Data processing method and device, equipment, medium and program product

Abstract

The embodiment of the invention discloses a data processing method and device, equipment, a medium and a program product. The method comprises the steps of obtaining at least one to-be-processed data source in response to a data source access operation; in response to an operator selection operation triggered in an operator warehouse based on a data processing demand, the selected target operator is displayed in an editing interface, each operator in the operator warehouse is a pre-packaged data processing operation, and different operators correspond to different data processing operations; in response to an editing operation triggered for the target operator in the editing interface, generating a data processing stream meeting a data processing requirement based on the target operator; and processing the data generated by the at least one data source by executing the data processing stream. Through the technical scheme of the embodiment of the invention, the flexibility of data processing can be improved, codes do not need to be written by relying on developers, the development efficiency is improved, and the development cost is reduced.

Description

Technical Field

[0001] The embodiments of the present invention relate to computer technology, and more particularly to a data processing method, apparatus, device, medium, and program product. Background Technology

[0002] With the continuous improvement of informatization, massive amounts of heterogeneous data assets are accumulating, and these data are complex and come from diverse sources. When processing this data, developers typically write code for the entire data processing procedure, executing the written code to achieve the corresponding data processing. However, developers need to write different code for different data processing procedures. This approach of using fixed code for data processing lacks flexibility and requires repeated code modifications or rewriting when data processing requirements change, thus reducing development efficiency and increasing development costs. Summary of the Invention

[0003] This invention provides a data processing method, apparatus, device, medium, and program product to improve the flexibility of data processing, and to improve development efficiency and reduce development costs by eliminating the need for developers to write code.

[0004] In a first aspect, embodiments of the present invention provide a data processing method, including:

[0005] In response to a data source access operation, obtain at least one data source to be processed;

[0006] In response to an operator selection operation triggered based on data processing requirements in the operator repository, the selected target operator is displayed in the editing interface. Each operator in the operator repository is a pre-encapsulated data processing operation, and different operators correspond to different data processing operations.

[0007] In response to the orchestration operation triggered on the target operator in the editing interface, a data processing stream that meets the data processing requirements is generated based on the target operator;

[0008] The data generated by the at least one data source is processed by executing the data processing stream.

[0009] Secondly, embodiments of the present invention also provide a data processing apparatus, comprising:

[0010] The data source acquisition module is used to acquire at least one data source to be processed in response to a data source access operation;

[0011] The target operator display module is used to respond to the operator selection operation triggered by data processing needs in the operator repository and display the selected target operator in the editing interface. Each operator in the operator repository is a pre-encapsulated data processing operation, and different operators correspond to different data processing operations.

[0012] A data processing stream generation module is used to generate a data processing stream that meets the data processing requirements based on the target operator in response to the arrangement operation triggered by the target operator in the editing interface.

[0013] A data processing module is used to process data generated by the at least one data source by executing the data processing stream.

[0014] Thirdly, embodiments of the present invention also provide an electronic device, the electronic device comprising:

[0015] At least one processor; and

[0016] A memory communicatively connected to the at least one processor; wherein,

[0017] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the data processing method provided in any embodiment of the present invention.

[0018] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing computer instructions, which are used to cause a processor to execute and implement the data processing method provided in any embodiment of the present invention.

[0019] Fifthly, embodiments of the present invention also provide a computer program product, including a computer program that, when executed by a processor, implements the data processing method provided in any embodiment of the present invention.

[0020] One embodiment of the above invention has the following advantages or beneficial effects:

[0021] By responding to data source access operations, at least one data source to be processed can be obtained, allowing dynamic access to different data sources to meet diverse data access needs. By responding to operator selection operations triggered in the operator repository based on data processing requirements, the selected target operator is displayed in the editing interface. Furthermore, by responding to orchestration operations triggered on the target operator in the editing interface, a data processing stream that meets the data processing requirements is generated based on the target operator. Thus, by simply selecting and orchestrating operators in the operator repository based on data processing requirements, a data processing stream that meets those requirements can be quickly generated. By executing the data processing stream, the data generated by the accessed data source is processed. This eliminates the need for developers to write processing code, and when data processing requirements change, only the operators in the data processing stream need to be adjusted, without modifying the code. This improves the flexibility of data processing, increases development efficiency, and reduces development costs.

