of repetition and making them candidate common representations. Similarly, for status fields, number fields, or fixed configuration fields that exist across multiple character objects, their repetition patterns can be used to determine their suitability for centralized representation. Furthermore, the sharing method for common attribute information is determined based on the coverage and repetition ratio of attribute information across different business data objects, eliminating the need for separate storage of recurring information across multiple business data objects. Instead, it is represented through a unified representation object or a shared representation identifier.

[0034] The arrangement of attribute information is used to determine the positional relationship of attribute information within the corresponding business data object. Specifically, when parsing multiple business data objects, not only are the field name and value of each attribute field recorded, but also the position of that attribute field in the object structure and its adjacency relationship with other attribute fields are recorded. For example, for a character object, the character content field `char` can be recorded as being in the first position in the object's attribute set, the ID field `id` as being in the second position, and the status field `checked` as being in the third position. For compressed data after attribute dimension transformation, the original attribute positions corresponding to different elements in the compressed attribute sequence can be determined based on the above positional relationships. By analyzing the attribute arrangement relationships, similar attributes in multiple business data objects can be grouped together during compression, and the object to which each attribute belongs can be determined based on the attribute positional relationships during the recovery process, avoiding the offset of the recovered object structure due to attribute rearrangement.

[0035] The association of attribute information is used to determine whether different attribute information jointly describes the same business object. Specifically, the degree of association between attribute information can be determined based on the co-occurrence relationship of multiple attribute fields in the same business data object, the reference relationship between fields, and the inclusion relationship between business objects. For example, the character content field `char`, the number field `id`, and the status field `checked` usually describe the same character object, thus having an object-level association relationship; the chapter list field and the chapter attribute field in a course object have a hierarchical association relationship; and the text content field and the text annotation field in a text object have a content association relationship. Based on the above association analysis results, the attribution relationship between attribute information and business objects can be established, so that even if different attribute information is encoded separately or expressed collectively during subsequent compression, the business object corresponding to each attribute information can still be determined based on the association relationship.

[0036] Furthermore, based on the analysis results of the repetition, arrangement, and association situations described above, this embodiment determines the compressible representation portion and the structural association information that needs to be maintained in multiple business data objects. The compressible representation portion mainly includes recurring common attribute information, identifier attribute information with regular changing relationships, state attribute information that can be converted into continuous encoding form, and text attribute information that can be represented by indexes. The structural association information that needs to be maintained includes the hierarchical relationship between attributes and objects, the hierarchical relationship between parent and child objects, and the positional correspondence between different compressed elements. By simultaneously analyzing the compressible representation portion and the structural association information, the data consistency between business data objects before and after compression can be maintained while reducing the length of the business data object representation, providing a foundation for subsequent generation of compressed representation structures, establishment of structural mapping relationships, and generation of structural constraint information.

[0037] Furthermore, based on the analysis results of the data organization relationships and attribute expression characteristics among multiple business data objects, this embodiment generates a corresponding compressed expression structure. The compressed expression structure is not merely used to store the compressed data content, but rather to describe the data organization methods, attribute expression methods, and element association methods used in the process of converting multiple business data objects from their original object expression forms to compressed expression forms. Specifically, when generating the compressed expression structure, firstly, based on the compressible expression parts determined in multiple business data objects, different types of attribute information are reorganized. For example, for common attribute information that appears repeatedly in multiple business data objects, corresponding shared expression elements can be generated, allowing multiple business data objects to jointly reference the same common expression content; for identifier attribute information with continuous change patterns, range expression elements can be generated; for state attribute information with fixed values ​​or finite state changes, encoded expression elements can be generated; and for text attribute information with repetitive text content, index expression elements can be generated. Through the above processing, different types of business attribute information can be converted into expression forms suitable for compression processing according to their own data characteristics, rather than being simply compressed using a uniform method.

[0038] Furthermore, the compressed expression structure also records the positional, arrangement, and association relationships between different compressed expression elements. For example, when the character content, numbering information, and status information in multiple business data objects are converted into different compressed expression elements, the compressed expression structure needs to record the correspondence between the character content elements, numbering elements, and status elements, so that the server can determine the numbering information and status information corresponding to a certain character content. When there is a parent-child hierarchical relationship between multiple business data objects, the compressed expression structure also needs to record the hierarchical connection relationship between the compressed elements corresponding to the parent object and the compressed elements corresponding to the child object, so that the compressed data can still maintain the original organizational structure between business objects. By recording the above structural information, the compressed expression structure not only indicates "what data is saved after compression" but also "how the compressed data corresponds to the original business objects," thus providing a structural basis for subsequent data recovery.

[0039] For example, in a smart teaching scenario, multiple character objects can each contain character content attributes, number attributes, and status attributes. For the character content of multiple character objects, compressed character content elements can be generated; for consecutive numbering information, range expression elements can be generated; and for multiple status attributes, consecutive encoding elements can be generated. In this case, the compressed expression structure not only stores compressed expression elements such as the `c`, `id`, and `cv` fields, but also further records the character attribute in the original character object corresponding to the Nth position of the `c` field; the number attribute in the same character object corresponding to the Nth position of the `id` field; and the status attribute in the same character object corresponding to the Nth encoding position of the `cv` field. Through this structural description, data originally organized according to the object dimension is transformed into data organized according to the attribute dimension, while still maintaining the one-to-one correspondence between attribute information and business objects.

[0040] Simultaneously, a structural mapping relationship is established between the compressed expression structure and multiple business data objects. This structural mapping relationship describes the bidirectional correspondence between the original business data objects and the compressed expression structure; that is, it records how attribute information in the business data objects is converted into corresponding compressed expression elements, and how compressed expression elements should be remapped to their corresponding business data objects during the recovery process. Specifically, when establishing the structural mapping relationship, the source attribute, the business object to which it belongs, and the corresponding positional relationship can be recorded for each compressed expression element. For example, for a 'c' field generated from character attributes in multiple character objects, the record can be: 'c' field → character object → 'char' attribute; for an 'id' field generated from number attributes in multiple character objects, the record can be: 'id' field → character object → 'id' attribute; for a 'cv' field generated from a status attribute, the record can be: 'cv' field → character object → 'checked' attribute. Through the above mapping relationship, the original business attribute corresponding to each compressed expression element can be determined.

[0041] Furthermore, for multiple business data objects with hierarchical relationships, the structural mapping relationship is also used to record the mapping paths between objects at different levels. For example, a hierarchical mapping path can be formed between course objects, chapter objects, text objects, and character objects: course object → chapter object → text object → character object. During compression, when attribute information in objects at different levels is encoded separately or shared, the structural mapping relationship can record the object level and attribute attribution of the compressed element; during restoration, based on this mapping relationship, it can be determined whether a compressed element should be restored to a course object, chapter object, text object, or character object, thereby avoiding attribute attribution errors during restoration.

[0042] Furthermore, to improve subsequent parsing efficiency, this implementation can also establish an expression path mapping based on the arrangement order, reference relationships, and combination relationships between elements in the compressed expression structure. For example, for business data objects after attribute dimension transformation, the parsing path of compressed element position → attribute type → original field → business object can be recorded, so that the recovery process does not need to re-analyze the business data structure, but can directly determine the recovery method corresponding to each compressed element based on the structure mapping relationship.

[0043] Through the above processing, the established structural mapping relationship associates the compressed representation structure with multiple business data objects. This ensures that while the compressed business data alters the original object's representation, it retains the attribute relationships, object relationships, and hierarchical relationships within the original business objects. When subsequent recovery of the compressed representation data is required, the business objects and attribute information corresponding to the compressed representation elements can be determined based on the structural mapping relationship. Combined with structural constraint information, the business data objects can be recombined, thereby achieving a reversible conversion between the compressed representation data and the original business data objects.

[0044] S2. Based on the compressed expression structure, the multiple business data objects are compressed and encoded to obtain compressed expression data. The structural constraint information is determined according to the arrangement and association relationships between data elements in the compressed expression structure to maintain the association relationship between the compressed expression data and the multiple business data objects.

[0045] In this embodiment, after generating the compressed expression structure and the structure mapping relationship through S1, multiple business data objects are further compressed and encoded according to the compressed expression structure to form corresponding compressed expression data.

[0046] It should be noted that although the attribute information in multiple business data objects collectively constitutes the complete expression of a business object, the appropriate compression expression methods differ because the business meanings, change patterns, and distribution methods of different attribute information vary among objects. Specifically, some attribute information has a high degree of repetition across multiple business data objects. For example, multiple course objects may have the same course type, subject information, and content type; multiple character objects may have the same status description or fixed configuration attributes. If such attribute information is still expressed by saving it separately for each business object, the same data content will be repeatedly occupied. Therefore, it is more suitable to compress it using common expression objects, shared expression identifiers, or reference relationships. While some attribute information differs between different business data objects, its change process follows a certain pattern. For example, attributes such as object number, sequence number, or time identifier may increase continuously. For such attribute information, saving each attribute value individually would fail to utilize its change pattern. Therefore, it is more suitable to generate range expression information, sequence expression information, or other regularized expression forms based on the change relationships between attribute values. Some attribute information may only exist in specific business objects or be generated only in specific business states, such as certain interaction result fields, exception description fields, or extended attribute fields. For such attribute information, it is necessary to maintain its belonging relationship with the corresponding business object to avoid loss or incorrect association of attribute information due to centralized processing during the compression process.

[0047] Furthermore, different attribute information may have different data organization methods. For example, for text-based attribute information, its main characteristic is that there may be repeated text fragments or similar descriptive content among multiple business objects. Therefore, common text fragments can be extracted based on the repetition relationship between text content and expressed through indexing. For status-based attribute information, it usually has a limited range of values, such as status fields like "completed," "selected," and "correct." It can be converted into a continuous encoding form based on the arrangement relationship between multiple business objects. For identifier-based attribute information, it is usually used to distinguish different business objects. It can be expressed using range or list representation based on the continuous change relationship between identifier values. For hierarchical association-based attribute information, it is important to maintain the mapping relationship between parent and child objects so that different levels of business objects can still restore their original organizational structure after compression. Therefore, if a single compression method is used to uniformly process all attribute information, it can only reduce some repetitive data expression, but cannot fully utilize the data characteristics of different attribute information itself, and may also destroy the original association relationship between business objects.

[0048] Therefore, this implementation determines the expression methods corresponding to different attribute information based on the compressed expression structure, and reorganizes and encodes the attribute information according to the data organization relationship between multiple business data objects to obtain compressed expression data. Specifically, according to the aforementioned determined common attribute information, changing attribute information, and associated attribute information, corresponding compressed expression rules are matched respectively. For common attribute information that can be shared, the same attribute content in multiple business data objects is converted into a unified expression element, and the association relationship between the expression element and multiple business data objects is established; for attribute information with changing patterns, the corresponding encoding method is determined according to the attribute value arrangement relationship, the range of change, and the object order. For example, continuously changing number attributes are converted into range expressions, and multiple state attributes are converted into continuous encoding bits; for attribute information that needs to maintain the object belonging relationship, its belonging business object and corresponding position are recorded according to the structural mapping relationship, so that the attribute information can still be restored to the original object structure after reorganization.

