A file management method and system

Abstract

The present invention relates to the field of archive management technology, and in particular to an archive management method and system. The method comprises the following steps: performing archive text theme level analysis between different theme contents on text theme content data to obtain archive text theme level data; performing memory pointer structure design according to the archive text theme level data to obtain associated extended memory pointer structure data; performing heap space adaptive allocation control according to the associated extended memory pointer structure data to obtain heap space adaptive allocation control data; performing virtual memory temporary allocation scheduling based on the heap space adaptive allocation control data and performing archive scheduling memory strategy formulation to obtain archive scheduling memory strategy, and sending the archive scheduling memory strategy to the server bottom system to perform archive management. The present invention makes the archive management technology more perfect by optimizing the archive management technology.

Description

Technical Field

[0001] The present invention relates to the technical field of archive management, and in particular to an archive management method and system. Background Art

[0002] In modern society, archives have shifted from traditional paper forms to digital, electronic and intelligent development. This transformation is not only aimed at improving management efficiency, but also attempts to solve the problems of inconvenient information acquisition, limited storage space and physical damage in traditional archive management. The advancement of computer hardware provides the basis for the digitization of archives. The processing power, storage capacity and continuous optimization of computer operating systems make the creation and management of digital archives possible. The emergence of database management systems, especially the popularization of relational databases, has greatly improved the efficiency of archival data storage and management. Archival information no longer relies on traditional paper documents, but is stored in digital form on hard disks, tapes, optical disks and other media. Using DBMS, archives can be efficiently classified, queried and retrieved, realizing the structured storage and management of archival data. However, a traditional archive management method has the problems of low accuracy of archive query and low retrieval efficiency. Summary of the invention

[0003] Based on this, it is necessary to provide a file management method and system to solve at least one of the above technical problems.

[0004] To achieve the above object, a file management method is provided, the method comprising the following steps:

[0005] Step S1: Acquire archive text data; perform archive text theme level analysis between different theme contents on the text theme content data to obtain archive text theme level data;

[0006] Step S2: performing context perception processing between different subject content levels on the archive text data according to the archive text subject level data to obtain subject content context level perception data; performing associated extension word retrieval on the subject content context level perception data to obtain semantic associated extension word data; performing memory pointer structure design based on the semantic associated extension word data to obtain associated extension memory pointer structure data;

[0007] Step S3: performing heap space adaptive allocation control according to the associated extended memory pointer structure data to obtain heap space adaptive allocation control data; performing virtual memory temporary allocation scheduling based on the heap space adaptive allocation control data to obtain virtual memory temporary allocation scheduling data;

[0008] Step S4: Formulate an archive scheduling memory strategy based on the associated extended memory pointer structure data and the virtual memory temporary allocation scheduling data to obtain the archive scheduling memory strategy, and send the archive scheduling memory strategy to the underlying system of the server to execute archive management.

[0009] The present invention can extract the subject structure of the archive text by performing hierarchical analysis of different subject contents on the archive text data, thereby achieving in-depth understanding and classification of the text. Through this analysis, key information and themes in the archive can be identified, the logical relationship between different levels can be clarified, and a basis can be provided for subsequent data processing and storage. This analysis can not only help to effectively classify the archive, but also provide an organized data structure for processing complex information. After obtaining the subject hierarchical data of the archive text, the contextual relationship between different subject contents is deeply analyzed through context perception processing. This process can help the system understand the subtle connection between the subject contents, thereby laying the foundation for further vocabulary expansion and semantic retrieval. Based on these context perception data, the system will retrieve the associated extension words and obtain the extension word data related to the subject, which helps to improve the depth of understanding of the subject and provide more diverse information support for the storage and retrieval process. At the same time, by constructing a memory pointer structure, the efficiency and flexibility of the system in archive processing can be enhanced. According to the memory pointer structure data obtained in step S2, the system will perform adaptive allocation regulation of the heap space. This process ensures that the system can flexibly adjust the memory allocation strategy according to different subject levels and context levels, thereby maximizing storage efficiency and avoiding memory waste or insufficient problems. The adaptive regulation of heap space makes the access process of archive data more efficient, and can dynamically adjust memory resources under different data volumes and processing requirements to ensure stable operation of the system. By combining the heap space regulation data in step S3, the system performs temporary allocation scheduling of virtual memory. This process not only ensures that the system can effectively allocate resources under high load, but also optimizes the time efficiency of data storage and reading. On this basis, an archive scheduling memory strategy is formulated to improve the system's response speed and processing capabilities for archive management by adjusting the memory resource allocation and scheduling strategy. Finally, the generated archive scheduling memory strategy is sent to the underlying server system for execution, ensuring that the storage, retrieval and management of archives are efficient and stable.

[0010] Preferably, step S1 comprises the following steps:

[0011] Step S11: Obtaining archive text data;

[0012] Step S12: Fill missing values ​​in the archive text data to obtain archive text filling data;

[0013] Step S13: classifying the archive text filling data by text subject content to obtain text subject content data;

[0014] Step S14: performing archive text theme level analysis between different theme contents on the text theme content data according to the text theme content data to obtain archive text theme level data.

[0015] The acquisition of archival text data in the present invention is the starting point of the entire process, which provides raw materials for subsequent processing and analysis. By collecting relevant archival data from multiple channels or systems, the diversity and comprehensiveness of the data are ensured. This step not only ensures the reliability and integrity of the data source, but also lays the foundation for subsequent text analysis, classification and hierarchical division. In archive management, obtaining accurate original data is a key step to ensure system performance. There are some missing values ​​in the archival text data, which affect the subsequent data processing and analysis effects. By filling these missing values, the consistency and integrity of the data can be maintained. Missing value filling can not only prevent analysis deviations caused by missing data, but also reasonably supplement the missing parts through intelligent methods (such as interpolation, model prediction, etc.), so that subsequent analysis can be more accurate and efficient. This step ensures the integrity of the data during the processing process and reduces errors caused by incomplete data. After the missing values ​​are filled in the archival text, the next step is to classify the text data by subject content. This process uses natural language processing (NLP) technology to classify the text by subject based on the similarity and relevance of the text content, making the archival data more orderly and structured. Text subject classification helps to distinguish different types of archival data by subject, making information retrieval, analysis and management more efficient. The classified data provides a clear framework and foundation for subsequent hierarchical analysis and context perception. After subject content classification, the relationship and hierarchical structure between different subjects can be revealed by performing hierarchical analysis on text subject data. This analysis process helps the system identify the main themes, sub-themes and logical connections between them in the text, making the organization of archival data more reasonable and facilitating subsequent semantic processing and storage optimization. Thematic hierarchical analysis can improve the system's understanding of complex data, ensure that the archive management system can perform intelligent data storage, indexing and scheduling according to different subject levels, and provide strong support for subsequent context perception processing and memory allocation.

[0016] Preferably, step S2 comprises the following steps:

[0017] Step S21: performing context perception processing between different subject content levels on the archive text data according to the archive text subject level data to obtain subject content context level perception data;

[0018] Step S22: performing semantic enhancement segment analysis on the subject content context level perception data to obtain content level semantic key segments;

[0019] Step S23: performing a search for associated expansion words on the subject content context level perception data according to the content level semantic key fragments to obtain semantic associated expansion word data;

[0020] Step S24: Design a memory pointer structure based on the semantically associated extended word data to obtain associated extended memory pointer structure data.

[0021] According to the subject level data of the archive text, the system performs context perception processing between different subject levels on the text content. The core of this process is to identify the relationship between each level in the text through context analysis, so as to ensure that the system can understand the deep context information behind the subject content. Context perception processing helps to accurately grasp the meaning of the archive text, avoid misunderstanding or omission caused by simple literal understanding, and thus provide more accurate basic data for subsequent data analysis and expansion. After obtaining the subject content context level perception data, the system performs semantic enhancement fragment analysis, and further extracts the key semantic fragments in the content level. This analysis can not only improve the depth of text understanding, but also help to extract the most valuable semantic information, and prepare for subsequent expansion and storage. Semantic enhancement fragment analysis can highlight the important parts in the archive text, so that the system can focus on the core information, avoid the interference of too much irrelevant content, and improve the processing efficiency of data. According to the key fragment data obtained from the semantic enhancement fragment analysis, the system performs semantic association expansion word retrieval. This process is based on the extracted key fragments, and the system automatically searches for expansion words associated with these fragments. Through semantic association expansion, the system can further expand and improve the vocabulary of subject content to ensure that relevant vocabulary can be fully covered in archive management. This step not only enhances the system's semantic understanding ability, but also improves the retrieval accuracy of similar or related information, and further enhances the flexibility of archive storage and retrieval. Based on the semantic association extension word data, the system designs an association extension memory pointer structure. In this process, the system will build an efficient memory pointer structure based on the relationship between the subject content and the extension words, so as to access and store related data more quickly. The optimized design of the memory pointer structure enables the archive system to locate and schedule memory resources more efficiently when processing massive data, thereby improving the speed and stability of data processing. This design helps to enhance the flexibility of memory access, reduce storage bottlenecks, and improve the performance of the overall archive management system.