[0022] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart of a data processing method provided in one embodiment of the present invention;

[0025] Figure 2 This is an example diagram of a data processing flow according to an embodiment of the present invention;

[0026] Figure 3 This is a flowchart of another data processing method provided in one embodiment of the present invention;

[0027] Figure 4 This is a schematic diagram of the structure of a data processing device provided in one embodiment of the present invention;

[0028] Figure 5 This is a schematic diagram of the structure of an electronic device that implements the data processing method of the present invention. Detailed Implementation

[0029] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0030] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0031] Figure 1 This is a flowchart illustrating a data processing method according to an embodiment of the present invention. This embodiment is applicable to processing data generated in business scenarios. The method can be executed by a data processing device, which can be implemented in software and / or hardware, and can be configured in an electronic device. Figure 1 As shown, the method specifically includes the following steps:

[0032] S110. In response to a data source access operation, obtain at least one data source to be processed.

[0033] The data source access operation can be one or more data sources that the user enters in the data source access interface, which need to be accessed and processed, so as to connect to various structured or unstructured data storage systems, thereby simultaneously obtaining data from all data sources that need to be processed, realizing dynamic and unified access to data sources.

[0034] In some optional implementations, at least one data source in step S110 includes, but is not limited to, at least one of the following: text data source, image data source, video data source, and audio / video source. Specifically, a text data source is a business data source used to generate text data, such as a business data source for generating text documents. An image data source is a business data source used to generate image data. A video data source is a business data source used to generate video data. An audio / video source is a business data source used to generate audio data. Different types of data sources generate data of different modalities. All accessed data sources can be data sources that generate data of the same modality or data sources that generate data of different modalities, in order to perform multimodal data fusion processing.

[0035] S120. In response to an operator selection operation triggered based on data processing requirements in the operator repository, the selected target operator is displayed in the editing interface. Each operator in the operator repository is a pre-encapsulated data processing operation, and different operators correspond to different data processing operations.

[0036] The operator repository comprises multiple operators, each a pre-contained component of code used to perform data processing operations. Each operator is an independently running instance. Each operator can correspond to a microservice interface, allowing for operator scheduling and execution via these interfaces, thus enabling microservice operator scheduling. Each operator can be viewed as a single-function data processing module, representing a specific step in the data processing workflow. Different operators are used to implement different data processing operations.

[0037] In some optional implementations, the operator repository includes pre-built system operators and custom operators created by the user. Custom operators are created based on the input operator name, implementation code, and functional description. System operators refer to pre-built system functional modules. In addition to pre-built system operators, the operator repository also includes new custom operators created by users based on business needs. When adding an operator, users can enter the operator name, implementation code, and functional description in the operator management interface. The system containerizes the input operator implementation code and generates a callable microservice interface, thus forming an independently runnable custom operator. After registration, the operator is added to the operator repository for subsequent use. By adding custom operators, various data processing needs can be met, improving the flexibility and scalability of data processing.

[0038] In some optional implementations, the operator repository may include, but is not limited to: text feature extraction operators, image feature extraction operators, video feature extraction operators, audio feature extraction operators, multimodal feature fusion operators, cross-modal semantic alignment operators, multiple different content recognition operators, multiple different data augmentation operators, multiple different data chunking operators, multiple different format conversion operators, multiple different data management operators, and multiple different target storage operators.