[0049] In generating compressed representation data, this implementation method does not simply reduce the number of fields. Instead, it rearranges, aggregates, and encodes the original attribute information based on the inherent relationships between business data objects. For example, character content attributes in multiple character objects can be formed into continuous representation fields according to the character order; number attributes in multiple character objects can be converted into range representation fields based on the continuous relationship; and multiple status attributes can be converted into encoded fields based on the object order. Simultaneously, a structural mapping relationship is used to record the correspondence between different compressed representation elements and the original business attributes. Therefore, while reducing the length of the data representation, the compressed representation data still retains the attribute relationships, object relationships, and hierarchical relationships in the original business data objects, providing a foundation for subsequent business data object recovery based on structural mapping relationships and structural constraint information.

[0050] Specifically, for attribute information that appears repeatedly in multiple business data objects, this implementation method centrally expresses the common attribute information in multiple business data objects based on the pre-determined shared expression relationship in the compressed expression structure. Specifically, firstly, the field names, field types, and field values ​​in multiple business data objects are compared to determine attribute information that remains consistent or has the same expressive meaning across different business data objects, and this type of attribute information is identified as common attribute information. For example, in multiple course objects, the course type field, subject field, and content type field may all have the same value; in multiple character objects, some status fields or configuration fields may all maintain the same status. For the aforementioned common attribute information, instead of writing it separately to each business data object, a corresponding shared expression element is generated according to the compressed expression structure, and the association relationship between the shared expression element and multiple business data objects is established. For example, common attribute information can be converted into a common attribute identifier, and compressed expression can be completed by having multiple business data objects reference the same common attribute identifier. During the recovery process, the scope of business objects corresponding to the shared expression element is determined according to the structural mapping relationship, and the shared attribute content is re-associated to the corresponding business data object. Through the above processing, the attribute content that appears repeatedly in multiple business data objects only needs to be saved once, which reduces data duplication while maintaining the correspondence between common attribute information and the original business objects.

[0051] For attribute information appearing according to a certain pattern in multiple business data objects, this implementation determines the corresponding encoding method based on the correlation between attribute values, rather than directly saving each independent attribute value. Specifically, for identifier attributes, sequence attributes, or other attribute information with a pattern in multiple business data objects, the change relationship between the corresponding attribute values ​​of different business objects is first analyzed to determine whether the attribute values ​​change continuously, at fixed intervals, or in a preset arrangement. For example, in multiple character objects, the object number attribute may change continuously in the order of 1, 2, 3, 4. If we save id=1; id=2; id=3; id=4 respectively, multiple independent expression positions will be required. Therefore, this implementation converts multiple independent identifier attribute values ​​into range expression information based on the continuous change relationship between identifier attribute values. For example, the consecutive numbers 1, 2, 3, 4 are converted into id="1-4", and the range expression information is saved as a compressed expression element in the compressed expression data. For identifier attribute values ​​that do not meet the continuous change condition, a list expression method can be used according to the number of attribute values ​​and the arrangement relationship, such as id="1,3,7", or a single value expression method can be used for a single attribute value. By selecting the corresponding encoding method based on the attribute change pattern, attribute information with regular change characteristics can be expressed in a more compact form. At the same time, the correspondence between the range expression information and the original business object is recorded through structural mapping relationship, so that the original attribute value corresponding to each business object can be determined according to the range boundary during the subsequent recovery process.

[0052] For situations where attribute information in multiple business data objects is reorganized according to attribute type, this implementation method converts the attribute information originally organized by the object dimension into an attribute sequence organized by the attribute dimension based on the attribute arrangement relationship in the compressed expression structure. Specifically, in the original business data object, each business object typically stores all attribute information on an object-by-object basis, for example: Object 1: Attribute A; Attribute B; Attribute C.

[0053] Object 2: Attribute A; Attribute B; Attribute C.

[0054] After performing attribute dimension transformation, attributes of the same type in multiple business objects can be expressed separately, for example: Attribute A sequence: A1, A2; Attribute B sequence: B1, B2; Attribute C sequence: C1, C2.

[0055] The above conversion method allows data with the same attribute type to be centrally arranged, facilitating further compression processing using a unified encoding method. For example, for multiple character objects, the character content can be uniformly converted into a character sequence, the numbering information into a number sequence, and the status information into a status encoding sequence, thereby avoiding the repeated expression of field structures that appear repeatedly in multiple objects.

[0056] Because attribute dimension transformation alters the arrangement of attribute information in the original business data object, object boundaries and attribute attribution relationships in the original object may not be directly reflected in the compressed attribute sequence. Therefore, this implementation method establishes a correspondence between attribute sequences and business data objects simultaneously during attribute dimension transformation. For example, the first character in the record character sequence corresponds to the first business object, the first number in the number sequence corresponds to the first business object, and the first code in the status code sequence corresponds to the first business object, thus forming attribute positional constraints. During the recovery process, the correspondence between different attribute sequence elements is determined based on the attribute positional constraints, and the attribute information corresponding to different attribute types is recombined into the corresponding business data objects, ensuring that the compressed representation data after attribute dimension transformation can still be restored to the same business object structure as before compression.

[0057] Based on the above processing, multiple business data objects, after being reorganized and encoded, are converted into compressed representation data. The compressed representation data includes multiple compressed representation elements organized according to a compressed representation structure. These compressed representation elements are used to represent compressed attribute information, attribute value information, or other business data content.

[0058] However, because the compression encoding process alters the original data organization of business data objects, transforming attribute information from its original object-dimensional expression into shared expression, attribute-centralized expression, or encoded expression, the compressed expression data itself may not fully reflect the structural relationships between the original business objects. If parsing relies solely on the field content of the compressed expression data, it may be impossible to accurately determine the scope of business objects corresponding to each compressed expression element and the attribute attribution relationships. For example, when common attribute information in multiple business data objects is converted into shared expression elements, the compressed expression data only stores one copy of the common attribute content, failing to directly indicate which business data objects this common attribute content should be associated with. When attribute information in multiple business data objects is centrally arranged according to attribute type—for example, unifying character content from multiple objects into a character sequence, unifying numbering information from multiple objects into a numbering sequence, or unifying state information from multiple objects into a state encoding sequence—while the elements in different attribute sequences maintain their original arrangement order, the corresponding object relationships between attribute elements may be hidden. When there are hierarchical relationships between multiple business data objects, the compressed expression may also fail to directly represent the containment relationship between parent and child objects. Therefore, in order to ensure that the compressed data can still be restored to the same structure as the original business data object, it is necessary to synchronously record the structural information describing the relationship between compressed elements during the compression encoding process.

[0059] Therefore, this embodiment further determines structural constraint information based on the arrangement and association relationships between data elements in the compressed expression structure. Specifically, when generating the compressed expression structure, the spatial arrangement relationship, attribute correspondence relationship, and object association relationship between different compressed expression elements are analyzed, and structural information affecting the subsequent recovery process is extracted to form structural constraint information. The structural constraint information is not used to store the attribute content of the business data object itself, but is used to describe the organization rules between different data elements in the compressed expression data, so that the server can determine the combination method between compressed elements based on the structural constraint information during the recovery process. For example, for compressed expression data after attribute dimension transformation, the structural constraint information can record the positional correspondence between character sequences, number sequences, and state sequences; for common attribute information after shared expression processing, the structural constraint information can record the business object range corresponding to the shared expression element; for business data objects with hierarchical relationships, the structural constraint information can record the association path between the parent object compressed element and the child object compressed element.

[0060] The structural constraint information is used to describe the structural relationships between different compressed expression elements in the compressed expression data, so as to maintain the association between the compressed expression data and multiple business data objects. Specifically, the element boundary relationship can be determined according to the positional relationship between different compressed expression elements in the compressed expression structure, the attribute positional relationship can be determined according to the attribute arrangement relationship corresponding to different compressed expression elements, and the object belonging relationship can be determined according to the association between the compressed expression elements and the business data objects.

[0061] The element boundary relationship is used to describe the data range of different compressed expression elements in the compressed expression data. For example, when the character content, number information, and status information in multiple character objects are converted into 'c', 'id', and 'cv' fields respectively, the element boundary relationship is used to record the data range and parsing boundary of each field. This allows the server to accurately distinguish different types of compressed elements when parsing the compressed expression data, avoiding confusion between field contents due to continuous data arrangement.

[0062] Attribute positional relationships describe the positional correspondence between different attribute values ​​after attribute reorganization. For example, after converting attribute information from multiple business data objects into attribute sequences, if the first element of the character sequence, the first element of the number sequence, and the first digit of the status code all correspond to the same business data object, then the attribute positional relationship is used to record: Character sequence position 1; Number sequence position 1; Status code position 1; Commonly corresponding object 1. In this way, even if the original object-dimensional data is converted into attribute-dimensional data, the correspondence between different attribute values ​​can still be determined based on the attribute positional relationship.

[0063] Object attribution is used to describe the association between compressed representation elements and business data objects. Specifically, when common attribute information in multiple business data objects is converted into a shared representation element, the object attribution relationship is used to record the range of multiple business objects corresponding to that shared representation element; when differential attribute information in multiple business data objects is encoded separately, the object attribution relationship is used to record the specific business object to which different differential representation elements belong; when there is a parent-child hierarchical relationship between business data objects, the object attribution relationship further records the connection relationship between parent objects, child objects, and corresponding attribute information.

[0064] By establishing the aforementioned structural constraints, the compressed representation data not only includes compressed attribute content but also preserves the structural relationships between attribute content, compressed elements, and business objects. During subsequent parsing of the compressed representation data, the server can determine the range of data to be parsed based on element boundary relationships, determine the correspondence between different attribute information based on attribute position relationships, and determine the business object to which the attribute information belongs based on object affiliation relationships, thereby achieving accurate recovery of the compressed representation data from the original business data object.

[0065] S3. Based on the structural mapping relationship and the structural constraint information, the compressed expression data is parsed to determine the attribute information and object association relationship corresponding to the compressed expression element. Then, multiple attribute information is combined and reconstructed according to the object association relationship to obtain the restored business data object.

[0066] In this embodiment, after compressing and encoding multiple business data objects, when the business processing needs to call the corresponding business data again, it is necessary to restore the business data objects that can participate in subsequent business processing based on the compressed expression data. Since the compressed expression data in this embodiment is not saved according to the original arrangement of the business data objects, but rather the attribute information in multiple business data objects is reorganized according to the compressed expression structure, there is no simple one-to-one correspondence between the compressed expression elements in the compressed expression data and the original business data objects.

[0067] For example, when common attribute information in multiple business data objects is shared and expressed, the attribute information originally contained separately in each business object may be converted into a common compressed expression element. In this case, the compressed expression data only stores the common attribute content once, and no longer retains the independent expression form of that attribute information in each business object. For instance, when multiple chapter objects have the same course type attribute, content type attribute, or status description attribute, after sharing and expressing, they can form a common compressed expression element, which is associated with multiple chapter objects. At this point, if the compressed expression data is read directly, only the attribute content corresponding to the common compressed expression element can be obtained, but it is impossible to directly determine which business objects this attribute content should be restored to. As another example, when attribute information in multiple business data objects is rearranged according to attribute type, multiple attribute values ​​of the same type may be stored consecutively in the same attribute sequence. For example, the character content of multiple character objects is grouped into a character sequence, the numbering information of multiple character objects is grouped into a numbering sequence, and the status information of multiple character objects is grouped into a status encoding sequence. In this case, although the elements in different attribute sequences maintain their original arrangement order, the object correspondence between the attribute values ​​is no longer directly reflected through the data arrangement. Therefore, if we only read the data content in the compressed expression data, although we can obtain some attribute information, we cannot determine the business data object to which each attribute information belongs, nor can we accurately restore the original organizational structure, hierarchical relationship and attribute association relationship between multiple business data objects.