[0022] Preferably, step S24 comprises the following steps:

[0023] Step S241: performing context relevance evaluation on the subject content context level perception data based on the semantically associated extended word data to obtain extended word-subject content relevance data;

[0024] Step S242: Designing an expansion word chain pointer rule for the semantically associated expansion word data according to the expansion word-subject content correlation data to obtain an expansion word chain pointer design rule;

[0025] Step S243: matching the pointer allocation synchronization mechanism with the extended word chain pointer design rule to obtain the extended word chain pointer design synchronization rule;

[0026] Step S244: performing chain parallel partitioning on the extended word chain pointer design synchronization rule to obtain chain parallel partition pointer design data;

[0027] Step S245: Design a memory pointer structure according to the chained parallel partition pointer design data to obtain associated extended memory pointer structure data.

[0028] The present invention is based on semantically associated extended word data, and the system evaluates the contextual relevance of the context level of the subject content. The core purpose of this evaluation is to analyze the degree of association between the extended word and the subject content to determine which extended words can better reflect the contextual meaning of the text. By accurately evaluating the relevance of the extended word and the subject content, the system can screen out the information with the most semantic relevance to ensure accurate processing and analysis of the text. This process helps the system improve its understanding of the deep meaning of the archive text and provides more reliable data support for subsequent steps. According to the extended word-subject content relevance data, the extended word chain pointer rule is designed. The purpose of this rule design is to effectively structure the relationship between the extended word and the subject content. By constructing a chain pointer, the system can achieve efficient management of the mutual connection between multiple extended words and the subject content. The chain pointer can not only strengthen the association between semantics, but also optimize the efficiency of storage and retrieval. Through this design, the system can achieve flexible matching and dynamic updating of extended words, thereby more accurately reflecting the subject changes in the text. Based on the extended word chain pointer design rule, the pointer allocation synchronization mechanism matching is implemented. This process ensures that in the chain pointer mapping process between the extension word and the subject content, each pointer can be allocated and synchronized in a coordinated manner. Through the precise synchronization mechanism, the system can avoid pointer pointing errors or inconsistent information, thereby ensuring the stability and reliability of the data processing process. The design of the synchronization mechanism is particularly important for multi-threaded and high-concurrency data access, which can ensure the efficiency and accuracy of pointer allocation in a large-scale data processing environment. Chain parallel partitioning is performed on the synchronization rules of the extension word chain pointer design to generate chain parallel partition pointer design data. This process improves the efficiency of pointer access and update by parallel partitioning the related pointers according to certain rules to form multiple parallel processing areas. Chain parallel partitioning helps to divide a large amount of data into smaller modules, so that each module can be processed independently, reducing the processing burden of a single pointer block. Through this partition design, the system can achieve more efficient parallel computing when processing complex data and effectively improve the overall performance of the system. Based on the chain parallel partition pointer design data, the memory pointer structure is designed, and finally the associated extended memory pointer structure data is generated. The core purpose of this step is to optimize the pointer structure in memory based on the previous pointer design rules and partitioning strategies, so that it can store, access and update data more efficiently. Through the refined memory pointer design, the system can reduce memory usage and increase data processing speed when processing large-scale archival texts. The memory pointer structure design ensures the efficiency of information storage and retrieval, and can flexibly respond to various access requirements in a complex data environment.

[0029] Preferably, designing an expansion word chain pointer rule for semantically associated expansion word data according to expansion word-subject content correlation data includes the following steps:

[0030] Performing extended word context position difference analysis on semantically related extended word data to obtain extended word context position difference data;

[0031] Based on the extended word context position difference data, the extended word-topic content relevance data is quantified with respect to the content relevance weight, thereby obtaining the extended word-topic content relevance weight quantification data;

[0032] According to the weight quantification data of the correlation between the extended word and the subject content, a bidirectional linked list pointer structure is constructed for the semantically associated extended word data to obtain an extended word associated linked list pointer structure;

[0033] Based on the extended word associated linked list pointer structure, a pointer reconfigurable linked list is constructed to obtain a pointer reconfigurable linked list;

[0034] According to the pointer structure of the extension word association chain list and the pointer reconfigurable chain list, the extension word chain pointer rule is designed for the semantically associated extension word data, and the extension word chain pointer design rule is obtained.

[0035] The present invention performs context position difference analysis on semantically associated extended word data, with the purpose of identifying the position change of extended words in different text contexts and their influence on the understanding of subject content. By analyzing the position difference of extended words in different contexts, the system can find the semantic deviation and potential context association of extended words, and further reveal the different roles played by extended words in subject content. This analysis can effectively identify which extended words are more relevant or prominent in different contexts, and provide accurate data support for subsequent weight quantification and structural design. Based on the extended word context position difference data, the system performs weight quantification on the relevance of extended words-subject content. The process of weight quantification is to quantitatively evaluate the relevance between each extended word and subject content, and reflect the matching degree of extended words and subject content by assigning different weight values. This step helps the system to more accurately evaluate the importance of each extended word in the subject by calculating the influence of context difference on content. Through this quantitative analysis, the system can give priority to those extended words that are highly relevant to subject content, and improve the accuracy and efficiency of data processing. According to the weight quantification data of the relevance of extended words-subject content, a bidirectional linked list pointer structure is constructed. The bidirectional linked list establishes forward and backward pointers for the connection between each extension word and the related subject content, so that the system can more efficiently perform bidirectional navigation between the extension word and the subject content. The construction of this pointer structure enhances the dynamic access capability to semantic relationships and supports fast and flexible data retrieval and update between different extension words and subject content. The design of the bidirectional linked list structure can effectively improve storage and query efficiency, especially when processing large-scale data. The pointer reconfigurable linked list is constructed based on the pointer structure of the extension word association linked list. The purpose of this process is to enable the linked list structure to have the ability to dynamically reconstruct, and to adjust the connection relationship of the pointer in real time according to semantic requirements or data changes. Through this reconfigurable pointer linked list, the system can dynamically adjust the pointer structure according to different query requirements or the update of subject content, and improve the flexibility of data storage and access. The design of the reconfigurable linked list enables the system to cope with more complex semantic relationships, especially in the processing of changeable and dynamic data environments, which can effectively avoid the performance bottleneck caused by structural solidification. Based on the pointer structure of the extension word association linked list and the pointer reconfigurable linked list, the extension word chain pointer rule is designed for the semantically associated extension word data. This step integrates the previous linked list structure and reconstruction mechanism to develop a complete set of extended word chain pointer rules to manage and access the associated data between extended words and subject content. These pointer rules can ensure efficient storage while achieving accurate positioning and fast access to data. Through the regularized pointer design, the system not only optimizes the data storage structure, but also improves the semantic matching accuracy between subject content and extended words, further enhancing text retrieval and processing capabilities.

[0036] Preferably, the design of synchronization rules for chain pointers of extended words to perform chain parallel partitioning includes the following steps:

[0037] The pointer traversal retrieval time complexity is simulated for the synchronization rule of the extended word chain pointer design, and the pointer traversal retrieval time complexity is obtained;

[0038] According to the pointer traversal retrieval time complexity, the synchronization rule of the extended word chain pointer design is matched with the limit chain length to obtain the limit chain length matching data;

[0039] Based on the extreme chain length matching data and the expansion word-topic content relevance data, the expansion word chain pointer design synchronization rules are used to construct contextual parallel subchains to obtain the expansion word contextual parallel subchains;

[0040] Perform subchain incremental constraint analysis on the parallel subchains in the context of the extended word to obtain subchain incremental constraint data;

[0041] According to the extended word context parallel sub-chains and sub-chain incremental constraint data, the extended word chain pointer design synchronization rule is chained and partitioned to obtain chained parallel partition pointer design data.