[0039] The system includes several functional modules: text feature extraction operators, image feature extraction operators, and audio feature extraction operators. Text feature extraction operators are used to extract text features; for example, they are used based on the BERT (Bidirectional Encoder Representations from Transformers) model. Image feature extraction operators are used to extract image features; for example, they are used based on a convolutional neural network model. Video feature extraction operators are used to extract video features; for example, they are used based on a time-spectral convolutional network model. Different modalities are processed using matching models. Multimodal feature fusion operators are functional modules that fuse different modalities of data. For example, multimodal feature fusion operators are based on multi-head self-attention mechanisms to fuse data from multiple modalities. For instance, features extracted from each modality are linearly projected to generate query, key, and value matrices. These matrices are then divided into multiple heads, each independently calculating an attention score (through dot product or additive methods) to obtain a weighted feature representation, thus achieving multi-head attention computation. In the cross-attention stage, the query from one modality is similar to the key from another modality, focusing on the most relevant part of the other modality, achieving information interaction and alignment between modalities. Finally, the outputs of all heads are concatenated and fused through a linear layer to obtain the fused multimodal feature representation. This effectively fuses multi-source information by capturing feature interactions between different subspaces in parallel through multi-head self-attention and explicitly modeling intermodal relationships through cross-attention. The cross-modal semantic alignment operator calculates the cross-modal contrastive loss to ensure that semantically aligned samples are close together, thereby achieving semantic alignment between different modalities. The cross-modal contrastive loss is determined based on the cosine similarity between features of different modalities.

[0040] Content recognition operators are functional modules used to identify data content. For example, they may include plain text content recognition operators, high-precision text-image-table hybrid recognition operators, visual language model deep recognition operators, and adaptive recognition operators based on file type. Content recognition operators can also include multimodal content layout analysis operators, content parsing operators, standardized format conversion operators, semantic structure extraction operators, and object recognition operators. Data augmentation operators are functional modules that enhance data to improve its quality. For example, they may include text spelling correction operators, summary generation operators, named entity recognition operators, image denoising operators, contrast enhancement operators, image description generation operators, table completion operators, and semantic annotation operators to enhance data consistency and expressiveness. Data chunking operators are functional modules used to chunk data content to improve data processing efficiency and maintain semantic coherence. For example, data chunking operators can include chunking operators by title, chunking operators by character count, chunking operators by page, and chunking operators by similarity. Data chunking operators divide large files or long content into logical chunks to reduce the amount of data within a single chunk, effectively improving the efficiency and accuracy of subsequent knowledge retrieval. Format conversion operators are functional modules used to convert data of different modalities and structures into a searchable and computable standard format based on different downstream target storage systems. For example, format conversion operators can include operators that convert data to vector format, data to JSON format, data to database table format, data to graph format, and data to other custom formats, thus adapting to heterogeneous storage systems such as vector libraries, files, relational databases, and graph databases, and meeting different storage and retrieval needs. Data control operators are used for the security review and management of data content. For example, data control operators include data anonymization operators, data encryption operators, and data access control operators, thereby supporting various review functions such as keyword filtering, sensitive content identification, and malicious logic identification. This ensures the legality and compliance of output data before model training and knowledge construction, improving the quality of content output. Target storage operators are used to periodically synchronize processing results to the corresponding target storage system. For example, target storage operators include operators that store processing results in vector library systems, object storage systems, relational database systems, and graph database systems.

[0041] In this embodiment, the user can select the required operators from the operator repository based on data processing needs, such as by clicking on the required operators to trigger the selection. When the user-triggered operator selection is detected, all the target operators selected by the user are graphically displayed in the editing interface, so that all the displayed target operators can be arranged and processed in the editing interface.

[0042] It should be noted that the target operators displayed in the editing interface can include the same operator. For example, selecting the same operator twice will display the same two target operators in the editing interface, so that certain data processing operations can be repeated to meet various arrangement needs.

[0043] S130, In response to the orchestration operation triggered on the target operator in the editing interface, generate a data processing stream that meets the data processing requirements based on the target operator.

[0044] The orchestration operation involves dragging and connecting target operators displayed in the editing interface to arrange the data flow and execution order of these operators. The data processing flow is a data processing sequence formed by piecing together all displayed target operators in the orchestrated order. The execution order of target operators in the data processing flow can be sequential, parallel, cyclical, or with multiple branches, flexibly determined based on data processing requirements.