[0068] Therefore, in this embodiment, when restoring compressed expression data, the attribute information corresponding to the compressed expression elements is first determined based on the structural mapping relationship. The structural mapping relationship describes the correspondence between compressed expression elements in the compressed expression structure and the original business data objects. It records not only the attribute type and content of the compressed expression element, but also the business object to which the attribute information belongs, the object level, and its positional relationship within the original business object. This allows the restoration process to determine the corresponding business data object and the attribute information to be restored based on the compressed expression element. Specifically, during the generation of the compressed expression structure, a corresponding mapping record is established for each compressed expression element. For example, for the 'c' field generated from the character attributes of multiple character objects, a mapping can be established: 'c' field → character object → 'char' attribute; for the 'id' field generated from the number attributes of multiple character objects, a mapping can be established: 'id' field → character object → 'id' attribute; for the 'cv' field generated from the status attributes of multiple character objects, a mapping can be established: 'cv' field → character object → 'checked' attribute. Through this mapping relationship, the server can clearly identify the original attribute information corresponding to each compressed expression element during the restoration process without needing to re-analyze the business meaning of the compressed data itself.

[0069] Specifically, based on the structural mapping relationship, each compressed expression element in the compressed expression data is parsed and converted into corresponding attribute information. For common compressed expression elements formed through shared expression, the common attribute type and associated object range corresponding to the compressed expression element are first determined according to the structural mapping relationship, and the common attribute content is restored to the corresponding multiple business data objects. For example, when a common compressed expression element is associated with multiple chapter objects, the server determines the set of chapter objects corresponding to the common attribute according to the mapping relationship, and associates the restored attribute information with the corresponding chapter objects respectively. For attribute sequences formed through attribute dimension transformation, the business data objects and attribute positions corresponding to each attribute value in the attribute sequence are determined according to the structural mapping relationship. For example, when the character content sequence: c="ABCD"; number sequence: id="1-4"; and status code: cv="1111" are stored separately, the server determines according to the structural mapping relationship that the first element of the character sequence corresponds to the first character object, the first number of the number sequence corresponds to the same character object, and the first digit of the status code corresponds to the status attribute of the same character object, thereby restoring the complete business objects corresponding to character A, number 1, and status true.

[0070] Furthermore, for business data objects with hierarchical relationships, the structural mapping relationship is also used to determine the object level corresponding to the compressed expression element. For example, after course objects, chapter objects, text objects, and character objects are compressed, the attribute information in objects at different levels may be reorganized. The server determines the course level, chapter level, text level, or character level corresponding to a certain compressed expression element based on the structural mapping relationship, and restores the association relationship between parent and child objects according to the original hierarchical relationship. Through the above processing, although the compressed expression data changes the expression form of the original business data object, it is still possible to accurately determine the attribute information and business object corresponding to the compressed expression element based on the structural mapping relationship, providing a foundation for subsequent attribute combination and reconstruction based on object association relationships.

[0071] After determining the attribute information corresponding to the compressed expression elements, this embodiment further combines the structural constraint information to determine the object association relationships between different attribute information. Because the original business data objects are reorganized during the compression encoding process—for example, converting common attribute information in multiple business objects into shared expression elements, converting similar attribute information in different objects into attribute sequences, or converting multiple state attributes into continuous encoding forms—although the recovered attribute information can already determine the corresponding attribute types, it is still necessary to further determine the combination relationships between each attribute information and the business object to which it belongs. The structural constraint information is used to describe the structural relationships synchronously retained during the compression process of the compressed expression data. It does not contain specific business attribute content, but is used to record the association rules between compressed expression elements, between attribute information, and between attribute information and business data objects.

[0072] Specifically, the structural constraint information includes at least the boundary relationships between compressed expression elements, the positional relationships between different attribute information, and the object association relationships between attribute information and business data objects. The boundary relationships describe the data range of different compressed expression elements in the compressed expression data, enabling the server to accurately determine the start and end positions of different compressed expression elements when parsing continuously arranged data content. The attribute positional relationships describe the arrangement position of attribute information after attribute reorganization within the original business object, enabling data of different attribute types to be matched according to their original positions during the recovery process. The object association relationships describe the business data object to which the attribute information belongs, allowing multiple attribute information stored separately or expressed independently to be recombine into a complete business object.

[0073] Specifically, based on element boundary constraint information, the data range corresponding to different compressed expression elements can be determined, avoiding boundary confusion during the parsing process of multiple consecutively arranged compressed expression elements. For example, when the character content, number information, and status information in multiple character objects are converted into 'c', 'id', and 'cv' fields respectively, the compressed expression data may consecutively store the data content corresponding to multiple fields. In this case, by recording the data range corresponding to each compressed expression element through element boundary constraint information, the server can determine which data belongs to character content, which data belongs to number information, and which data belongs to status information. For example, it can determine the character sequence range corresponding to the 'c' field, the number sequence range corresponding to the 'id' field, and the status encoding range corresponding to the 'cv' field, thereby avoiding attribute parsing errors caused by the consecutive arrangement of different compressed elements.

[0074] Based on attribute position constraints, the arrangement of each attribute within a business data object can be determined, allowing the reorganized attribute information to be restored to its corresponding position. Specifically, when attribute information in multiple business data objects is converted from object-dimensional representation to attribute-dimensional representation—for example, grouping the character content of multiple character objects into a character sequence: c="ABCD", grouping numbering information into id="1-4", and grouping status information into cv="1111"—the attribute information is no longer arranged according to the original object structure. Therefore, it is necessary to determine the positional correspondence between different attribute sequences based on attribute position constraints. For example, the first position in a character sequence corresponds to the character attribute of the first character object, the first position in a numbering sequence corresponds to the numbering attribute of the same character object, and the first position in a status code corresponds to the status attribute of the same character object. Through these positional constraints, data of different attribute types can still be matched according to the original business object structure even after independent encoding and grouping.

[0075] Based on object association constraint information, the business data object to which multiple attribute information belongs can be determined, enabling separately stored or independently expressed attribute information to be recombinated into the correct object. Specifically, when multiple business data objects are compressed, different attribute information within the same business object may be converted into different compressed expression elements. For example, the `char`, `id`, and `checked` attributes in a character object may correspond to a specific character in the `c` field, a specific number in the `id` field, and a specific encoding position in the `cv` field, respectively. In this case, parsing any single compressed expression element alone cannot determine the complete business object; the correspondence between these attributes needs to be recorded using object association constraint information. During the recovery process, based on the object association constraint information, it is determined that the Nth element of the `c` field, the Nth element of the `id` field, and the Nth encoding position of the `cv` field all correspond to the same business data object, and these multiple attribute information are further combined to form a complete object.

[0076] Furthermore, for multiple business data objects with hierarchical relationships, object association constraint information is also used to describe the relationships between objects at different levels. For example, in a course generation scenario, a course object contains chapter objects, a chapter object contains text objects, and a text object contains character objects. The object association constraint information can then be recorded as: Course object → Chapter object; Chapter object → Text object; Text object → Character object. During the recovery process, the object level to which different compressed expression elements belong is determined based on the above hierarchical relationships, and the parent-child object relationships are re-established according to the original business structure.

[0077] Through the above processing, this implementation method, when restoring compressed expression data, does not simply concatenate the data based on the attribute types corresponding to the compressed expression elements. Instead, it determines "which attributes to restore" by combining structural mapping relationships and "which object the attributes belong to and how they are combined" by combining structural constraint information. Therefore, even after the compressed expression data has undergone processing such as shared expression, attribute recombination, and encoding conversion, it can still accurately restore the attribute relationships, object relationships, and hierarchical relationships between multiple business data objects, achieving consistency between the compressed expression data and the original business data objects.

[0078] Furthermore, after determining the attribute information and object associations corresponding to the compressed expression elements, this embodiment combines and reconstructs multiple attribute information based on the object associations to obtain the restored business data object. Since the compression encoding process has changed the original representation of the business data object, some attribute information may be stored in a shared representation format, some in an attribute sequence format, and some in an independent and differentiated representation format. Therefore, during the restoration process, it is not possible to simply splice the compressed expression elements according to their order. Instead, it is necessary to determine the business objects to which different attribute information belongs based on the object associations and re-combine the attributes according to the original data organization method of the business objects.

[0079] Specifically, during object reconstruction, the original attribute types corresponding to each compressed expression element are first determined based on the structural mapping relationship, and the business data objects corresponding to the attribute information are determined based on the object association relationship. For example, for compressed expression elements after field abbreviation processing, the corresponding original attribute names can be recovered based on the field mapping relationship; for common attribute elements formed through shared expression, the common attribute information can be associated with multiple business data objects based on the object range corresponding to the shared expression element; for independent attribute elements formed through differentiated expression, their belonging object and corresponding attribute position are determined based on the differentiated expression relationship. Through the above processing, attribute information from different sources can be recombined according to the attribute composition relationship in the original business object, rather than being recovered solely based on the compressed storage format.

[0080] For example, for common attribute information shared by multiple business data objects, the same attribute content can be converted into shared expression elements during compression. When this shared expression element is parsed during recovery, the multiple business data objects corresponding to the common attribute information are determined based on object association relationships, and the common attribute information is associated with the corresponding objects. For instance, when multiple chapter objects share the same course type information and content type information, the recovery process determines the following relationships based on shared expression relationships: shared expression element A → chapter object 1; shared expression element A → chapter object 2; shared expression element A → chapter object 3. This restores the same common attribute content to the corresponding chapter object, ensuring that the recovered business data object retains its complete attribute structure.

[0081] For the attribute sequence after attribute dimension transformation, this implementation method reallocates multiple attribute values ​​in the attribute sequence to the corresponding business data objects according to the attribute position relationship and object association relationship. For example, after multiple character objects are compressed, the character content may form c="ABCD", the number information may form id="1-4", and the status information may form cv="1111".

[0082] During the recovery process, the number range corresponding to each character object is first determined based on the `id` field, and the corresponding status information for each object is determined based on the `cv` field. Then, the positional correspondence between character content, number information, and status information is determined based on attribute position constraints. For example, the first character in `c`, the first number in `id`, and the first status code in `cv` all correspond to the first character object. Through this method, the attribute information originally stored separately in different compressed expression elements is recombined to form a complete character business object composed of character content, object number, and status information.

[0083] For business data objects with hierarchical relationships, this implementation further determines the inclusion relationship between parent and child objects based on the object hierarchy mapping to restore business data objects with nested structures. For example, in a smart teaching scenario, a course object contains chapter objects, a chapter object contains text objects, and a text object contains character objects. During the restoration process, the chapter affiliation is first established based on the hierarchical mapping between course objects and chapter objects. Then, the chapter to which the text object belongs is determined based on the mapping between chapter objects and text objects. Finally, the character object set is restored based on the mapping between text objects and character objects. Through the above processing, the compressed business data object can still be restored as a nested object consistent with the original business structure.

[0084] Furthermore, during the recovery process, this embodiment can also form a parsing path from compressed expression elements back to business data objects based on structural mapping relationships and structural constraint information. The parsing path describes the parsing order, attribute association method, and object combination method of each compressed expression element during the compressed expression data recovery process, enabling the recovery process to determine not only the attribute content but also the corresponding business object and its position within the object.

[0085] Specifically, the parsing path can record the correspondence between compressed expression elements, corresponding attribute types, belonging business objects, attribute recovery positions, and object combination orders. For example, for the character object recovery process, the parsing paths can be: c field → char attribute → character object; id field → id attribute → character object; cv field → checked attribute → character object.