[0042] The present invention can accurately evaluate the time consumption required for pointer traversal operation under different input conditions by simulating the pointer traversal retrieval time complexity of the extended word chain pointer design synchronization rule. This process can reveal the bottleneck of pointer traversal efficiency and provide a theoretical basis for subsequent optimization. Specifically, the simulation process will take into account the complexity of different data scales and chain structures, help designers understand the performance limitations of pointer traversal, and thus provide a direction for optimizing pointer operations. This step provides important data support for subsequent chain design and rule adjustment. According to the pointer traversal retrieval time complexity obtained in the previous step, the maximum tolerable length of the chain structure under certain specific conditions can be derived through the extreme chain length matching. This matching process takes into account the balance between chain length and traversal efficiency, ensuring that the chain structure can handle long extended word chains without sacrificing retrieval efficiency. The result of the extreme chain length matching provides a data basis for further rule optimization and chain length control, which is helpful to design a more scalable pointer structure. On the basis of obtaining the extreme chain length matching data and the extended word-subject content correlation data, the relevant extended words in different contexts are connected into parallel subchains through the context parallel subchain construction. This process effectively integrates the subject content and the word chain structure, so that the chain pointer is not limited to vocabulary matching, but can also perform deeper structural optimization based on content relevance. Guided by the context, the constructed parallel subchain can better reflect the semantic relationship between the extended words, thereby improving the expressiveness and adaptability of the pointer chain. After completing the construction of the extended word context parallel subchain, the subchain incremental constraint analysis is performed. This step explores the performance bottlenecks or data inconsistency problems that occur during the chain structure expansion process by analyzing the incremental changes of each subchain. By setting reasonable incremental constraints, it is ensured that the chain pointer can maintain stability and efficiency during expansion, avoiding inefficiency or excessive memory consumption caused by too long chains or unreasonable increments. Incremental constraint analysis provides necessary guarantees for the balance and maintainability of the chain structure. Finally, chain parallel partitioning is performed based on the extended word context parallel subchain and subchain incremental constraint data. This process divides the complex pointer chain into several relatively independent parallel partitions to optimize storage, management and retrieval. Chain parallel partitioning can effectively reduce the coupling between chains, improve the efficiency of pointer operations, and help to better perform parallel computing in distributed or large-scale data environments. Through this partitioning strategy, the overall performance of the system can be significantly improved, ensuring that the extended word chain pointer design can meet the needs of large-scale data applications.

[0043] Preferably, step S3 comprises the following steps:

[0044] Step S31: performing concurrent search efficiency simulation on the associated extended memory pointer structure data to obtain pointer structure concurrent search efficiency data;

[0045] Step S32: performing heap space adaptive allocation control according to the pointer structure concurrent retrieval efficiency data to obtain heap space adaptive allocation control data;

[0046] Step S33: Perform virtual memory temporary allocation scheduling based on the heap space adaptive allocation control data and the pointer structure concurrent retrieval efficiency data to obtain virtual memory temporary allocation scheduling data.

[0047] The present invention simulates the concurrent retrieval efficiency of the associated extended memory pointer structure data, with the purpose of evaluating the retrieval efficiency of the pointer structure in a multi-threaded or multi-tasking environment. By simulating the concurrent retrieval scenario, the performance of the pointer structure under high load conditions, especially the response time and throughput of the retrieval operation, can be accurately predicted. This step can help the system identify potential performance bottlenecks, and then provide a basis for subsequent memory management and optimization, ensuring that the system can efficiently process a large number of concurrent query requests in actual operation. Based on the pointer structure concurrent retrieval efficiency data obtained in step S31, step S32 performs adaptive allocation and regulation of the heap space. By dynamically analyzing the memory requirements of the system in concurrent retrieval, the system can intelligently adjust the allocation strategy of the heap space according to the actual load and the access mode of the pointer structure. The heap space adaptive allocation and regulation can automatically adjust the size and structure of the memory allocation according to the hotspot area and concurrent load of data access, thereby improving memory utilization, reducing unnecessary memory waste, and ensuring that the retrieval operation can maintain high efficiency and stability under different loads. In step S33, based on the heap space adaptive allocation and regulation data and the concurrent retrieval efficiency data, the system further performs temporary allocation scheduling of virtual memory. Temporary allocation and scheduling of virtual memory can flexibly and dynamically schedule resources between physical memory and virtual memory, ensuring that when memory pressure is high, the system can efficiently use virtual memory to supplement the shortage of physical memory. This process ensures that when the system faces high concurrency and large-scale data, it can avoid memory overflow or performance degradation through reasonable memory allocation strategies, thereby improving the overall stability and response speed of the system.

[0048] Preferably, step S32 includes the following steps:

[0049] Step S321: performing efficiency fluctuation range analysis on the pointer structure concurrent retrieval efficiency data to obtain the concurrent retrieval efficiency fluctuation range;

[0050] Step S322: extracting the number of search results from the pointer structure concurrent search efficiency data to obtain the number of search results;

[0051] Step S323: mapping the search result quantity value according to the concurrent search efficiency fluctuation range to obtain efficiency fluctuation-result quantity association mapping data;

[0052] Step S324: dynamically adjust the heap space according to the efficiency fluctuation-result quantity association mapping data to obtain heap space dynamic adjustment data;

[0053] Step S325: performing release switching fragment optimization processing on the heap space dynamic adjustment data to obtain release switching fragment optimization data;

[0054] Step S326: performing heap space adaptive allocation control according to the heap space dynamic adjustment data and the release switching fragment optimization data to obtain heap space adaptive allocation control data.

[0055] The present invention performs efficiency fluctuation interval analysis on the pointer structure concurrent retrieval efficiency data, and the main purpose is to identify the efficiency fluctuation range in the concurrent retrieval operation. By quantitatively analyzing the retrieval efficiency in different time periods, the fluctuation range of the efficiency performance of the pointer structure under different loads or different conditions can be clearly identified. Identifying the efficiency fluctuation interval helps to reveal the potential performance bottleneck of the system under high concurrency, and provides a specific basis for subsequent optimization work, ensuring that the system can always maintain a relatively stable retrieval efficiency under different load conditions. The number of retrieval results, that is, the amount of valid data returned by each retrieval operation, is extracted from the pointer structure concurrent retrieval efficiency data. The number of retrieval results is an important indicator for evaluating retrieval efficiency and performance, because it is directly related to the effectiveness of each retrieval operation and the amount of data returned. By extracting the number of retrieval results, the system can more accurately judge the load of the retrieval operation and provide key data support for the adjustment and optimization of the heap space. This step lays a data foundation for subsequent efficiency optimization and resource scheduling. According to the efficiency fluctuation interval analysis results and the extracted number of retrieval results, the correlation mapping between efficiency fluctuation and the number of retrieval results is performed. By associating the values ​​of the number of search results under different fluctuation intervals, the system can identify how the number of results returned changes under different search efficiency conditions. This step can reveal the intrinsic connection between the fluctuation of search efficiency and the returned results, help the system optimize the heap space and memory allocation strategy, and achieve more accurate resource allocation. Through this step, the growth trend of search results under certain fluctuation conditions can be predicted in advance, thereby avoiding over-allocation or under-allocation of resources. According to the efficiency fluctuation-result number association mapping data obtained in the previous step, the heap space is dynamically adjusted. Specifically, the system dynamically adjusts the allocation of heap space according to the fluctuation of search efficiency and the change in the number of returned results. Under high load, the system needs to allocate more heap space to accommodate more results; while under low load, the use of heap space can be reduced. This dynamic adjustment can effectively avoid over-allocation or waste of heap space, ensure that the system can maintain good memory usage efficiency under different load conditions, and reduce performance degradation caused by insufficient memory. The heap space dynamic adjustment data is released and switched to optimize fragmentation. In the process of dynamic allocation of heap space, memory fragmentation will be generated, especially after frequent heap space adjustment and memory release operations, there are some memory blocks that are not effectively used in the system. By switching to release fragment optimization, the system can process these fragments so that they can be effectively recycled or reallocated, avoiding memory waste and performance degradation caused by fragmentation. This step can effectively improve the utilization of heap space, avoid the reduction of memory management efficiency due to excessive fragmentation, and keep the overall memory resources of the system efficient and stable. According to the dynamic adjustment data of heap space and the release and switch fragment optimization data, the system performs adaptive allocation and regulation of heap space.This step intelligently adjusts the memory allocation strategy by comprehensively considering the dynamic demand for heap space and fragmentation optimization, so that the heap space can be adaptively optimized according to the actual load, retrieval efficiency fluctuations and memory usage. Through adaptive allocation control, the system can ensure efficient use of memory, and can automatically adjust resource allocation according to changes in retrieval efficiency and the number of results, avoiding performance bottlenecks caused by insufficient or excessive memory. Ultimately, this process greatly improves the overall stability and concurrent processing capabilities of the system and ensures efficient memory management.