[0045] In some optional implementations, step S130 includes: in response to a drag operation triggered on the target operator in the editing interface, synchronously adjusting the placement position of the target operator in the editing interface; in response to a connection operation triggered on the target operator in the editing interface, performing connection processing on the target operator to generate a data processing stream that meets the data processing requirements.

[0046] Specifically, users can freely drag and drop target operators in the editing interface based on their data processing needs to change their placement. Users can then connect the placed target operators to create a complete data processing flow. This drag-and-drop and connection operation allows for flexible combination and arrangement of operators. For example, combining data augmentation and data partitioning operators can improve data quality and retrieval efficiency, while combining format conversion and target storage operators can achieve standardized output and multi-database synchronization. Dynamic arrangement of operators allows for flexible adjustment of the data processing flow according to needs, improving the flexibility and scalability of data processing and facilitating refined management and optimization of each data processing stage.

[0047] For example, Figure 2 An example diagram of a data processing flow is given, such as Figure 2 As shown, when the accessed data source is a data source with three modalities, such as a text data source, an image data source, and a video data source, the generated data processing stream is as follows: Figure 2 As shown, the data is first anonymized using a data anonymization operator from three different data sources. Then, text feature extraction, image feature extraction, and video feature extraction operators are used to extract features from the anonymized data in the three modalities. Finally, a multimodal feature fusion operator is used to fuse the extracted features between different modalities. The fused feature information is then input into an object recognition operator to identify a specified object. Finally, a target storage operator is used to store the recognition results in a specified system, thus completing the entire data processing process.

[0048] It's worth noting that users can define the execution order, branching conditions, and loop logic of data processing flows by dragging and dropping graphical operator components, enabling flexible construction and rapid iteration of multimodal data pipelines. After generating the data processing flow, it can be stored as a DSL (Domain Specific Language) script, supporting version management and historical rollback for convenient continuous maintenance. Each data processing flow script is stored as a versioned object, supporting version management and historical rollback, allowing users to view and restore previous versions at any time. This operator orchestration method effectively lowers the development threshold of data processing flows, improves business response speed, and supports rapid adaptation to constantly changing business scenarios. Dynamic parameter passing and global variable management are supported within the data processing flow, ensuring high process flexibility. Global variables are set for data sharing and transfer across operators and processes; global variables can be modified or updated during the execution of any operator and can be invoked in other operators.

[0049] S140. Process data generated by at least one data source by executing a data processing stream.

[0050] In this embodiment, by executing the data processing stream—that is, executing the code corresponding to all operators in the data processing stream according to the execution order of the operators—the data generated by the access data source is processed accordingly, thereby completing the corresponding data processing process. For example, the scheduling engine parses the operator execution flow in the data processing stream and realizes asynchronous distribution and parallel execution of operators through message queues, thereby supporting parallel and pipelined scheduling and ensuring flexible, efficient, and reliable operator execution. This embodiment can monitor the resources of operators in the data processing stream, realize dynamic load balancing and elastic scaling, and ensure stable operation of operators. The scheduling engine can simulate the execution of the data processing stream in real time, view data flow and operator output, and quickly locate potential anomalies.

[0051] The technical solution of this embodiment obtains at least one data source to be processed in response to a data source access operation, thereby dynamically accessing different data sources to meet different data access needs. In response to an operator selection operation triggered based on data processing needs in the operator repository, the selected target operator is displayed in the editing interface. In response to an orchestration operation triggered on the target operator in the editing interface, a data processing stream that meets the data processing needs is generated based on the target operator. Thus, only the operators in the operator repository need to be selected and orchestrated based on the data processing needs to quickly generate a data processing stream that meets the data processing needs. By executing the data processing stream, the data generated by the accessed data source is processed. This does not require developers to write processing code, and when the data processing needs change, only the operators in the data processing stream need to be adjusted without modifying the code. This improves the flexibility of data processing, increases development efficiency, and reduces development costs.