[0086] During the recovery process, the attribute information corresponding to each compressed expression element is first determined based on the parsing path. Then, the object to which the attribute information belongs is determined based on the object association relationship. Finally, the business data object is reconstructed according to the object combination order. For complex business objects, the recovery order between objects at different levels can also be determined based on the parsing path. For example, course objects can be recovered first, followed by chapter objects, then text objects, and finally character objects, thereby ensuring that the recovered business objects maintain their original organizational logic.

[0087] Through the above processing, this implementation method achieves the reconstruction of business data objects from compressed representation to their original object form. Unlike methods that rely solely on data decoding results for recovery, this implementation method establishes a structural mapping relationship during the compression encoding stage and synchronously generates structural constraint information based on the data organization relationships between business data objects. This ensures that although the compressed representation data undergoes processing such as shared representation, attribute reorganization, range encoding, and differential representation, it still retains the structural basis required for the recovery process. During the recovery stage, the server can not only obtain the corresponding attribute content based on the compressed representation elements but also determine the attribute source based on the structural mapping relationship and determine the attribute attribution and combination method based on the structural constraint information. This avoids problems such as loss of object boundaries, attribute mismatch, and destruction of hierarchical relationships caused by compression processing, achieving consistent recovery of business data objects before and after compression.

[0088] Preferably, in order to reduce data redundancy caused by duplicate attribute information in multiple business data objects, when generating the compressed expression structure, common attribute information can be determined according to the attribute expression relationship between multiple business data objects, and a shared expression relationship can be established based on the common attribute information.

[0089] In this implementation, multiple business data objects are typically used to describe different business entities under the same business type. Therefore, different business data objects usually have a certain degree of data similarity. For example, multiple business data objects generated in the same business scenario may simultaneously contain attribute information such as object type, status description, and business category, while different business data objects only differ in some business attribute values. If each business data object is stored separately in the traditional way, the same attribute information will be repeatedly stored as the number of business objects increases, causing a large amount of identical content to occupy additional data space.

[0090] To address the aforementioned issues, this implementation method does not directly compress multiple business data objects as a whole when generating the compressed expression structure. Instead, it first analyzes the attribute expression relationships between multiple business data objects to identify common attribute information with shared expressive value across different business data objects. Specifically, it performs correlation analysis on the attribute names, attribute types, and attribute value expression methods of multiple business data objects to determine whether the attribute information in different business data objects has the same data expression characteristics. When a certain attribute information in multiple business data objects maintains consistency in attribute type, expression method, and business meaning, that attribute information is identified as common attribute information.

[0091] After determining the common attribute information, a shared representation relationship is further established based on the distribution relationship of the common attribute information in different business data objects. Since the common attribute information originally existed in different business data objects, but only a unified representation is retained after sharing, it is necessary to record the correspondence between the common attribute information and multiple business data objects simultaneously. To this end, this embodiment generates corresponding attribute mapping information based on the reference relationship, attribute position relationship, and object association relationship of the common attribute information in different business data objects, and adds the attribute mapping information to the compressed representation structure.

[0092] For example, when multiple business data objects contain the same business type attribute, the business type attribute can be converted into a common expression element and represented by a compressed expression identifier. At the same time, the multiple business data objects corresponding to the compressed expression identifier are recorded through attribute mapping information, so that the corresponding attribute content can be determined according to the compressed expression identifier during the subsequent recovery process, and the attribute content can be re-associated to the corresponding business data object according to the attribute mapping information.

[0093] Furthermore, during compression encoding based on the aforementioned compressed expression structure, common attribute information in multiple business data objects can be converted into corresponding compressed expression identifiers according to the shared expression relationship, without needing to repeatedly save the same attribute content for each business data object. Thus, repetitive attribute information in multiple business data objects is converted into a single-object, multi-object association expression, reducing data redundancy while preserving the mapping relationship between common attribute information and business data objects, ensuring that the compressed data still has a clear basis for recovery.

[0094] Through the above processing, this implementation method solves the problem that traditional compression methods only focus on reducing data content while ignoring the structural relationships of business objects. It enables public attribute information to still be accurately located to the corresponding business data object after compression, providing a foundation for subsequent reconstruction of business data objects based on structural mapping relationships and structural constraint information.

[0095] Preferably, in order to further improve the ability to compress and express multiple business data objects under the condition of attribute organization changes, when generating the compressed expression structure, the attribute information organized according to the object dimension can be converted into an expression form organized according to the attribute dimension based on the attribute dimension relationship between multiple business data objects, and an index relationship between the attribute information and the business data objects can be established.

[0096] In this implementation, multiple business data objects are typically organized according to the object dimension, meaning each business data object corresponds to a complete set of attribute information. However, in scenarios where multiple business data objects are processed simultaneously, different business data objects often contain a large amount of attribute information of the same type. If each business data object is still stored separately according to the object dimension, the same type of attribute information will be scattered across different object representation structures, which not only increases the amount of duplicated content but also reduces the overall compression processing efficiency among multiple business data objects.

[0097] For example, when multiple business data objects all contain name, status, category, and time attributes, the traditional object-dimensional representation method requires recording the name, status, category, and time information for each business data object separately. However, when viewed as a whole, these attribute information actually exhibit clear attribute type patterns, and there are strong correlations between attribute information of the same type in different business data objects. Therefore, the attribute information in multiple business data objects can be reorganized according to attribute types, centralizing the expression of similar attribute information that was originally scattered across different business objects.

[0098] Specifically, this implementation classifies and analyzes the attribute information in multiple business data objects, determines the corresponding attribute set based on attribute name, attribute type, and attribute expression relationship, and aggregates multiple attribute information belonging to the same attribute type to form an attribute information sequence organized according to attribute dimensions. In this process, although the attribute information in multiple business data objects is transformed from an object-dimensional arrangement to an attribute-dimensional arrangement, the attribute information itself is not directly changed, but rather the organizational form of the attribute information is altered.

[0099] Because attribute dimension transformation hides the object boundary relationships in the original business data object, it is necessary to simultaneously establish an index relationship between attribute information and business data objects while building the attribute dimension representation structure. This index relationship describes the correspondence between attribute information and business data objects before and after attribute dimension transformation, ensuring that the rearranged attribute information can still be traced back to the original business data object.

[0100] For example, after grouping the status attributes of multiple business data objects into a status attribute sequence, the multiple status values ​​in the status attribute sequence may be stored consecutively according to the attribute arrangement order, which cannot directly reflect the business data object to which each status value belongs. In this case, an index relationship can be established based on the positional relationship of the status attributes in the original business data objects, so that during the recovery process, the business object corresponding to each status attribute can be determined according to the index relationship, and then recombined to form the original business data object.

[0101] Furthermore, when generating compressed representation data based on the attribute dimension representation, the attribute positions, object correspondences, and recovery order in the attribute sequence can be recorded according to the index relationship, so that the attribute dimension representation can be converted back to the object dimension representation according to the index relationship in the subsequent parsing process.

[0102] Through the above processing, this implementation method does not simply store the attribute information of multiple business data objects in a centralized manner, but retains the association path between attribute information and business objects while changing the attribute organization form, so that business data objects can be reversibly converted between object-dimensional expression form and attribute-dimensional expression form, thereby improving the efficiency of compression expression while maintaining the integrity of data structure.

[0103] Preferably, after generating the compressed expression structure, the shared expression and differential expression of multiple business data objects can be performed based on the relationship between the common expression part and the changed expression part among multiple business data objects.

[0104] In this implementation, multiple business data objects typically share some identical data content as well as some variable data content. If a completely independent representation is used, the identical parts will be repeatedly saved; if a completely merged representation is used, the differences between different objects may not be accurately distinguished. Therefore, it is necessary to represent the consistent data content and the changed data content separately based on the data change relationships between multiple business data objects.

[0105] Specifically, the changes in attribute information across multiple business data objects are analyzed to identify the common representations that remain consistent across the multiple business data objects, as well as the differences that vary depending on the object. The common representations describe the data content shared by the multiple business data objects, while the differences describe the data content that differs between different business data objects.

[0106] After determining the common and differential representation parts, the common representation parts are uniformly represented as shared representation objects, and independent representation relationships are established for the differential representation parts corresponding to different business data objects. These independent representation relationships describe the association between different differential representation parts and their corresponding business data objects, ensuring that while multiple business data objects share some representation content, they can still distinguish their unique data content.

[0107] For example, if multiple business data objects contain the same business rule information but differ in their corresponding business number, processing status, or time information, the business rule information can be used as a shared representation object and saved only once. At the same time, the business number, processing status, and time information can be used as different representation parts to establish corresponding relationships. This allows the recovery process to first use the shared representation object to form the basic object content, and then supplement the different attributes of the business objects based on the different representation parts.

[0108] Furthermore, when restoring business data objects, the common and variable expressions are recombined based on the association between shared and differing expressions to form a complete business data object. Through this process, the compression not only reduces redundant expressions but also preserves the differences between different business objects, enabling multiple similar business data objects to be compressed and expressed in a way that better suits their data characteristics.

[0109] Through the above processing, this embodiment utilizes the similarity and differences between business data objects to establish a hierarchical expression relationship, enabling the compressed expression structure to adapt to the actual situation where multiple business data objects have both commonalities and differences, thereby improving the expression flexibility in the compression encoding process.

[0110] Preferably, after generating compressed expression data, structural constraint information can be generated based on the arrangement and association relationships between data elements in the compressed expression structure to maintain the structural interpretability of the compressed expression data during the parsing process.

[0111] In this embodiment, after multiple business data objects are compressed and encoded, their original data organization typically changes. For example, common attribute information in multiple business data objects may be converted into shared expression elements, similar attribute information in multiple business data objects may be rearranged according to attribute dimensions, and multiple variable attribute information may be organized according to differential expression methods. While the above processing reduces repetitive content in business data objects, some structural relationships in the original business data objects are no longer directly reflected in the compressed data.

[0112] If only the compressed representation data itself is preserved, while some attribute content can be obtained during the recovery process, the boundary relationships, attribute positional relationships, and object affiliation relationships between different compressed representation elements cannot be determined. For example, when multiple attribute values ​​from multiple business data objects are arranged in a centralized manner, consecutively arranged attribute values ​​may belong to different business data objects; when multiple business data objects share the same common attribute information, a single common representation element cannot reflect the scope of objects it needs to be associated with. Therefore, it is necessary to synchronously retain information that can describe the above structural relationships during the compression process.

[0113] Based on this, this embodiment generates structural constraint information according to the arrangement and association relationships between data elements in the compressed expression structure. The structural constraint information is used to describe the correspondence between each compressed expression element in the compressed expression data and between the compressed expression elements and business data objects, so that the compressed expression data retains the structural basis required for recovery even after the original expression form is changed.

[0114] Specifically, based on the arrangement relationship between multiple compressed expression elements, the data boundaries corresponding to different compressed expression elements can be determined, and element boundary constraint information can be generated. The element boundary constraint information is used to represent the start position, end position, and separation relationship between adjacent elements of different compressed expression elements in the compressed expression data, enabling accurate identification of independent compressed expression elements during the parsing process.

[0115] For example, after performing attribute dimension transformation on attribute information from multiple business data objects, multiple attribute values ​​of the same type may be arranged consecutively according to attribute type to form an attribute sequence. Since the multiple attribute values ​​in the attribute sequence are no longer stored separately according to the original business data objects, if element boundary constraint information is missing, it may be impossible to accurately distinguish the data range corresponding to different compressed expression elements during the parsing process. Therefore, by recording the boundary relationships between compressed expression elements, erroneous connections between different data elements can be avoided during the recovery process.