[0056] Preferably, the release switching fragment optimization process for dynamically adjusting the heap space data includes the following steps:

[0057] According to the efficiency fluctuation-result quantity association mapping data, the heap space dynamic adjustment data is evaluated for the space switching release fragment intensity to obtain the space release fragment intensity data;

[0058] Marking the high-frequency address area of ​​the spatial release fragmentation intensity data to obtain the fragmentation intensity high-frequency address area;

[0059] Estimate the fragment release overhead for the high-frequency address area of ​​fragment intensity to obtain fragment release overhead data;

[0060] Perform adjacent region fragment merging processing on the fragment intensity high frequency address region according to the fragment release overhead data to obtain adjacent region fragment merging data;

[0061] Perform multi-threaded fragment compression processing on the adjacent region fragment merged data to obtain multi-threaded fragment compressed data;

[0062] The release switching fragment optimization processing is performed according to the adjacent area fragment merging data and the multi-thread fragment compression data to obtain the release switching fragment optimization data.

[0063] The present invention evaluates the intensity of fragmentation released by space switching for dynamic adjustment data of heap space according to "efficiency fluctuation-result quantity association mapping data". The core of this process is to analyze the intensity of fragmentation in the heap space, specifically, to evaluate which areas of memory fragmentation cause performance degradation or memory waste during the switching process. Through this evaluation, the system can identify highly fragmented memory areas and provide accurate data basis for subsequent optimization processing, avoiding excessive fragmentation release operations in unnecessary areas, thereby improving the overall memory management efficiency. High-frequency address areas are marked according to the space release fragmentation intensity data obtained by the evaluation. This process identifies which memory address areas have a greater impact on system performance by identifying frequently used areas in the memory. High-frequency address areas usually involve hot data or frequently accessed cache areas, and the fragmentation release of these areas requires special care to avoid frequent memory operations affecting performance. By marking these high-frequency address areas, the system can prioritize the optimization of fragmentation processing in these areas, thereby achieving more sophisticated memory management. The fragmentation release overhead of the marked high-frequency address areas with fragmentation intensity is estimated. This estimation takes into account the computing resources, time overhead and other memory operation burdens required for releasing fragments. Through this estimation, the system can identify high-overhead areas in the process of releasing fragments and make optimization decisions. For example, some areas have excessive overhead due to frequent memory operations, which affects the overall performance. The system will decide whether to delay processing or merge and optimize some high-overhead areas based on these data. Based on the fragment release overhead data, adjacent area fragments are merged for the high-frequency address area with high fragment intensity. Memory fragments are often distributed in different areas. Merging adjacent fragment areas can reduce the number of fragments and improve memory utilization. By merging adjacent fragment areas, the system can not only reduce the number of fragments, but also reduce the complexity of memory allocation, thereby improving system performance. This process is particularly suitable for scenarios where memory access patterns are relatively concentrated and fragmentation is severe. Multi-threaded fragment compression is performed on the merged adjacent area fragments. This process significantly improves the efficiency of memory fragment compression by using multiple threads to process fragment compression in parallel. Multi-threaded processing can effectively reduce system blockage or delay caused by defragmentation, especially on multi-core processors, which can make full use of parallel computing resources. This compression process can not only speed up the defragmentation process, but also reduce the performance bottleneck caused by single-threaded operation of the system, further improving the response speed and efficiency of memory management. The adjacent area fragment merging data and multi-threaded fragment compression data are combined to perform release switching fragment optimization processing. By integrating the optimization data of all the above steps, the system can formulate the best fragmentation release strategy to ensure that the fragmentation processing process is both efficient and does not cause additional performance loss when memory resources are tight. This step ensures efficient recovery of memory resources and avoids memory leaks or performance degradation caused by excessive fragmentation.The optimized data not only improves memory utilization, but also provides an accurate basis for subsequent memory management decisions.

[0064] Preferably, the present invention further provides a file management system for executing the file management method as described above, the file management system comprising:

[0065] The subject level analysis module is used to obtain archive text data; the subject level analysis of different subject contents is performed on the text subject content data to obtain the subject level data of the archive text;

[0066] The memory pointer structure design module is used to process the context perception between different subject content levels of the archive text data according to the archive text subject level data to obtain subject content context level perception data; perform associated expansion word retrieval on the subject content context level perception data to obtain semantic associated expansion word data; perform memory pointer structure design based on the semantic associated expansion word data to obtain associated expansion memory pointer structure data;

[0067] A virtual memory allocation module is used to perform heap space adaptive allocation control according to the associated extended memory pointer structure data to obtain heap space adaptive allocation control data; perform virtual memory temporary allocation scheduling based on the heap space adaptive allocation control data to obtain virtual memory temporary allocation scheduling data;

[0068] The archive management execution module is used to formulate an archive scheduling memory strategy based on the associated extended memory pointer structure data and the virtual memory temporary allocation scheduling data, obtain the archive scheduling memory strategy, and send the archive scheduling memory strategy to the server underlying system to execute archive management.

[0069] The invention has the beneficial effect that the invention optimizes a traditional archive management method, solves the problems of low accuracy of archive query and low retrieval efficiency in the traditional archive management method, improves the accuracy of archive query and improves the retrieval efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 A schematic diagram of the steps of a file management method;

[0071] Figure 2 for Figure 1 Detailed implementation steps of step S2 in FIG.

[0072] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0073] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0074] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0075] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0076] To achieve this, please refer to Figure 1 to Figure 2 , a file management method, the method comprising the following steps:

[0077] Step S1: Acquire archive text data; perform archive text theme level analysis between different theme contents on the text theme content data to obtain archive text theme level data;

[0078] Step S2: performing context perception processing between different subject content levels on the archive text data according to the archive text subject level data to obtain subject content context level perception data; performing associated extension word retrieval on the subject content context level perception data to obtain semantic associated extension word data; performing memory pointer structure design based on the semantic associated extension word data to obtain associated extension memory pointer structure data;

[0079] Step S3: performing heap space adaptive allocation control according to the associated extended memory pointer structure data to obtain heap space adaptive allocation control data; performing virtual memory temporary allocation scheduling based on the heap space adaptive allocation control data to obtain virtual memory temporary allocation scheduling data;

[0080] Step S4: Formulate an archive scheduling memory strategy based on the associated extended memory pointer structure data and the virtual memory temporary allocation scheduling data to obtain the archive scheduling memory strategy, and send the archive scheduling memory strategy to the underlying system of the server to execute archive management.

[0081] In the embodiment of the present invention, reference Figure 1 The above is a schematic diagram of the steps of a file management method of the present invention. In this example, the file management method includes the following steps:

[0082] Step S1: Acquire archive text data; perform archive text theme level analysis between different theme contents on the text theme content data to obtain archive text theme level data;

[0083] In an embodiment of the present invention, archival text data is extracted from a specified data source through a standardized document acquisition interface, and the data source includes structured and unstructured text forms, such as files stored in a document management system, records in a database, etc. The extracted archival text data is first cleaned to remove blank lines, invalid characters, and redundant content. The cleaning process is completed using a regular expression matching pattern and a text preprocessing rule set to ensure data quality. Subsequently, the text data is subjected to word segmentation and word frequency statistical analysis, and a set of subject words in the text is determined by a graph theory method based on a word co-occurrence network. According to the subject word set, the text is divided into fragments of different subject contents, and the fragments are clustered and analyzed, and a subject hierarchical relationship of the text is extracted using a subject modeling method based on latent semantic analysis (LSA). In the modeling process, the semantic similarity of each subject is calculated, and a hierarchical subject content tree structure is generated through a cosine similarity matrix, thereby obtaining archival text subject hierarchical data.