[0052] Figure 3 This is a flowchart illustrating another data processing method according to an embodiment of the present invention. Based on the above embodiments, this embodiment describes in detail the process of determining and adjusting the target credibility corresponding to the data processing stream before executing the data processing stream. Explanations of terms that are the same as or corresponding to those in the above embodiments are not repeated here.

[0053] See Figure 3 Another data processing method provided in this embodiment specifically includes the following steps:

[0054] S310. In response to a data source access operation, obtain at least one data source to be processed.

[0055] S320, In response to an operator selection operation triggered by data processing requirements in the operator repository, the selected target operator is displayed in the editing interface.

[0056] S330, In response to the orchestration operation triggered on the target operator in the editing interface, generate a data processing stream that meets the data processing requirements based on the target operator.

[0057] S340. The credibility of the data processing flow is evaluated from the dimensions of accuracy, completeness, consistency within the same modality, timeliness, cross-modal consistency, and security compliance, so as to obtain the target credibility of the data processing flow.

[0058] In this embodiment, the credibility evaluation metrics include the accuracy, completeness, intramodal consistency, timeliness, cross-modal consistency, and security compliance of data processing. This allows for a multi-dimensional evaluation of credibility, resulting in a more accurate target credibility score. Specifically, data processing accuracy characterizes the accuracy of model-based data processing within the data processing flow. Completeness characterizes the integrity of the processed data, such as whether any fields are missing. Intramodal consistency characterizes whether there are inconsistencies in the format, semantics, and logical relationships of data within the same modality during the processing flow. Timeliness characterizes the acceptable time consumed in data processing. Cross-modal consistency characterizes the degree of similarity between data from different modalities. Security compliance characterizes the compliance and security of sensitive information processing during the data processing phase.

[0059] In some optional implementations, step S340 includes: determining the accuracy of the data processing stream based on the confidence level of the model output involved in the data processing stream; determining the completeness of the data processing stream based on the data coverage in the data processing stream; determining the same-modality consistency of the data processing stream based on whether there are conflicts within the same-modality data in the data processing stream; determining the timeliness of the data processing stream based on the data generation time and data processing completion time in the data processing stream; determining the semantic similarity between different modal data based on the feature information corresponding to each modal data in the data processing stream, and determining the cross-modal consistency of the data processing stream based on the semantic similarity; determining the security compliance of the data processing stream based on the data false negative rate and de-identification compliance in the data processing stream; and obtaining the target credibility of the data processing stream by weighted fusion of accuracy, completeness, same-modality consistency, timeliness, cross-modal consistency, and security compliance.

[0060] Specifically, the confidence scores of each model output involved in the data processing stream are obtained, such as text recognition confidence score, speech recognition accuracy, and entity recognition confidence score. All model output confidence scores are summed, and the sum is taken as the accuracy of the data processing stream. The number of fields in the data processing stream with no missing values ​​is determined, and this number is divided by the total number of fields to obtain data coverage. Data coverage is taken as the completeness of the data processing stream. Based on whether there are conflicts within the same modality data in the data processing stream, the number of fields without conflicts is obtained, and this number is divided by the total number of fields. The result is taken as the same modality consistency of the data processing stream. The data processing completion time is subtracted from the data generation time in the data processing stream, and the result is taken as the timeliness of the data processing stream. Based on the feature information corresponding to each modality data in the data processing stream, the cosine similarity between every two modality features is determined. This cosine similarity is taken as semantic similarity, and all semantic similarities are averaged to obtain the average semantic similarity, which is taken as the cross-modal consistency of the data processing stream. The data false negative rate is obtained by dividing the number of missed detections in the data processing stream by the total number of sensitive items. The sensitive data detection rate is obtained by subtracting the false negative rate from 1. This rate checks whether the data processing process correctly de-identifies data according to rules, such as whether named entities are replaced with "XXX" or hash values. The de-identification rate is obtained by dividing the replaced items by the total number of sensitive items. The sensitive data detection rate and the sensitive data de-identification rate are weighted and summed; the sum is used as the security compliance of the data processing stream. The target credibility of the data processing stream is obtained by weighting and summing the accuracy, completeness, intramodal consistency, timeliness, cross-modal consistency, and security compliance. The weights for accuracy, completeness, intramodal consistency, timeliness, cross-modal consistency, and security compliance are summed to 1, which can be obtained through machine learning. The final target credibility value is [0, 1].