[0116] Furthermore, based on the positional relationships between attribute information in the compressed representation structure, attribute position constraint information can be generated. This attribute position constraint information describes the arrangement of attribute information within the corresponding business data object, ensuring that the reorganized attribute information can be restored to its correct position within the original business object.

[0117] For example, when the state attributes of multiple business data objects are grouped into a state encoding sequence, the compressed representation data only stores consecutive encoded bits formed according to the encoding rules, and cannot directly reflect the business object and attribute position corresponding to each encoded bit. In this case, by recording the positional relationship between the encoded bits and the business data objects through attribute position constraint information, the object to which the state attribute belongs can be determined based on the corresponding position during the recovery process, thereby avoiding mismatches in state information between different business objects.

[0118] Furthermore, based on the association relationships between different attribute information and business data objects in the compressed expression structure, object association constraint information can be generated. This object association constraint information describes the range of business data objects corresponding to the compressed expression elements and the combination relationships between multiple attribute information, enabling separately stored or independently expressed attribute information to be recombinated into the corresponding business data objects.

[0119] For example, when multiple business data objects jointly reference a common attribute information, only one common expression element is retained in the compressed expression data. At this time, the object association constraint information is used to record the multiple business data objects corresponding to the common expression element. When the attribute information in multiple business data objects is converted into different compressed expression elements, the object association constraint information is used to record the business objects that need to be formed by combining the compressed expression elements.

[0120] Through the above processing, this implementation method simultaneously establishes structural constraint information in addition to compressing the data, ensuring that the compressed data not only contains the compressed data content but also retains the structural relationships between data elements. Therefore, during the subsequent recovery process, the location of compressed elements can be determined based on element boundary constraint information, the location of attribute information within an object can be determined based on attribute position constraint information, and the business object to which different attribute information belongs can be determined based on object association constraint information, thus forming a complete basis for data recovery.

[0121] Preferably, when generating structural constraint information, attribute attribution constraint information can also be determined based on the attribute combination relationship between multiple business data objects to further improve the accuracy of the object reconstruction process.

[0122] In this implementation, different attribute information in multiple business data objects typically follows certain combination rules. Multiple attribute information in the same business object jointly describe the corresponding business entity, while the attribute combination relationships between different business objects usually differ. If recovery is performed based solely on a single attribute information, incorrect combinations may occur due to the presence of the same type of attribute in multiple business objects.

[0123] Therefore, this implementation analyzes the attribute combination relationship in multiple business data objects, determines the combination relationship and attribute association order when multiple attribute information appears together, and generates attribute attribution constraint information based on the above relationship.

[0124] The attribute attribution constraint information is used to describe the attribution relationship between different attribute information, so that during the recovery process, it is possible to determine which business data object multiple attribute information should be combined to form based on the attribute combination relationship.

[0125] For example, multiple business data objects may contain name and status attributes. If recovery is performed solely based on the name or status attributes, it's possible to combine the name attribute of one business object with the status attribute of another. By using attribute attribution constraint information, the combination relationships between name, status, and other related attributes can be recorded, allowing attribute matching to be performed according to the original business object structure during the recovery process.

[0126] Through the above processing, this implementation method further enhances the interpretability of compressed expression data, enabling business data objects to be accurately recovered based on the retained structural constraint information even after shared expression, attribute conversion, and differential encoding.

[0127] Preferably, when restoring the compressed expression data, a parsing association path between the compressed expression elements and the business data objects can be established based on the structural mapping relationship and the structural constraint information, and the business data objects can be reconstructed based on the parsing association path.

[0128] In this embodiment, after the compressed expression data is compressed and encoded, its data organization method is different from that of the original business data object. Therefore, the recovery process cannot simply read in reverse order according to the data arrangement in the compressed expression data. Instead, it is necessary to determine the business meaning and recovery position of each compressed expression element based on the structural relationship preserved during the compression process.

[0129] Specifically, the structural mapping relationship describes the correspondence between the compressed representation structure and the original business data object, recording the mapping relationship between compressed representation elements and attribute information, and between attribute information and business data objects. The structural constraint information describes the structural constraints in the compressed representation data, recording the boundary relationships, attribute positional relationships, and object association relationships between compressed representation elements. By combining the structural mapping relationship and the structural constraint information, a parsed association path can be formed from the compressed representation elements back to the business data object.

[0130] The parsing association path does not represent a simple data reading order, but rather the correspondence between different technical objects during the recovery of compressed expression data. Specifically, the parsing association path includes the process of determining the attribute information corresponding to the compressed expression elements, the process of determining the business data objects corresponding to the attribute information, and the process of combining multiple attribute information to form business data objects.

[0131] For example, for compressed representation data after sharing common attributes, the recovery process first determines the common attribute information corresponding to the common compressed representation elements based on the structural mapping relationship. Then, it determines the range of business data objects that the common attribute information needs to be associated with based on the object association constraint information. Finally, the common attribute information is restored to the corresponding business data object. For compressed representation data after attribute dimension transformation, the attribute information in the attribute sequence is determined based on the structural mapping relationship. Then, the arrangement position of each attribute information in the business data object is determined based on the attribute position constraint information. Finally, the object reorganization is completed based on the object association relationship.

[0132] Furthermore, after determining the attribute information corresponding to the compressed expression elements, the attribute recovery order and object combination relationship can be determined based on the structural constraint information. Since different types of business data objects may have different data organization forms, some attribute information needs to be recovered first, while other attribute information needs to be determined in conjunction with other attribute information. Therefore, this embodiment determines the recovery order of multiple attribute information based on the attribute association relationships and object combination relationships recorded in the structural constraint information, enabling the recovery process to proceed according to the data organization logic before compression.

[0133] For example, when multiple business data objects have hierarchical nesting relationships, it is not possible to simply combine all attribute information flatly. Instead, it is necessary to first determine the parent business data object, and then determine the belonging relationship of the corresponding child business data objects according to the object hierarchy. At this time, the object hierarchy to which different compressed expression elements belong can be determined according to the object hierarchy mapping relationship, and the object to which each attribute information belongs can be determined according to the attribute belonging constraint information, thereby restoring the business data object with a nested structure.

[0134] Furthermore, to improve traceability during the recovery process, a bidirectional structural mapping relationship can be established. Specifically, during the generation of the compressed expression structure, a forward mapping is established from business data objects to compressed expression elements, enabling the compression process to determine the corresponding compressed expression method based on the business data objects. During the recovery process, a reverse mapping is established based on the correspondence between compressed expression elements and attribute information, allowing the recovery process to return the corresponding business data object based on the compressed expression elements.

[0135] Through the above two-way mapping relationship, a corresponding relationship can be established between the compressed expression state and the original expression state of the same business data object, so that any compressed expression element can be traced back to its source business data object, and at the same time, any business data object can also determine its corresponding compressed expression content.

[0136] Furthermore, after completing the attribute information parsing and object association determination, multiple attribute information is reconstructed at the object level based on the object association relationships. Specifically, based on the object identifier, attribute location, attribute association relationship, and object hierarchy relationship corresponding to the business data object, the parsed multiple attribute information is recombined to restore the data content after shared expression, attribute dimension transformation, or differentiated expression to the original business data object.

[0137] For example, when some attribute information in multiple business data objects is stored using a shared representation method, the shared attribute information is loaded into the corresponding business data objects according to the object association relationship corresponding to the shared representation elements; when the attribute information in multiple business data objects is stored using a differentiated representation method, the differentiated attribute information is supplemented into the corresponding shared representation object according to the mapping relationship between the differentiated representation part and the business object, forming a complete business data object.

[0138] Through the above processing, this embodiment does not rely on a complete copy of the data before compression during the data recovery process. Instead, it utilizes the structural mapping relationships and structural constraint information established during compression to recover the business data object along a pre-preserved parsing association path. Therefore, even if the business data object undergoes shared representation, attribute reorganization, or differentiated encoding processing, it can still maintain the consistency of data content, attribute relationships, and object structure.

[0139] Preferably, the compressed data can also be expressed in a range based on the relationship between the changes in the identifier attribute values ​​in multiple business data objects, so as to reduce the repetitive expression of continuous identifier information.

[0140] In this embodiment, business data objects typically contain identifying attribute information to distinguish different objects, such as object number, record sequence number, or business transaction identifier. When multiple business data objects are compressed simultaneously, some identifying attribute information may change continuously according to a certain pattern. If the corresponding identifying attribute value is still stored separately for each business data object, a large amount of continuously changing identifying information will be repeatedly expressed, reducing compression efficiency.

[0141] Therefore, this embodiment analyzes the relationship between the changes in identifier attribute values ​​in multiple business data objects when generating compressed representation data, and determines the set of identifier attribute values ​​that meet the continuous change condition. The continuous change condition can be used to determine whether there is a fixed change pattern among multiple identifier attribute values, such as adjacent identifier attribute values ​​increasing or decreasing according to a preset step size.

[0142] After determining the set of identifier attribute values ​​that meet the condition of continuous change, range expression information is generated based on the starting and ending identifier attribute values ​​in the set. This range expression information is then used to replace the representation of multiple independent identifier attribute values. Compared to storing each identifier attribute value individually, range expression information can represent multiple continuously changing identifier attribute values ​​with less data, thereby reducing the space occupied by identifier attribute information in compressed representation data.

[0143] Furthermore, to avoid losing the correspondence between business objects due to the range expression information only representing a numerical range, this embodiment also establishes an association relationship based on the arrangement relationship between the range expression information and the business data objects. For example, when multiple business data objects correspond to consecutive identifier attribute values ​​according to their generation order, the number of objects corresponding to the range expression information and the object arrangement order can be recorded, so that the identifier attribute value corresponding to each business data object can be determined based on the range expression information during the recovery process.

[0144] When restoring business data objects, the restoration range of the identifier attribute value is determined according to the range expression information, and the business data objects corresponding to different identifier attribute values ​​are determined in combination with the structural mapping relationship, thereby restoring the range expression form to the identifier attribute information corresponding to multiple independent business data objects.

[0145] Through the above processing, this embodiment utilizes the changing patterns between identifier attribute values ​​for compressed expression. While reducing the repeated occupation of continuous identifier information, it maintains the correspondence between identifier attribute information and business data objects, so that the data after range expression can still be accurately recovered.

[0146] Preferably, the compressed expression data can also generate continuous encoding bits according to the arrangement relationship of state attribute information in multiple business data objects, so as to realize the centralized expression of state attribute information.

[0147] In this implementation, the status attribute information in a business data object is typically used to represent the object's current state, such as business processing status, running status, or review status. When multiple business data objects exist simultaneously, the status attribute information corresponding to different objects may only contain a limited number of status types. If the status information of each business data object is still stored separately using complete attribute values, it will result in duplicate status representations.

[0148] Therefore, this implementation analyzes the status attribute information in multiple business data objects to determine the status type corresponding to different status attribute information, and establishes the positional correspondence between status attribute information and business objects based on the arrangement order of multiple business data objects.

[0149] Specifically, the state attribute information corresponding to different business objects can be converted into consecutive encoded bits according to the arrangement order of multiple business data objects, so that multiple state attribute information forms a state encoding sequence. The state encoding sequence does not directly store the complete state description content, but uses encoded bits to represent the corresponding state type, thereby reducing the expression length of state attribute information.

[0150] Because the status encoding sequence alters the organization of status attribute information in the original business data object, it is necessary to synchronously record the positional correspondence between the encoded bits and the business data object. This positional correspondence describes the business object and status attribute information corresponding to each encoded bit, enabling the corresponding business data object to be determined based on the encoded position during the recovery process.