[0084] Step S2: performing context perception processing between different subject content levels on the archive text data according to the archive text subject level data to obtain subject content context level perception data; performing associated extension word retrieval on the subject content context level perception data to obtain semantic associated extension word data; performing memory pointer structure design based on the semantic associated extension word data to obtain associated extension memory pointer structure data;

[0085] In an embodiment of the present invention, according to the archival text subject hierarchical data generated in step S1, context perception processing is performed between different subject content hierarchies of the archival text data. The specific operation includes extracting a context relationship diagram from the subject hierarchical data, and using a text context modeling method based on a Markov chain to perceive and model the context of the text fragment. By calculating the probability of text context association, subject content context hierarchical perception data is generated. Next, the subject content context hierarchical perception data is retrieved for associated extended words. First, a semantically similar word set is generated using a word vector-based embedding technology (such as Word2Vec or GloVe), and an extended vocabulary is constructed in combination with keywords matched from a preset domain dictionary. By matching the extended vocabulary with the context of the perception data, the correlation between semantics is strengthened to obtain semantically associated extended word data. Based on the semantically associated extended word data, a memory pointer structure is designed. First, the extended words are sorted according to the strength of association between the extended words and the subject content, and then a pointer structure of a bidirectional linked list is constructed. Each node stores an extended word and its context weight. Finally, a chain rule is designed to support fast traversal and insertion operations, and associated extended memory pointer structure data is generated.

[0086] Step S3: performing heap space adaptive allocation control according to the associated extended memory pointer structure data to obtain heap space adaptive allocation control data; performing virtual memory temporary allocation scheduling based on the heap space adaptive allocation control data to obtain virtual memory temporary allocation scheduling data;

[0087] In an embodiment of the present invention, adaptive allocation and control of the heap space is first performed according to the associated extended memory pointer structure data. The access frequency and the size of the occupied memory block of the associated extended memory pointer data are analyzed, and an adaptive heap allocation algorithm based on a statistical dynamic distribution model is used to dynamically adjust the allocation strategy of the heap space. In this process, the access probability and the average occupied memory of each pointer structure are calculated, and the size and allocation position of the heap space block are adjusted according to these parameters, and finally the heap space adaptive allocation control data is generated. Subsequently, virtual memory temporary allocation scheduling is performed based on the heap space adaptive allocation control data. The priority queue algorithm is used to sort the memory pages, and the pointer data with high occupancy rate and frequent access is given priority. The scheduling algorithm is combined with the virtual memory paging strategy to realize the allocation and release scheduling of the memory page, and ensure the maximum access efficiency. Generate virtual memory temporary allocation scheduling data.

[0088] Step S4: Formulate an archive scheduling memory strategy based on the associated extended memory pointer structure data and the virtual memory temporary allocation scheduling data to obtain the archive scheduling memory strategy, and send the archive scheduling memory strategy to the underlying system of the server to execute archive management.

[0089] In an embodiment of the present invention, an archive scheduling memory strategy is formulated based on the associated extended memory pointer structure data and the virtual memory temporary allocation scheduling data. The storage location and allocation strategy of the pointer structure are tuned by a memory access optimization algorithm, and the segmented memory allocation method is used to minimize memory fragmentation. The specific implementation steps include: analyzing the bottleneck in the memory access path and redefining the pointer access priority by combining static analysis with dynamic analysis. Formulate time slice allocation rules for memory scheduling in combination with the paging strategy of the virtual memory scheduling data. Through the method of offline optimization and online update, ensure that the scheduling strategy can adapt to dynamically changing memory requirements. Finally, the archive scheduling memory strategy is written into the specified storage path in binary file format, and sent to the underlying server system through the interface for automatic execution to ensure the efficient operation of the archive management task.

[0090] Step S1 includes the following steps:

[0091] Step S11: Obtaining archive text data;

[0092] Step S12: Fill missing values ​​in the archive text data to obtain archive text filling data;

[0093] Step S13: classifying the archive text filling data by text subject content to obtain text subject content data;

[0094] Step S14: performing archive text theme level analysis between different theme contents on the text theme content data according to the text theme content data to obtain archive text theme level data.

[0095] In the embodiment of the present invention, the acquisition of archival text data depends on the strict design of the data collection process. By establishing a direct connection with the folder path, database table structure or designated path of the external storage medium in the archival storage system, data extraction is realized by using a standardized interface protocol (such as HTTP / RESTful interface or SQL query language). Archival data includes structured text (such as metadata table) and unstructured text (such as text converted from scanned files). During the extraction process, a list index is first established to record the metadata of each archival text with a unique identifier, including the creation date, version number, storage path and file name. For unstructured data, optical character recognition (OCR) technology is used to convert image content into text. The extracted text content is uniformly encoded according to the UTF-8 encoding standard to ensure data format consistency and is stored as a set of original archival text data for subsequent operations. The acquired archival text data is checked for integrity, and each paragraph of content is parsed one by one and the missing paragraphs, missing punctuation or damaged characters in the data are counted. The identification of missing values ​​is completed by combining regular expression rules with string matching algorithms. For example, when there is a missing title in the preset template or some key fields are not correctly extracted, an error log record is triggered. Missing value filling is implemented in two ways: for missing template fixed content, the archival metadata is directly used to complete the completion; for missing semantic integrity, the potential semantic connection between adjacent sentences is inferred by the filling algorithm based on the word co-occurrence matrix to generate semantic continuation paragraphs. In addition, for the character damage in small text segments, the character-level spelling correction algorithm is used to automatically repair it. After the processing is completed, the archival text filling data is generated and stored. The text subject content classification is achieved through a multi-stage text analysis method. First, the keyword extraction algorithm based on the word frequency-inverse document frequency (TF-IDF) weight is used to extract the important feature words of each text segment in the filling data, and the results are mapped into feature vectors. Then, the hierarchical clustering method is used to divide these feature vectors into several preliminary subject categories. In the clustering process, the Euclidean distance is used to calculate the similarity between feature vectors, and the subject level is recursively analyzed based on the hierarchical tree clustering structure. Finally, the preliminary classification results are verified by domain expert knowledge rules, and the clustering results are corrected in combination with the predefined classification dictionary, and the classified text content is stored as text subject content data. According to the text subject content data generated in step S13, the hierarchical analysis of the archive text subject is further performed. First, the association between subjects is calculated using a hierarchical modeling algorithm based on semantic similarity (such as Latent Dirichlet Allocation, LDA). The semantic similarity of the subject is achieved through word vector embedding calculation, and the text content is mapped to a high-dimensional space to analyze the semantic distance between subjects. Then, a subject association matrix is ​​generated based on the semantic distance, and a subject tree hierarchy is constructed based on the matrix. In this process, each node represents a specific subject, and each edge represents the association weight between subjects.Finally, the important nodes in the tree structure are weighted hierarchically to obtain the hierarchical data of the archival text topics, which are then output and stored for use in subsequent steps.

[0096] Step S2 includes the following steps:

[0097] Step S21: performing context perception processing between different subject content levels on the archive text data according to the archive text subject level data to obtain subject content context level perception data;

[0098] Step S22: performing semantic enhancement segment analysis on the subject content context level perception data to obtain content level semantic key segments;

[0099] Step S23: performing a search for associated expansion words on the subject content context level perception data according to the content level semantic key fragments to obtain semantic associated expansion word data;

[0100] Step S24: Design a memory pointer structure based on the semantically associated extended word data to obtain associated extended memory pointer structure data.

[0101] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0102] Step S21: performing context perception processing between different subject content levels on the archive text data according to the archive text subject level data to obtain subject content context level perception data;

[0103] In the embodiment of the present invention, the context perception processing is based on the archival text subject level data and is realized by analyzing the contextual relationship between the text subject levels. First, the context window of each subject paragraph in the archival text data is extracted using the sliding window method, and the window length is set to a dynamic value related to the subject relevance. Then, based on a predefined semantic similarity evaluation method (such as cosine similarity), the vocabulary in the window is calculated and the highly associated context paragraphs are marked. Next, in combination with the semantic association matrix, the core vocabulary in the context is matched with the semantic core words of the adjacent subject level to generate an association score table. For the paragraphs in the association score table whose association degree is higher than the threshold, a context enhancement strategy is added to make the semantics between the subject paragraphs more coherent by inserting dynamically generated connecting sentences. Finally, the processed subject content context level perception data is output and stored for subsequent analysis.