[0061] S350. If the target confidence level is less than the preset confidence level, the target operator in the data processing stream is adjusted until the target confidence level of the adjusted data processing stream is greater than or equal to the preset confidence level.

[0062] The preset credibility level is a pre-set minimum credibility level that allows the execution of a data processing stream. Specifically, if the target credibility level of a data processing stream is less than the preset credibility level, it indicates that the processing result of that data processing stream has poor credibility. In this case, the target operator in the data processing stream needs to be adjusted, such as replacing it with another operator, adding a new operator (e.g., a data augmentation operator), or modifying the implementation code of the target operator. The credibility of the adjusted data processing stream is then re-evaluated until the redefined target credibility level is greater than or equal to the preset credibility level.

[0063] S360. Process data generated by at least one data source by executing a data processing stream.

[0064] Specifically, if the target credibility of the data processing stream is greater than or equal to the preset credibility, the data processing stream is allowed to be executed, thereby ensuring the accuracy and reliability of the data processing results.

[0065] The technical solution of this embodiment evaluates the credibility of the data processing stream from the dimensions of accuracy, completeness, consistency within the same modality, timeliness, cross-modal consistency, and security compliance. It obtains the target credibility of the data processing stream. If the target credibility is less than the preset credibility, the target operator in the data processing stream is adjusted until the target credibility of the adjusted data processing stream is greater than or equal to the preset credibility. Then, the data processing stream is allowed to be executed, thereby ensuring the accuracy and reliability of data quality and data processing results.

[0066] In some optional implementations, a modal data lifecycle governance mechanism manages the entire lifecycle of multimodal data, from collection, processing, storage, archiving to destruction, ensuring data security, compliance, and high quality. This mechanism supports automatic scanning of specified data source paths, interfaces, or databases via scheduled or event-triggered methods. Combining semantic embedding and similarity comparison technologies, it accurately identifies newly added or changed data information, avoiding redundant access. Simultaneously, newly identified data is automatically integrated into the processing flow, undergoing modality recognition, semantic enhancement, sensitive information review, and standardization transformation before being written to the target storage system. This mechanism not only enables dynamic data access and continuous updates but also improves data freshness, system response efficiency, and knowledge evolution capabilities in lifecycle management. It is particularly suitable for building knowledge bases, question-answering systems, and multimodal integration platforms with automatic update capabilities, enhancing the system's perception and adaptation to changing information while ensuring data quality and security, forming a closed-loop governance path for the entire data lifecycle. The modal data lifecycle governance mechanism manages data security and compliance throughout its entire lifecycle, ensuring that data meets relevant regulatory requirements during processing, thereby achieving efficient, intelligent, and scalable multimodal data processing and significantly improving the value and management capabilities of data assets.

[0067] The following are embodiments of the data processing apparatus provided in this invention. This apparatus and the data processing methods in the above embodiments belong to the same inventive concept. For details not described in detail in the embodiments of the data processing apparatus, please refer to the embodiments of the above data processing methods.

[0068] Figure 4This is a schematic diagram of a data processing device provided in an embodiment of the present invention. This embodiment is applicable to processing data generated in any business scenario. Figure 4 As shown, the device specifically includes: a data source acquisition module 410, a target operator display module 420, a data processing stream generation module 430, and a data processing module 440.

[0069] The system includes: a data source acquisition module 410, used to acquire at least one data source to be processed in response to a data source access operation; a target operator display module 420, used to display the selected target operator in the editing interface in response to an operator selection operation triggered based on data processing requirements in the operator repository, wherein each operator in the operator repository is a pre-encapsulated data processing operation, and different operators correspond to different data processing operations; a data processing stream generation module 430, used to generate a data processing stream that meets the data processing requirements based on the target operator in response to an orchestration operation triggered on the editing interface; and a data processing module 440, used to process the data generated by the at least one data source by executing the data processing stream.