[0151] During the recovery process, the state attribute information corresponding to each coding bit in the state coding sequence is determined according to the position correspondence, and the recovered state attribute information is loaded into the corresponding business data object to form a complete business data object.

[0152] Through the above processing, this embodiment converts the state attribute information in multiple business data objects into a continuous encoded expression. While reducing the expression of repetitive state content, it maintains the association between state information and business objects through position correspondence, avoiding mismatch of state attribute information during the recovery process.

[0153] Preferably, the compressed expression data can also generate a text fragment index based on the repetition relationship between text attribute information in multiple business data objects, so as to reduce the expression of repetitive text content.

[0154] In this implementation, some business data objects contain text attribute information, such as descriptions, remarks, or business instructions. Since multiple business data objects may be generated from the same business template, the text attribute information in different objects may contain a large number of repetitive text fragments. Saving the complete text content for each business data object separately would result in the same text content being stored multiple times.

[0155] Therefore, this implementation analyzes the text attribute information in multiple business data objects and determines the shareable text fragments based on the repeated text content and text boundary relationships among different text attribute information.

[0156] Specifically, content matching is performed on multiple text attribute information to determine the text content that appears repeatedly in different text attribute information, and the corresponding text boundaries are determined by combining the positional relationship of the text fragments in the original text. A text fragment index is generated based on the determined text fragments, and the corresponding repeated text content is represented by the text fragment index.

[0157] After generating the text fragment index, the text attribute information from multiple business data objects is converted into a serialized representation based on index references. At this point, the compressed representation data does not need to repeatedly save the same text content; only the text fragment index and its corresponding reference relationship need to be saved.

[0158] Furthermore, to ensure accurate recovery of text attribute information, this embodiment establishes an association between text fragment indexes and business data objects, enabling the determination of the business object and text location corresponding to a certain text fragment index during the recovery process.

[0159] When restoring business data objects, the corresponding text fragment is determined based on the text fragment index, and multiple text fragments are combined according to the original text boundary relationships to restore the complete text attribute information.

[0160] Through the above processing, this embodiment uses the repetitive relationships between text content for indexing and expression. While reducing the redundant storage of text attribute information, it retains the text content recovery path through text fragment indexing, so that the compressed text data can still be restored into a complete business object.

[0161] Preferably, when there are nested or hierarchical relationships among multiple business data objects, an object hierarchy mapping relationship can be established based on the hierarchical connection relationship between objects to maintain the structural consistency of complex business data objects before and after compression processing.

[0162] In this implementation, some business data objects are not composed of independent attribute information, but rather have a multi-level data organization. For example, a business data object can contain multiple sub-business data objects, and each sub-business data object can further contain corresponding attribute information. In such business scenarios, if only the attribute information at each level is compressed and expressed without preserving the inclusion relationship between objects at different levels, problems such as incorrect attribution of sub-objects, attribute information being attached to the wrong level, or the object structure being unable to be restored can easily occur during the recovery process.

[0163] Therefore, this implementation analyzes the object nesting relationships between multiple business data objects when generating the compressed expression structure to determine the connection relationships between business data objects at different levels. Specifically, based on the inclusion relationship, reference relationship, and attribute affiliation relationship between business data objects, the hierarchical connection relationship between parent business data objects and child business data objects is determined, and the hierarchy to which the attribute information corresponding to each child business data object belongs is further determined.

[0164] Based on the hierarchical connection relationships and attribute attribution relationships, an object hierarchical mapping relationship is established. This mapping relationship describes the correspondence between compressed expression elements and the hierarchical levels of business data objects, ensuring that the attribute information, even after compression, still identifies its object hierarchy. For example, when a parent business data object contains multiple child business data objects, the compressed expression structure can record the compressed expression elements corresponding to the parent object, the compressed expression elements corresponding to the child objects, and the hierarchical association between them, enabling the parent-child relationship to be re-established according to the original object structure during the recovery process.

[0165] Furthermore, during encoding based on the compressed expression structure, attribute information in different levels is associated and expressed according to the object hierarchy mapping relationship, ensuring that attribute information belonging to the same level of objects maintains a corresponding relationship, while attribute information between objects at different levels remains isolated. During the recovery process, the object hierarchy to which each compressed expression element belongs is determined according to the object hierarchy mapping relationship, and the corresponding attribute information is recombined according to the connection relationship between parent and child objects, thereby recovering the business data object with a nested structure.

[0166] For example, in a scenario where a business data object contains multiple sub-business objects, the business identification information in the parent object can form a complete business structure together with the business attribute information in the multiple sub-objects. Through the object hierarchy mapping relationship, the recovery process can first determine the parent object, and then load the corresponding attribute information into the correct sub-object according to the sub-object mapping relationship, avoiding the loss of hierarchy relationship due to the rearrangement of compressed attribute information.

[0167] Through the above processing, this implementation method not only restores the attribute content in the business data object, but also maintains the hierarchical organizational relationship between objects during the compression process. This allows complex business data objects to be reconstructed according to the original hierarchical structure after compression, thus improving the adaptability of business data object compression processing to complex data organization forms.

[0168] Preferably, based on the object hierarchy mapping, attribute attribution constraint information can also be generated according to the attribute combination relationship in multiple business data objects, so as to avoid attribute cross-combination between different business data objects during the recovery process.

[0169] In this implementation, when multiple business data objects are compressed simultaneously, even if the object hierarchy has been determined, different business data objects may still have the same type of attribute information. If the recovery process only matches based on attribute names or attribute types, it may incorrectly combine attribute information belonging to different business objects, resulting in a deviation between the recovered business data object and the original object.

[0170] Therefore, this embodiment further analyzes the attribute combination relationship in multiple business data objects, determines the combination rule corresponding to the co-occurrence of different attribute information, and generates attribute attribution constraint information based on the combination rule.

[0171] The attribute attribution constraint information is used to describe the attribution relationship between attribute information and business data objects, as well as the combination relationship between multiple attribute information within the same business data object. For example, when multiple business data objects contain both name and status attributes, it is impossible to determine which status attribute a certain name attribute should be combined with based solely on the attribute name. However, by analyzing the attribute combination relationships in the original business data objects, the correspondence between name attributes, status attributes, and other related attributes can be determined.

[0172] During the recovery process, the parsed attribute information is matched against the attribute attribution constraint information, allowing different attribute information to be re-associated according to the combination relationships in the original business data objects. For multiple business data objects with the same attribute type, the attribute attribution constraint information can be used to limit the attribute matching range, preventing attribute information from one business object from being incorrectly combined into another business object.

[0173] Through the above processing, this embodiment further supplements the attribution judgment mechanism in the business data object recovery process, so that after the compressed data has undergone attribute sharing, attribute reorganization and hierarchical transformation, it can still recover accurate business objects based on the preserved structural relationships.

[0174] Preferably, when there are duplicate attribute omissions or default attribute expressions among multiple business data objects, default recovery constraint information can be generated based on the attribute relationships among multiple business data objects to ensure that the recovered business data objects maintain a complete attribute structure.

[0175] In this implementation, some attribute information in a business data object does not always need to be explicitly expressed. For example, in the same type of business data object, some attribute information may have fixed default values, or some attribute information in multiple business data objects may maintain the same value. In such cases, if all attribute information is still fully stored for each business data object, it will result in a large amount of data that does not need to be repeatedly expressed.

[0176] Therefore, when generating compressed representation data, attribute information that does not need to be repeatedly written can be identified based on the attribute relationships between multiple business data objects, and converted into a default representation. Specifically, the attribute information in multiple business data objects is analyzed to determine attribute information that persists in different business objects but has the same representation content, or attribute information that can be derived from the structure of the business data objects, and this attribute information is determined as omitted attribute information.

[0177] After determining which attribute information to omit, the corresponding attribute content is not directly deleted. Instead, default recovery constraint information is generated based on the correspondence between the omitted attribute information and the business data object. This default recovery constraint information describes which attribute information was not directly written into the compressed data and how to supplement the corresponding attribute information during the recovery process based on existing data content or preset relationships.

[0178] For example, if a certain state attribute is represented by a default state in multiple business data objects, repeatedly saving this state attribute in each business object would increase the amount of data in the compressed representation. Therefore, the direct representation of this default state attribute can be omitted during the compression process, and the scope of the business objects corresponding to this attribute information and the recovery rules can be recorded through default recovery constraint information. During the recovery process, the attribute information that needs to be supplemented for the corresponding business data object is determined according to the default recovery constraint information, and the default state attribute is restored to the corresponding business object.

[0179] Furthermore, for cases where some attribute information is derived from the relationships between other attribute information, the attribute recovery logic can be maintained through default recovery constraint information. For example, when a certain attribute information in multiple business data objects can be determined based on object type, business category, or other recovered attribute information, it is not necessary to repeatedly save this attribute information in the compressed representation data. Instead, the corresponding recovery method can be determined based on the default recovery constraint information during the recovery process.

[0180] Through the above processing, this implementation method does not simply delete some attribute information from the business data object. Instead, it retains the corresponding recovery constraints while omitting attribute expression content, enabling the compressed expression data to maintain a complete data structure while reducing redundant expressions. Therefore, even if some attribute information is not directly stored in the compressed expression data, it can still be accurately supplemented into the corresponding business data object based on the default recovery constraint information, avoiding attribute loss issues in the recovered business data object.

[0181] Preferably, the data object compression encoding method provided in this application can be applied to an intelligent teaching content generation platform based on a large language model. In this application scenario, the online education platform calls the large language model based on the user's input of course learning requests, learning stage information, and interaction requirements. The large language model generates corresponding teaching business data, and the generated results are returned to the client via the SSE streaming protocol on the server side. At the same time, the finally processed business data is saved to a cache environment for course display, learning analysis, and subsequent teaching service calls. Since the generated results in intelligent teaching scenarios are usually not single text, but contain multiple business data objects such as course objects, chapter objects, text objects, and character objects, the generated data usually has the characteristics of complex hierarchical relationships, a large number of objects, and many repetitive attributes.

[0182] For example, in an English course generation scenario, the large language model generates structured JSON data based on the course content. Course objects can include a course type field (curClassType), a subject field (subject), an expanded state field (isExpand), a character count field (char_count), a reading count field (reading_count), a selected count field (selectedCount), an expression field (expression), a speech field (speech), a title field (title), and a content type field (content_type). Text objects can include a text name field (text_name), a total number of items field (total_items), and a text identifier field (text_id). The character object collection is organized by a words field, and each character object further includes attributes such as a character content field (char), an object number field (id), a selected state field (checked), an added state field (isAdd), a corrected state field (isCorrection), a user state field (isCustomer), an edited state field (isEdit), and a corrected state field (isRight). In traditional generation methods, these field names and object structures will repeat as the number of objects increases. For example, multiple character objects may require repeated output of fields such as char, id, and checked, resulting in a large number of duplicate tokens during the large language model generation process and increasing the amount of data transmitted via SSE streaming.

[0183] To address the aforementioned issues, this embodiment generates a compressed expression structure during the business data object generation process based on the data organization relationships and attribute expression characteristics among multiple business data objects. It also establishes corresponding data expression methods based on the repetition, variation, and object association relationships among different attribute information. For example, course type, subject information, and content type information shared by multiple chapter objects can be identified as common expression parts, and a shared expression object can be established, allowing multiple business objects to jointly reference the same common expression content. Changes in learning status, selection status, and processing results for different users are recorded independently as differential expression parts. Through this method, fixed information in business data objects does not need to be repeatedly saved in each object, while changes between different objects can still be expressed independently.