[0104] Step S22: performing semantic enhancement segment analysis on the subject content context level perception data to obtain content level semantic key segments;

[0105] In an embodiment of the present invention, the goal of semantically enhanced fragment analysis is to extract key fragments that have a significant impact on the semantic level from the subject content context level perception data. First, the input context level perception data is decomposed into multiple topic paragraphs, and a list of core words for each paragraph of text is generated based on TF-IDF weight calculation. Then, the sentence structure in the text is parsed using a syntactic analysis tool to extract the subject-predicate-object structure and modifying components to ensure that key semantics are not missed. The parsed sentences are sorted according to semantic completeness and relevance to the context, and a bidirectional semantic hierarchical algorithm is used to align high-level fragments with underlying semantic features. After the sorting is completed, the fragments with significance higher than the average level are marked as semantic key fragments, and the content-level semantic key fragment data is output.

[0106] Step S23: performing a search for associated expansion words on the subject content context level perception data according to the content level semantic key fragments to obtain semantic associated expansion word data;

[0107] In an embodiment of the present invention, the associated extended word retrieval is based on the semantic key fragments at the content level, and the extended words are generated by searching the associated dictionary and the similar word library. First, the core words in the semantic key fragments at the content level are used as the seed words for retrieval, and the initial matching is performed in combination with the predefined semantic classification dictionary to record all the extended word groups directly related to the seed words. Secondly, based on the second-order word co-occurrence algorithm, the co-occurrence frequency of the extended words and other words in the subject content context level perception data is counted to determine the semantic groups with high frequency co-occurrence. Finally, these extended word groups are input into the semantic verification link to verify the contextual applicability of each extended word, and filter out low-relevance words to form the final semantic associated extended word data.

[0108] Step S24: Design a memory pointer structure based on the semantically associated extended word data to obtain associated extended memory pointer structure data.

[0109] In an embodiment of the present invention, the memory pointer structure is designed to optimize the storage and access efficiency of the associated extended word data. First, based on the lexical frequency statistics of the semantically associated extended word data, the phrases are sorted according to the access priority, and the high-frequency phrases are allocated to the efficient storage segment in the memory. Then, a prefix tree (Trie) structure is used to establish a fast retrieval index for the extended words, and each phrase is mapped to the pointer table to form a pointer linked list. For phrases with subset relationships, the dynamic hash partitioning method is used to split the prefix tree nodes to avoid storage redundancy. Finally, a bidirectional link mechanism is added to the pointer structure to support reverse tracing queries from the extended words to the original semantic words, ensuring the multi-directional access capability of the data. After the design is completed, the associated extended memory pointer structure data is generated and stored, providing a basis for subsequent storage optimization and retrieval operations.

[0110] Step S24 includes the following steps:

[0111] Step S241: performing context relevance evaluation on the subject content context level perception data based on the semantically associated extended word data to obtain extended word-subject content relevance data;

[0112] Step S242: Designing an expansion word chain pointer rule for the semantically associated expansion word data according to the expansion word-subject content correlation data to obtain an expansion word chain pointer design rule;

[0113] Step S243: matching the pointer allocation synchronization mechanism with the extended word chain pointer design rule to obtain the extended word chain pointer design synchronization rule;

[0114] Step S244: performing chain parallel partitioning on the extended word chain pointer design synchronization rule to obtain chain parallel partition pointer design data;

[0115] Step S245: Design a memory pointer structure according to the chained parallel partition pointer design data to obtain associated extended memory pointer structure data.

[0116] In an embodiment of the present invention, the context relevance evaluation first aligns the vocabulary in the semantically associated extended word data with the context fragments in the subject content context level perception data. The extended word data is segmented by a semantic segmentation method, high-frequency phrases are extracted, and a context matrix is ​​constructed in combination with the subject fragments in the context level perception data. The rows of the context matrix represent extended words, and the columns represent context subject fragments. Subsequently, the point mutual information (PMI) method is used to calculate the relevance score between each extended word and the subject fragment. The word pairs whose relevance scores exceed the set threshold are recorded in the extended word-subject content relevance table. Next, the data in the extended word-subject content relevance table is subjected to clustering-based dimensionality reduction processing to optimize the data storage structure. Finally, the extended word-subject content relevance data is output and saved for the next step. The extended word chain pointer rule is designed to efficiently organize and manage the extended word and its associated subject content, and the chain pointer rule is used to realize the fast access and semantic navigation of dynamic extended words. In the implementation process, the extended word-subject content relevance data must first be preprocessed, the non-important associated items in the relevance score matrix are removed, and only the mapping relationship between the extended word and the subject content with high relevance is retained. The threshold of high correlation is determined according to the standard distribution of statistical analysis, and is usually taken as the mean of the correlation score plus two standard deviations. Next, the processed data is grouped. Based on the semantic association between the extension word and the subject content, the agglomerative hierarchical clustering algorithm is used to divide the extension word into multiple semantic groups. In the specific implementation, the semantic distance between the extension words is first calculated, and the cosine similarity formula is used to measure the correlation between the two extension words. After the distance matrix is ​​generated, semantic clusters are formed by gradually merging the extension words with high similarity, and each cluster finally corresponds to a semantic subchain. After completing the semantic clustering, a chain pointer rule is designed for each subchain. The design of the rule includes key contents such as node structure definition, pointer jump logic and storage address allocation. First, the standard structure of the linked list node is defined for each extension word, and the node must contain the following fields: extension word identifier (uniquely marking the word), subject content mapping pointer (pointing to the corresponding subject content data block address), front and back pointers (pointing to the previous and next nodes in the linked list) and extended metadata fields (used to record auxiliary information such as word frequency and semantic tags). Then, the pointer jump logic is designed according to the structure of the linked list node. The jump logic needs to consider the efficiency of linked list access, and the skip table structure is used to optimize the traditional linear access path of the linked list. In the skip table design, multiple jump levels are added to each linked list node, and each layer records the cross-chain pointer from the node to the distant node. The access efficiency is improved by setting different jump intervals. For example, the first-level jump pointer of the linked list node can skip one node, the second level can skip two nodes, and so on. The specific jump interval is determined by a dynamic programming algorithm. Finally, the storage address is allocated for the linked list and its skip table structure.The partition memory allocation method is adopted to allocate semantically similar linked list nodes to physically adjacent storage areas in priority, so as to reduce memory fragmentation and improve data access performance. The memory address is realized by hash mapping, and the extended word identifier is converted into a partition address through a hash function to ensure that the logical and physical distribution of the linked list nodes in the memory are highly consistent. After the rule design is completed, the extended word chain pointer rule is output in the form of structured data and stored in a dedicated extended word index file. The file format adopts variable length encoding, which can not only effectively compress the data volume, but also support fast decoding and real-time access, laying the foundation for the subsequent dynamic operation and expansion of chain pointers. The purpose of pointer allocation synchronization mechanism matching is to ensure the synchronization and consistency of pointer access in a multi-task environment based on the chain pointer rule. First, the linked list node distribution and jump conditions in the extended word chain pointer design rule are parsed to establish a priority queue for pointer access. Secondly, the optimistic locking mechanism is introduced to synchronize the pointer allocation and access process. In the specific implementation, a timestamp field is embedded in the pointer data of each linked list node to record the time of the most recent access operation. When the pointer allocation task is executed, the timestamp is compared to confirm whether the pointer state is available, and the timestamp is updated after the task is completed. Finally, the extended word chain pointer design synchronization rule is output to ensure the stable operation of the linked list structure in a concurrent environment. Chain parallel partitioning takes the extended word chain pointer design synchronization rule as input and optimizes the access efficiency by partitioning the linked list nodes. First, the access frequency and association strength of each node in the linked list are analyzed, and the high-frequency access nodes are marked as key nodes. Then, according to the distribution of access frequency and association strength, the linked list is divided into multiple independent parallel partitions, each partition contains one or more key nodes and their associated nodes. The partitioning process uses the K-means clustering algorithm to assign nodes to the optimal partition, while ensuring that the pointer jump path within each partition is the shortest. For the connection between partitions, a bidirectional linked list structure is used to ensure efficient cross-region access between partitions. Finally, the chain parallel partition pointer design data is generated. The specific implementation of chained parallel partitioning of chained pointer design synchronization rules for extended words needs to be based on the theory of distributed data structure optimization, and adopts a recursive partitioning algorithm combined with a binary tree partitioning strategy to ensure the efficiency and balance of the pointer structure. First, the data in the chained pointer design synchronization rules for extended words are hierarchically classified according to semantic relevance, frequency of use, and topic priority. The classification criteria are determined by a preset multidimensional matrix, in which each dimension corresponds to the core characteristics of the extended words. Subsequently, the hash partitioning strategy is used to generate initial partition nodes for the classification results, and high-frequency extended words and low-frequency extended words are assigned to different node groups to reduce conflicts in memory operations. Next, chain reconstruction is performed on each partition node, and a cross-partition chain connection is gradually established through a pointer recursive algorithm to ensure that fast jumps between parallel nodes can be achieved during efficient retrieval.In this process, it is necessary to calculate the access time weight for each pointer path, and use the dynamic programming optimization algorithm to adjust the priority path order of the chain pointer, and finally generate the chain parallel partition pointer design data. The specific implementation of the memory pointer structure design based on the chain parallel partition pointer design data needs to take memory management technology as the core, and combine memory mapping and address conversion mechanism to ensure data access efficiency. First, the chain parallel partition pointer design data is loaded into the memory space, and a virtual memory mapping table is constructed at the same time, and the address conversion function is used to realize the efficient conversion from logical address to physical address. Subsequently, the linked list structure is used to extend the memory pointer, and the pointer nodes of the chain parallel partition are organized in the form of a dynamic linked list, and the memory page size is allocated according to the access weight. For high-priority pointers, fixed page frames are allocated to reduce the frequent page scheduling overhead; for low-priority pointers, an elastic page allocation algorithm is used to dynamically adjust the allocation size by real-time monitoring of usage. Finally, the recursive lock mechanism is used to implement thread-safe control of pointer access to prevent memory conflicts and data errors caused by concurrent access. After the above operations, the generated associated extended memory pointer structure data can ensure efficiency and stability in complex archive management scenarios.