[0070] The technical solution of this embodiment obtains at least one data source to be processed in response to a data source access operation, thereby dynamically accessing different data sources to meet different data access needs. In response to an operator selection operation triggered based on data processing needs in the operator repository, the selected target operator is displayed in the editing interface. In response to an orchestration operation triggered on the target operator in the editing interface, a data processing stream that meets the data processing needs is generated based on the target operator. Thus, only the operators in the operator repository need to be selected and orchestrated based on the data processing needs to quickly generate a data processing stream that meets the data processing needs. By executing the data processing stream, the data generated by the accessed data source is processed. This does not require developers to write processing code, and when the data processing needs change, only the operators in the data processing stream need to be adjusted without modifying the code. This improves the flexibility of data processing, increases development efficiency, and reduces development costs.

[0071] Optionally, the operator repository includes preset system operators and created custom operators, wherein the custom operators are created based on the input operator name, operator implementation code, and operator function description information.

[0072] Optionally, the at least one data source includes at least one of the following: a text data source, an image data source, a video data source, and an audio / video source;

[0073] The operator repository includes: text feature extraction operators, image feature extraction operators, video feature extraction operators, audio feature extraction operators, multimodal feature fusion operators, cross-modal semantic alignment operators, multiple different content recognition operators, multiple different data augmentation operators, multiple different data segmentation operators, multiple different format conversion operators, multiple different data management operators, and multiple different target storage operators.

[0074] Optionally, the data processing stream generation module 430 is specifically used for:

[0075] In response to a drag operation triggered on the target operator in the editing interface, the placement position of the target operator in the editing interface is adjusted synchronously.

[0076] In response to a connection operation triggered on the target operator in the editing interface, the target operator is connected to generate a data processing stream that meets the data processing requirements.

[0077] Optionally, the device further includes:

[0078] The credibility assessment module is used to assess the credibility of the data processing stream from the dimensions of accuracy, completeness, consistency within the same modality, timeliness, cross-modal consistency, and security compliance, and to obtain the target credibility of the data processing stream.

[0079] The data processing stream adjustment module is used to adjust the target operator in the data processing stream if the target confidence level is less than the preset confidence level, until the target confidence level of the adjusted data processing stream is greater than or equal to the preset confidence level.

[0080] Optionally, the credibility assessment module is specifically used for:

[0081] The accuracy of the data processing stream is determined based on the confidence level of the model output involved in the data processing stream.

[0082] The completeness of the data processing stream is determined based on the data coverage in the data processing stream.

[0083] The consistency of the same modality corresponding to the data processing stream is determined based on whether there are conflicts within the same modal data in the data processing stream;

[0084] The timeliness of the data processing stream is determined based on the data generation time and data processing completion time in the data processing stream.

[0085] Based on the feature information corresponding to each modality of data in the data processing stream, the semantic similarity between different modalities is determined, and based on the semantic similarity, the cross-modal consistency corresponding to the data processing stream is determined;

[0086] Based on the data miss rate and anonymization compliance in the data processing stream, the security compliance corresponding to the data processing stream is determined;

[0087] The target credibility of the data processing stream is obtained by weighted fusion of the accuracy, completeness, intramodal consistency, timeliness, cross-modal consistency, and security compliance.

[0088] The data processing apparatus provided in the embodiments of the present invention can execute the data processing method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the data processing method.

[0089] It is worth noting that in the embodiments of the above data processing device, the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of the present invention.

[0090] Figure 5 A schematic diagram of an electronic device 10, which can be used to implement embodiments of the present invention, is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0091] like Figure 5As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0092] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0093] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various data processing methods and processes described above.

[0094] In some embodiments, the data processing method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or mounted on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the data processing method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the data processing method by any other suitable means (e.g., by means of firmware).

[0095] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0096] Computer programs for implementing the data processing methods of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The computer programs can be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0097] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0098] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0099] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0100] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0101] This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the data processing method provided in any embodiment of this invention.