[0184] Furthermore, for nested relationships among course objects, chapter objects, text objects, and character objects, this embodiment establishes an object hierarchy mapping relationship based on the inclusion relationship between objects. For example, if a course object corresponds to multiple chapter objects, a chapter object corresponds to multiple text objects, and a text object corresponds to multiple character objects, then a hierarchy mapping relationship of course object → chapter object → text object → character object is established. During compression, even if attribute information in objects at different levels is reorganized or centrally expressed, the object level to which the compressed element belongs can still be determined based on this object hierarchy mapping relationship, enabling the parent-child object relationship to be re-established according to the original business structure during the restoration process. Simultaneously, for multiple attribute information in objects at the same level, attribute attribution constraint information is generated based on field combination relationships, allowing attributes such as character content, number information, and status information to be re-matched to the corresponding character object, avoiding attribute cross-combinations between multiple objects.

[0185] After forming the above compressed expression structure, this embodiment further compresses and encodes the business data objects according to the attribute type and attribute expression characteristics. For course basic fields, field name expression can be reduced through field abbreviation mapping. For example, a mapping relationship can be established as follows: curClassType → cct, subject → sub, isExpand → exp, selectedCount → sc, content_type → ct, total_items → tot, words → w. This allows the large language model output stage to directly use short field representations, while the server saves the corresponding structure mapping relationship so that the original business attributes can be restored later based on the compressed fields. For the character object array words, since multiple character objects contain the same type of attribute, the original object-dimensional expression method is converted into an attribute-dimensional expression method. The character content in multiple character objects is uniformly converted into a c field, the number information is uniformly converted into an id field, the status information is uniformly converted into a cv field, and the pinyin or annotation information is uniformly converted into a py field, thus forming a compressed expression structure composed of c / id / cv / py.

[0186] For example, the character content A, B, C, D in multiple character objects can be converted to c="ABCD"; multiple consecutive numbers 1, 2, 3, 4 can be converted to id="1-4"; and multiple status messages true, false, true, true can be converted to cv="1011". Through the above encoding method, the repeated field names and object structures in multiple business data objects are converted into attribute set expressions, while maintaining the correspondence between compressed elements and original objects through field position relationships and object association relationships.

[0187] For text-based business data, this embodiment further generates a text fragment index based on the repetitive content among multiple text attributes. For example, when multiple character objects or multiple text objects contain the same pinyin, annotations, or descriptive text, the repetitive text content is extracted as a shared text fragment and referenced through an index, so that multiple business objects do not need to repeatedly save the same text content. During the recovery process, the corresponding text content is determined based on the text fragment index, and the complete text attribute is regenerated by combining the text boundary relationships.

[0188] For identifier attribute information, this embodiment determines the corresponding expression method based on the identifier change relationship among multiple business data objects. When multiple object numbers change continuously, a range expression method is used, for example, id="1-4" represents 1, 2, 3, 4. When the numbers are not continuous, id="1,3,7" is used to represent multiple independent numbers; when only one number exists, a single value expression is used. Through the above method, continuously changing identifier attribute information is transformed from multiple independent fields into range expression information, while the correspondence between the number and the business object is maintained by using structural mapping relationship.

[0189] For status attribute information, this embodiment generates consecutive encoded bits based on the arrangement relationship between multiple business objects. For example, multiple character objects corresponding to status information true, false, true, true are converted to cv="1011". Each bit of the encoding corresponds to a character object. The positional relationship between the encoded bits and the business objects is recorded through attribute position constraint information, so that the object to which each status information belongs can be accurately determined during the recovery process.

[0190] After the large language model completes the compressed output and sends it to the server via the SSE protocol, the server reconstructs the business data object based on the structural mapping relationship and structural constraint information. Specifically, the server first performs reverse conversion based on the field mapping relationship, for example, restoring cct to curClassType; sub to subject; and exp to isExpand. Then, it parses the attribute set in the compressed data, restoring character content based on the 'c' field, number information based on the 'id' field, status information based on the 'cv' field, and annotation information based on the 'py' field, and recombines them into a complete character object based on the object association constraint information. For default attributes not directly expressed during compression, such as fixed status fields or default configuration fields, the corresponding attributes are automatically supplemented based on the default restoration constraint information, ensuring that the restored business data object maintains its complete structure.

[0191] After restoring the business objects, the server can generate subsequent teaching business data based on the restored object set. For example, it can generate a result array `result` based on character objects and maintain information such as object serial number, learning status, and processing result. Subsequently, the restored complete JSON data is written to the Redis cache, allowing the client and downstream teaching business modules to still obtain complete data according to the original business interface without being aware that a compressed expression method was used in the large language model generation stage.

[0192] In this application scenario, business data objects undergo three stages of data representation transformation: First, a compressed representation structure is used during the large language model generation and SSE streaming stages, reducing the output data volume through field abbreviations, attribute aggregation, range representation, and index referencing. Second, the server internally performs parsing and reconstruction based on structural mapping relationships and structural constraint information. Third, the data is restored to its complete business data object format during cache storage and downstream business consumption. Furthermore, when exceptions occur during field mapping, object reconstruction, or cache writing, the server can perform degradation processing, saving the original compressed data to the cache and recording exception information to ensure the continuous operation of the teaching business process.

[0193] Through the above application scenarios, this application extends the business data object compression processing from traditional data stream compression to the stage of generating structured business data from large language models. While reducing redundant information in the model output, it maintains the attribute relationships, hierarchical relationships, and recovery paths between business objects, so that the compressed business data can still be restored to a complete object structure that meets the needs of downstream business without loss.

[0194] like Figure 3 As shown, this application also provides a data object compression encoding system, including a business data generation server, a compression expression processing server, a compression data management server, and a business data recovery server; The business data generation server is used to acquire multiple business data objects and send the multiple business data objects to the compressed expression processing server; The compressed expression processing server is used to generate a corresponding compressed expression structure based on the data organization relationship and attribute expression characteristics among the multiple business data objects, establish a structural mapping relationship between the compressed expression structure and the multiple business data objects, and compress and encode the multiple business data objects based on the compressed expression structure to obtain compressed expression data, while generating structural constraint information to maintain the association relationship between the compressed expression data and the multiple business data objects. The compressed data management server is used to store the compressed expression data and the corresponding structural mapping relationship and structural constraint information; The business data recovery server is used to acquire the compressed expression data, the structural mapping relationship, and the structural constraint information, parse the compressed expression data, determine the attribute information and object association relationship corresponding to the compressed elements, and combine and reconstruct multiple attribute information according to the object association relationship to obtain the recovered business data object.

[0195] In summary, the data object compression encoding method and system provided in this application, by analyzing the data organization relationships and attribute expression characteristics among multiple business data objects, establishes a compressed expression structure suitable for the characteristics of business data, and combines structural mapping relationships and structural constraint information to achieve reversible conversion between compressed data and original business objects. This reduces the length of business data expression while maintaining the attribute relationships, object relationships, and hierarchical relationships between business data objects, specifically offering the following beneficial effects: First, this application improves the ability to utilize data duplication features during the compression processing of business data objects, reducing data redundancy caused by duplicate attribute information. In existing business data processing, multiple business data objects are usually generated from the same business template or the same processing flow, thus often containing a large amount of duplicate attribute information between different business objects. For example, in the scenario of intelligent teaching content generation, multiple chapter objects may have the same course type, subject information, and content type, and multiple character objects may have the same status fields, attribute structures, and field composition forms. If a traditional object-level storage method is used, each business data object needs to independently save its complete field structure. Even if some attribute contents in multiple objects are completely identical, the same field names and field values ​​still need to be repeatedly expressed.

[0196] To address the aforementioned issues, this application analyzes the redundancy of attribute information among multiple business data objects, identifies shared attribute information, and generates corresponding shared representation elements based on the compressed representation structure. This eliminates the need for multiple business data objects to separately store identical attribute content, instead using a unified representation through a common method. For example, common attributes such as course type, content type, or status description shared by multiple business objects are converted into shared representation objects, and the scope of business objects corresponding to these shared representation objects is recorded through structural mapping relationships. This process reduces the space occupied by duplicate attribute fields in the compressed representation data, lowering the overall length of the business data object representation. Furthermore, since the common attribute information is not simply deleted but preserved through shared representation relationships, the scope of business objects corresponding to this common attribute information can still be accurately determined during subsequent recovery, preventing the loss of business information due to compression processing.

[0197] Secondly, this application improves the adaptability of differentiated compression processing for different types of attribute information, avoiding the problem of reduced compression effect caused by using a single compression method.

[0198] In real-world business scenarios, different attribute information exhibits different data characteristics. For example, identifier attributes typically exhibit continuous change patterns, status attributes usually have a limited range of values, text attributes often contain repetitive content or similar expressions, while hierarchical attributes are used to describe the organizational relationships between objects. If a uniform data representation method is used to process all attribute information, the inherent data patterns of different attribute information cannot be fully utilized. To address these issues, this application determines corresponding representation methods based on the data characteristics of attribute information. For identifier attribute information with continuous change patterns, range representation information is generated based on the changing relationships between multiple attribute values. For example, consecutive numbers 1, 2, 3, 4 are converted to id="1-4", thereby reducing the repetitive expression of independent identifier fields by utilizing the continuous relationships between attribute values. For status attribute information, continuous encoding information is generated based on the arrangement relationships between multiple business objects. For example, multiple object-corresponding states true, false, true, true are converted to cv="1011". Through the correspondence between encoding bits and object positions, multiple status attributes can be expressed in a more compact form. For text attribute information, common text fragments are extracted based on the repetition relationship between multiple text contents, and referenced by indexing to reduce the repeated saving of duplicate text content in multiple business objects.

[0199] Through the above processing, this application does not apply the same compression rules to all business data. Instead, it selects the corresponding compression method according to the attribute type, change pattern, and expression characteristics, so that different attribute information can be effectively compressed according to its own characteristics, thereby improving the pertinence and effectiveness of compression processing.

[0200] Furthermore, this application maintains the structural integrity of business data objects during the compression process, avoiding the problems of object boundary loss and attribute mismatch caused by traditional compression methods.

[0201] Because this application may perform shared representation, attribute dimension transformation, and differentiated representation processing on the original business objects during the compression process, the compressed data organization may differ from the original object structure. For example, the character content, numbering information, and status information of multiple character objects may be stored separately in different compressed representation elements after attribute dimension transformation; common attribute information of multiple business objects may only retain one common representation element after shared representation. If recovery is based solely on the compressed representation data itself, it is impossible to directly determine the business objects to which different attribute information belongs, nor can the combination relationship between multiple compressed elements be determined, which can easily lead to attribute mismatches or structural changes in the recovered object.

[0202] To address the aforementioned issues, this application establishes a structural mapping relationship while generating the compressed expression structure, and generates structural constraint information based on the positional, arrangement, and association relationships between the compressed expression elements. The structural mapping relationship describes the correspondence between compressed expression elements and the original business objects, enabling the server to determine the attribute type and the business object to which a given compressed element belongs. The structural constraint information describes element boundary relationships, attribute positional relationships, and object affiliation relationships, ensuring that the reorganized attribute information can still be restored according to the original object structure. For example, for character object data after attribute dimension transformation: c="ABCD", id="1-4", cv="1111", this application determines, based on the attribute positional constraint relationship, that the first element of the c field, the first number of the id field, and the first status code of the cv field all correspond to the first character object. This avoids misalignment between different attribute sequences, ensuring that the restored business object maintains the original attribute correspondence.