[0117] Designing extension word chain pointer rules for semantically associated extension word data based on extension word-subject content relevance data includes the following steps:

[0118] Performing extended word context position difference analysis on semantically related extended word data to obtain extended word context position difference data;

[0119] Based on the extended word context position difference data, the extended word-topic content relevance data is quantified with respect to the content relevance weight, thereby obtaining the extended word-topic content relevance weight quantification data;

[0120] According to the weight quantification data of the correlation between the extended word and the subject content, a bidirectional linked list pointer structure is constructed for the semantically associated extended word data to obtain an extended word associated linked list pointer structure;

[0121] Based on the extended word associated linked list pointer structure, a pointer reconfigurable linked list is constructed to obtain a pointer reconfigurable linked list;

[0122] According to the pointer structure of the extension word association chain list and the pointer reconfigurable chain list, the extension word chain pointer rule is designed for the semantically associated extension word data, and the extension word chain pointer design rule is obtained.

[0123] In the embodiment of the present invention, when the semantically associated extended word data is analyzed for the difference in the extended word context position, it is necessary to combine the natural language processing technology, analyze the contextual relationship of the semantically associated extended words in different text topics, and extract the position and frequency distribution of each extended word in its context. First, the semantically associated extended word data is segmented into sentences using the text segmentation technology, and the position distribution of the extended words in each subject content is recorded by constructing a word frequency-context matrix. Subsequently, the cosine similarity calculation method is used to compare the context difference values ​​of the extended words in different contexts, thereby forming a context position difference matrix for each extended word. Through the statistical analysis of multidimensional difference data, the key difference dimensions are extracted in combination with the principal component analysis method, and finally the extended word context position difference data is generated to support the calculation of subsequent correlation weights. When the content correlation weight is quantified for the extended word-subject content correlation data based on the extended word context position difference data, it is necessary to design a weight allocation algorithm to quantify the context difference value between the extended word and the subject content into a comparable correlation weight. First, the extended word context position difference data is used as an input variable, matched with the extended word-subject content correlation data, and a weighted mapping relationship matrix is ​​constructed. Then, the TF-IDF (term frequency-inverse document frequency) algorithm is used, combined with the difference value correction coefficient, to calculate the contribution of each extension word in different topics and output the weighted relevance value. Finally, the weight values ​​of all extension words are adjusted to a unified interval using the normalization method to generate the weight quantification data of the extension word-topic content relevance. When constructing the bidirectional linked list pointer structure of the semantically associated extension word data based on the weight quantification data of the extension word-topic content relevance, it is necessary to efficiently organize the extension word pointer linked list based on the data relevance. First, the extension words are arranged in descending order according to the weight quantification data to determine the order of the linked list nodes. Subsequently, the linked list construction algorithm is used to assign a storage address to each extension word node, and its front and back nodes are connected respectively through bidirectional pointers to realize the bidirectional traversal function. At the same time, by embedding the relevance weight and context position difference information of the extension word in the linked list node, each node can carry multi-dimensional data. After the construction is completed, the structural integrity of the linked list is verified for consistency to ensure that the pointer connection is error-free, and finally the extension word associated linked list pointer structure is generated. When constructing a pointer reconfigurable linked list based on the extended word associative linked list pointer structure, it is necessary to introduce a dynamic reconstruction algorithm to enhance the flexibility and adaptability of the linked list. First, the nodes in the associative linked list pointer structure are monitored in real time, and the node access frequency and response time are recorded. The frequently accessed nodes are grouped through the hash mapping mechanism and the priority path pointer is set to reduce the average time complexity of the retrieval. Subsequently, the low-frequency nodes are dynamically grouped using the reconstruction rules, and the nodes with higher association are merged into short chains to optimize the overall linked list structure. During the reconstruction process, the recursive locking technology is used to ensure the security of the pointer update operation and avoid data conflicts during concurrent access.Finally, the construction of the pointer reconfigurable linked list is completed, and the corresponding linked list data is generated. When designing the extended word chain pointer rules for semantically associated extended word data based on the extended word associated linked list pointer structure and the pointer reconfigurable linked list, it is necessary to design rules to guide the specific operation of the extended word linked list. First, according to the characteristics of the linked list structure and the pointer reconstruction rule, the construction logic of the chain pointer is defined, including the node sorting rules, the pointer path selection strategy and the dynamic update trigger condition. Subsequently, the finite state machine model is introduced to describe the state transition rules of the linked list in different operation scenarios to ensure the universality and consistency of the rule design. After the rule design is completed, its feasibility is verified, the actual linked list operation process is simulated and potential conflicts are detected. Finally, the rules are optimized based on the verification results, and the extended word chain pointer design rules are output.

[0124] Designing synchronization rules for chain pointers of extended words to perform chain parallel partitioning includes the following steps:

[0125] The pointer traversal retrieval time complexity is simulated for the synchronization rule of the extended word chain pointer design, and the pointer traversal retrieval time complexity is obtained;

[0126] According to the pointer traversal retrieval time complexity, the synchronization rule of the extended word chain pointer design is matched with the limit chain length to obtain the limit chain length matching data;

[0127] Based on the extreme chain length matching data and the expansion word-topic content relevance data, the expansion word chain pointer design synchronization rules are used to construct contextual parallel subchains to obtain the expansion word contextual parallel subchains;

[0128] Perform subchain incremental constraint analysis on the parallel subchains in the context of the extended word to obtain subchain incremental constraint data;

[0129] According to the extended word context parallel sub-chains and sub-chain incremental constraint data, the extended word chain pointer design synchronization rule is chained and partitioned to obtain chained parallel partition pointer design data.