[0102] In the implementation of a computer program product, computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​as well as conventional procedural programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0103] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0104] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A data processing method, characterized in that, include: In response to a data source access operation, obtain at least one data source to be processed; In response to an operator selection operation triggered by data processing requirements in the operator repository, the selected target operator is displayed in the editing interface. Each operator in the operator repository is a pre-encapsulated data processing operation, and different operators correspond to different data processing operations. In response to the orchestration operation triggered on the target operator in the editing interface, a data processing stream that meets the data processing requirements is generated based on the target operator; The data generated by the at least one data source is processed by executing the data processing stream.

2. The method according to claim 1, characterized in that, The operator repository includes pre-set system operators and created custom operators, wherein the custom operators are created based on the input operator name, operator implementation code, and operator function description information.

3. The method according to claim 1, characterized in that, The at least one data source includes at least one of the following: text data source, image data source, video data source, and audio / video source; The operator repository includes: text feature extraction operators, image feature extraction operators, video feature extraction operators, audio feature extraction operators, multimodal feature fusion operators, cross-modal semantic alignment operators, multiple different content recognition operators, multiple different data augmentation operators, multiple different data segmentation operators, multiple different format conversion operators, multiple different data management operators, and multiple different target storage operators.

4. The method according to claim 1, characterized in that, The step of responding to an orchestration operation triggered on the target operator in the editing interface, and generating a data processing stream that meets the data processing requirements based on the target operator, includes: In response to a drag operation triggered on the target operator in the editing interface, the placement position of the target operator in the editing interface is adjusted synchronously. In response to a connection operation triggered on the target operator in the editing interface, the target operator is connected to generate a data processing stream that meets the data processing requirements.

5. The method according to any one of claims 1-4, characterized in that, When the at least one data source includes at least two of the following: a text data source, an image data source, a video data source, and an audio / video source, the data processing stream is further included before execution: The credibility of the data processing stream is evaluated from the dimensions of accuracy, completeness, consistency within the same modality, timeliness, cross-modal consistency, and security compliance, to obtain the target credibility of the data processing stream. If the target confidence level is less than the preset confidence level, the target operator in the data processing stream is adjusted until the target confidence level of the adjusted data processing stream is greater than or equal to the preset confidence level.

6. The method according to claim 5, characterized in that, The credibility assessment of the data processing stream is performed from the dimensions of accuracy, completeness, consistency within the same modality, timeliness, cross-modal consistency, and security compliance to obtain the target credibility of the data processing stream, including: The accuracy of the data processing stream is determined based on the confidence level of the model output involved in the data processing stream. The completeness of the data processing stream is determined based on the data coverage in the data processing stream. The consistency of the same modality corresponding to the data processing stream is determined based on whether there are conflicts within the same modal data in the data processing stream; The timeliness of the data processing stream is determined based on the data generation time and data processing completion time in the data processing stream. Based on the feature information corresponding to each modality of data in the data processing stream, the semantic similarity between different modalities of data is determined, and based on the semantic similarity, the cross-modal consistency corresponding to the data processing stream is determined; Based on the data miss rate and anonymization compliance in the data processing stream, the security compliance corresponding to the data processing stream is determined; The target credibility of the data processing stream is obtained by weighted fusion of the accuracy, completeness, intramodal consistency, timeliness, cross-modal consistency, and security compliance.

7. A data processing apparatus, characterized in that, include: The data source acquisition module is used to acquire at least one data source to be processed in response to a data source access operation; The target operator display module is used to respond to the operator selection operation triggered by data processing needs in the operator repository and display the selected target operator in the editing interface. Each operator in the operator repository is a pre-encapsulated data processing operation, and different operators correspond to different data processing operations. A data processing stream generation module is used to generate a data processing stream that meets the data processing requirements based on the target operator in response to the arrangement operation triggered by the target operator in the editing interface. A data processing module is used to process data generated by the at least one data source by executing the data processing stream.

8. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the data processing method according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the data processing method according to any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the data processing method as described in any one of claims 1-6.

Images