[0203] Furthermore, this application improves the accuracy of recovering complex business data objects, enabling data objects with hierarchical relationships to maintain their original organizational structure.

[0204] In actual business processing, multiple business data objects often have hierarchical relationships. For example, a course object contains chapter objects, chapter objects contain text objects, and text objects contain character objects. Traditional compression methods usually focus on reducing data length, but struggle to maintain the hierarchical relationships between complex objects simultaneously.

[0205] To address this issue, this application establishes an object hierarchy mapping relationship based on the object inclusion relationships among multiple business data objects. During compression, the connection relationships between parent and child objects are recorded. During recovery, the object hierarchy mapping relationship is used to determine the object level to which different compressed expression elements belong, and the object associations are re-established according to the original hierarchical structure. For example, in a smart teaching scenario, the hierarchical relationship between course objects → chapter objects → text objects → character objects can still be recovered through the object hierarchy mapping relationship after compression, ensuring that the recovered business data objects not only contain correct attribute information but also maintain the original business organizational structure. Therefore, this application not only achieves data content recovery but also business semantic structure recovery, improving the usability of compressed data in complex business scenarios.

[0206] Furthermore, this application improves the compatibility between compressed data and business systems, and reduces the cost of business system transformation.

[0207] Existing compression solutions typically require business systems to perform additional parsing and processing for the compressed format, necessitating client-side or downstream business modules to adapt to the new data structure. This application, however, establishes a structure mapping relationship during the compression process and completes the parsing of compressed data and reconstruction of business objects on the server side, ensuring that the compression process primarily occurs during the data generation and transmission stages. For example, in a scenario where a large language model generates teaching data, the large language model can output compressed fields cct, sub, exp, and w. The server restores curClassType, subject, isExpand, and words based on the field mapping relationship and ultimately writes the complete JSON business data to a Redis cache. Clients and downstream teaching business modules still obtain the complete business object using their original business interfaces, without needing to modify the compressed representation.

[0208] By employing the above methods, this application reduces the amount of model output data and transmission costs while maintaining compatibility with existing business system interfaces, thereby improving the ease of deployment of the technical solution in actual business environments.

[0209] Furthermore, this application improves processing efficiency in large-scale business data processing. In business data generation scenarios based on large language models, the amount of output data directly affects model generation time, streaming efficiency, and server processing pressure. This application reduces redundant expressions in structured business data through various methods such as field abbreviation, shared public attributes, range representation, status encoding, and text indexing, enabling large language models to generate business results in a more compact data format.

[0210] Meanwhile, since this application does not use simple deletion or irreversible compression, but instead saves the structural mapping relationship and structural constraint information synchronously during the compression process, it can ensure the accurate recovery of business objects while reducing the data size, thus achieving a balance between compression efficiency and data integrity.

[0211] Therefore, this application establishes a compressed expression structure oriented towards the characteristics of business data objects, and combines structural mapping relationships and structural constraint information to achieve efficient compressed encoding and accurate reconstruction of business data objects. This not only reduces data redundancy caused by repetitive attribute expressions and improves the efficiency of structured business data generation and transmission, but also maintains the attribute relationships, association relationships, and hierarchical relationships between business objects. It is suitable for large language model structured output, intelligent teaching data generation, and other application scenarios that require batch processing of complex business data objects.

Claims

1. A data object compression encoding method, characterized in that, include: Multiple business data objects are acquired, and a corresponding compressed expression structure is generated based on the data organization relationship and attribute expression characteristics among the multiple business data objects. A structural mapping relationship is established between the compressed expression structure and the multiple business data objects. The multiple business data objects are compressed and encoded based on the compressed expression structure to obtain compressed expression data. Structural constraint information is determined according to the arrangement and association relationships between data elements in the compressed expression structure to maintain the association relationship between the compressed expression data and the multiple business data objects.

2. The data object compression encoding method according to claim 1, characterized in that, The method further includes: Based on the structural mapping relationship and the structural constraint information, the compressed expression data is parsed to determine the attribute information and object association relationship corresponding to the compressed expression element. Then, multiple attribute information is combined and reconstructed according to the object association relationship to obtain the restored business data object.

3. The data object compression encoding method according to claim 1, characterized in that, The generation of the corresponding compressed expression structure includes: The attribute information in the multiple business data objects is traversed to obtain the attribute name, attribute type and attribute value expression relationship of the corresponding attributes in different business data objects, and the common attribute information used for shared expression is determined based on the repeated occurrence relationship of attribute information among multiple business data objects. Based on the attribute position, attribute reference relationship and attribute arrangement order of the public attribute information in different business data objects, establish the attribute mapping relationship between the public attribute information and the compressed expression identifier; Based on the attribute mapping relationship, common attribute information in multiple business data objects is converted into corresponding compressed expression identifiers, and the correspondence between the compressed expression identifiers and the original attribute positions is stored in the compressed expression structure, so that when the compressed expression data is restored, the original attribute information corresponding to the compressed expression identifier can be determined according to the attribute mapping relationship.

4. The data object compression encoding method according to claim 1, characterized in that, The generation of the corresponding compressed expression structure includes: The attribute information in the multiple business data objects is classified and processed to determine the set of attribute values ​​with the same attribute type in the multiple business data objects. Based on the arrangement relationship of business data objects corresponding to different attribute values ​​in the attribute value set, the original attribute values ​​organized according to business data objects are converted into a sequence of attribute values ​​organized according to attribute type; An index relationship is established between the attribute values ​​in the attribute value sequence and the corresponding business data objects, so that after multiple business data objects undergo attribute dimension transformation, the original correspondence between the business data objects can still be restored based on the index relationship.

5. The data object compression encoding method according to claim 1, characterized in that, The obtained compressed expression data includes: Obtain the identification attribute values ​​from the multiple business data objects, analyze the numerical continuity relationship between the multiple identification attribute values, and determine the set of identification attribute values ​​that meet the condition of continuous change. Range expression information is generated based on the start and end identifier attribute values ​​in the set of identifier attribute values, and an association is established between the range expression information and the corresponding business data object; The expression method of replacing multiple independent identifier attribute values ​​based on the range expression information is used to reduce the duplication of identifier attribute values ​​in the compressed expression data, and the corresponding identifier attribute values ​​are restored according to the range expression information during the recovery process.

6. The data object compression encoding method according to claim 1, characterized in that, The obtained compressed expression data includes: Obtain the status attribute information from the plurality of business data objects, and determine the arrangement position of the status attribute information according to the arrangement order among the plurality of business data objects; Multiple state attribute information is converted into consecutive encoded bits according to the arrangement position, and a state encoding sequence is formed according to the positional correspondence between the consecutive encoded bits and the business data object; When restoring the business data object, the state attribute information corresponding to each encoding bit in the state encoding sequence is determined according to the position correspondence, so as to restore the state information corresponding to different business data objects.

7. The data object compression encoding method according to claim 1, characterized in that, The obtained compressed expression data includes: Obtain text attribute information from the multiple business data objects, analyze the repeated text content and text boundary relationships between different text attribute information, and determine the text fragments that can be shared and expressed. A text segment index is generated based on the reference relationships between the text segments, and multiple text attribute information is serialized and expressed according to the text segment index; The serialized expression result is associated with the corresponding business data object to reduce the repetitive expression of text attribute information while recovering the complete text attribute information in the corresponding business data object based on the text fragment index.

8. The data object compression encoding method according to claim 1, characterized in that, The structural constraint information includes element boundary constraint information, and the generation of structural constraint information includes: The boundary positions between adjacent compressed expression elements are determined based on the arrangement relationship between multiple compressed expression elements in the compressed expression structure. Element boundary constraint information is generated based on the boundary positions, and the element boundary constraint information is associated with the compressed representation data; When parsing the compressed expression data, independent compressed expression elements are determined based on the element boundary constraint information to avoid boundary mismatches between the corresponding attribute information of different business data objects during the recovery process.

9. The data object compression encoding method according to claim 8, characterized in that, The structural constraint information includes attribute location constraint information, and the generation of structural constraint information includes: Based on the arrangement relationship between multiple attribute information in the compressed expression structure, the positional relationship of different attribute information in the corresponding attribute sequence is determined; Attribute position constraint information is generated based on the positional relationship, and the attribute position constraint information is associated with the compressed expression data; During the recovery process, the correspondence between different attribute information is determined based on the attribute location constraint information, so that the multiple attribute information encoded separately can be mapped to the correct business data object.

10. The data object compression encoding method according to claim 9, characterized in that, The structural constraint information includes object association constraint information, and the generation of structural constraint information includes: Based on the combination relationship between different attribute information in the multiple business data objects, determine the attribution relationship between attribute information and business data objects; Based on the attribution relationship, object association constraint information is generated, and the object association constraint information is associated with the compressed representation data; During object reconstruction, attribute information stored in different compressed expression structures is recombined according to the object association constraint information to restore multiple business data objects corresponding to those before compression.

11. The data object compression encoding method according to claim 1, characterized in that, The compressed expression structure includes encoding rule information, and generating the corresponding compressed expression structure includes: Based on the attribute types, attribute arrangement relationships, and attribute value expression characteristics in the multiple business data objects, determine the encoding expression method corresponding to different attribute information; Generate a corresponding compressed expression structure based on the encoding expression method, and associate the encoding expression method with the structure mapping relationship so that different types of attribute information are compressed and expressed in the corresponding way; During the recovery process, the parsing method for the corresponding attribute information is determined based on the encoding expression method.

12. The data object compression encoding method according to claim 2, characterized in that, The step of parsing the compressed representation data based on the structural mapping relationship and the structural constraint information includes: The attribute information corresponding to the compressed expression elements in the compressed expression data is determined based on the structural mapping relationship. The positional relationships between different attribute information and their respective business data objects are determined based on the structural constraint information. Based on the location relationship and the business data object to which it belongs, multiple attribute information is combined to recover the recovery business data object corresponding to the multiple business data objects.

13. The data object compression encoding method according to claim 12, characterized in that, The step of parsing the compressed representation data based on the structural mapping relationship and the structural constraint information includes: The correspondence between the compressed expression elements in the compressed expression data and the multiple business data objects is determined based on the structural mapping relationship; Determine the attribute recovery order, attribute association, and object composition relationship in the correspondence based on the structural constraint information; According to the attribute recovery order, attribute association, and object combination relationship, multiple compressed expression elements in the compressed expression data are parsed and combined to maintain the structural consistency between the compressed expression data and the recovered business data object.

14. A data object compression encoding system, characterized in that, This includes a business data generation server, a compressed data processing server, a compressed data management server, and a business data recovery server. The business data generation server is used to acquire multiple business data objects and send the multiple business data objects to the compressed expression processing server; The compressed expression processing server is used to generate a corresponding compressed expression structure based on the data organization relationship and attribute expression characteristics among the multiple business data objects, establish a structural mapping relationship between the compressed expression structure and the multiple business data objects, and compress and encode the multiple business data objects based on the compressed expression structure to obtain compressed expression data, while generating structural constraint information to maintain the association relationship between the compressed expression data and the multiple business data objects. The compressed data management server is used to store the compressed expression data and the corresponding structural mapping relationship and structural constraint information; The business data recovery server is used to acquire the compressed expression data, the structural mapping relationship, and the structural constraint information, parse the compressed expression data, determine the attribute information and object association relationship corresponding to the compressed elements, and combine and reconstruct multiple attribute information according to the object association relationship to obtain the recovered business data object.

Images