[0130] In the embodiment of the present invention, when simulating the time complexity of pointer traversal retrieval for the synchronization rule of chain pointer design for extended words, a simulation environment needs to be constructed to measure the traversal performance of the chain list structure of extended words. First, according to the current chain list structure of the synchronization rule of chain pointer design for extended words, each chain list node is assigned the extended words and correlation data it contains. Subsequently, the linear scanning algorithm is used to access the pointers one by one from the starting node to the tail node of the chain list, and the relationship between the number of nodes and the access time during the traversal process is recorded. By performing multiple traversal tests under different chain list length conditions, the total access time of each traversal is extracted, and the growth trend of the access time is fitted using a piecewise linear regression analysis method, thereby estimating the time complexity of the chain list. Based on the time complexity obtained from the experiment, the pointer traversal retrieval time complexity data is output. When matching the limit chain length of the synchronization rule of chain pointer design for extended words according to the time complexity of pointer traversal retrieval, it is necessary to set the optimal segmentation strategy of the chain list to balance the retrieval efficiency and the complexity of the chain list structure. First, the time complexity data is used to determine the critical point of the access performance of the current chain list at different lengths, that is, the chain list length at which the traversal time increases significantly. Combined with this critical point, the limit value of the chain segment length is set, and the linked list nodes are regrouped through a recursive segmentation algorithm. Each group of nodes needs to keep the most relevant words closely connected, while minimizing the number of pointers to optimize access efficiency. The limit chain length matching data is generated by statistics and used to guide subsequent partitioning operations. When designing synchronization rules for the extended word chain pointer based on the limit chain length matching data and the extended word-topic content relevance data to construct contextual parallel subchains, it is necessary to design an algorithm to group the nodes in the linked list in parallel according to topic relevance and contextual position. First, the extended word linked list nodes are divided into several initial subchains according to the limit chain length matching data. Subsequently, the nodes in each subchain are semantically analyzed through the extended word-topic content relevance data, and the semantically closely related nodes are adjusted to the same subchain using the semantic similarity calculation method (such as WordNet or Jaccard coefficient), and the word order consistency is ensured in the context. After the subchain grouping is completed, a bidirectional pointer structure is constructed for each subchain for fast access, and finally the extended word contextual parallel subchain is output. When performing incremental constraint analysis on the extended word context parallel subchains, it is necessary to evaluate and limit the dynamic expansion behavior of the subchain structure to ensure the stability of the linked list. First, set the maximum and minimum thresholds for the number of subchain nodes, and mark the subchains that exceed the threshold as target subchains by calculating the rate of change of the number of nodes in each subchain. For the target subchain, analyze the semantic features and contextual relevance of its extended nodes, and calculate the impact of the addition of new nodes on the complexity of the linked list structure. Combined with the incremental constraint conditions, the dynamic adjustment algorithm is used to prune or merge the target subchain to ensure that its structure remains controllable in the incremental case. After the analysis is completed, the subchain incremental constraint data is generated to guide subsequent partitioning.When chaining and parallel partitioning the synchronization rules of chain pointer design of extended words according to the contextual parallel subchains of extended words and the incremental constraint data of subchains, it is necessary to divide the entire linked list into independent but interrelated partitions to optimize its storage and retrieval efficiency. First, using the contextual parallel subchain as the initial partition unit, the subchain is clustered according to the node characteristics in its incremental constraint data. Subchains with similar correlation distribution are assigned to the same partition through the K-means clustering algorithm. Subsequently, a global index pointer is added to each partition for fast switching between partitions, and the access priority of the linked list within the partition is defined. After the partition is completed, the structural integrity is traversed to ensure that there is no conflict in the partition logic, and finally the chained parallel partition pointer design data is generated.

[0131] Step S3 includes the following steps:

[0132] Step S31: performing concurrent search efficiency simulation on the associated extended memory pointer structure data to obtain pointer structure concurrent search efficiency data;

[0133] Step S32: performing heap space adaptive allocation control according to the pointer structure concurrent retrieval efficiency data to obtain heap space adaptive allocation control data;

[0134] Step S33: Perform virtual memory temporary allocation scheduling based on the heap space adaptive allocation control data and the pointer structure concurrent retrieval efficiency data to obtain virtual memory temporary allocation scheduling data.

[0135] In the embodiment of the present invention, when performing concurrent retrieval efficiency simulation, it is first necessary to conduct a detailed analysis of the node distribution and storage layout of the associated extended memory pointer structure data to clarify the access path and overlapping parts of the nodes in the linked list. A multi-threaded simulation environment is constructed, and a fixed thread pool technology is used to define multiple threads to perform concurrent operations on the pointer structure. The tasks of each thread include randomly reading pointer nodes, inserting new nodes, and deleting specific nodes. The efficiency of different operations in the concurrent environment is quantified by recording the execution time of each thread and the frequency of resource conflicts between threads. The thread scheduling is managed by a priority queue to simulate the access of different priority tasks to the pointer structure. Through these operations, the performance bottleneck caused by competition between threads is captured, and the granular lock mechanism is used to control the access to shared resources, and the lock waiting time and unlocking time are recorded to measure the overall retrieval efficiency. Finally, the time data of all thread executions are counted and classified, and the pointer structure concurrent retrieval efficiency data is generated by analyzing the response time distribution and task completion rate. In the process of adaptive allocation and regulation of heap space, key performance indicators such as retrieval delay, thread conflict rate, and memory allocation failure rate are first extracted from the pointer structure concurrent retrieval efficiency data. Based on these data, a set of allocation strategies based on dynamic programming is designed. This strategy manages the partition of heap space and reserves memory blocks of different sizes to match different types of pointer node operations. The memory pool technology is used to divide multiple memory sub-pools, each of which corresponds to a common pointer operation mode, and the memory allocation ratio is dynamically adjusted according to actual needs. In the implementation process, a fast allocation algorithm (such as a memory block algorithm) is used to prioritize the allocation of small objects, and a partition memory merging strategy is introduced for large objects to reduce fragmentation. By monitoring the real-time data of memory allocation and release, the allocation ratio of heap space is adjusted to achieve the optimal memory usage efficiency. The output data of this step is the heap space adaptive allocation control data, which contains detailed information such as the partition ratio, allocation mode, and partition memory status. When performing virtual memory temporary allocation scheduling, it is necessary to integrate the heap space adaptive allocation control data and the pointer structure concurrent retrieval efficiency data to construct a set of virtual memory allocation and scheduling rules. First, by analyzing the memory areas frequently accessed in pointer operations, high-frequency access addresses are extracted and preferentially allocated to physical memory to improve access speed. For low-frequency access addresses, page replacement algorithms (such as LRU algorithms) are used to temporarily store them in virtual memory to free up physical memory space. The memory access pattern prediction technology is used to predict the access area at the next moment by counting the historical memory access data, and dynamically adjust the page exchange frequency between virtual memory and physical memory. In order to improve the scheduling efficiency, combined with the memory partitioning strategy, a temporary memory area is allocated to store the pointer data required for large-scale short-term linked list node operations. The optimal use of memory resources is achieved by adjusting the page size and allocation ratio of virtual memory in real time.Finally, the virtual memory temporary allocation scheduling data is output, which contains information such as the specific location, allocation time and scheduling results of each page allocation for subsequent operations.

[0136] Step S32 includes the following steps:

[0137] Step S321: performing efficiency fluctuation range analysis on the pointer structure concurrent retrieval efficiency data to obtain the concurrent retrieval efficiency fluctuation range;

[0138] Step S322: extracting the number of search results from the pointer structure concurrent search efficiency data to obtain the number of search results;

[0139] Step S323: mapping the search result quantity value according to the concurrent search efficiency fluctuation range to obtain efficiency fluctuation-result quantity association mapping data;

[0140] Step S324: dynamically adjust the heap space according to the efficiency fluctuation-result quantity association mapping data to obtain heap space dynamic adjustment data;

[0141] Step S325: performing release switching fragment optimization processing on the heap space dynamic adjustment data to obtain release switching fragment optimization data;

[0142] Step S326: performing heap space adaptive allocation control according to the heap space dynamic adjustment data and the release switching fragment optimization data to obtain heap space adaptive allocation control data.

[0143] In the embodiment of the present invention, the implementation process of efficiency fluctuation interval analysis first needs to extract efficiency indicators from the pointer structure concurrent retrieval efficiency data, including but not limited to retrieval time, thread scheduling time, resource occupancy rate, etc. The above data is segmented into time series by piecewise linear fitting method, and the efficiency change amplitude of each segment is calculated to identify the fluctuation interval. The efficiency data is analyzed in frequency domain by Fourier transform technology to determine the periodic characteristics and trend characteristics of the fluctuation interval. The high-frequency fluctuation area is marked by sliding window technology, and the maximum, minimum and mean values ​​of the efficiency in each time period are recorded in a statistical manner. At the same time, in order to enhance the accuracy of the analysis results, the monotonicity analysis method is used to detect the rising or falling trend of the fluctuation interval, and it is classified as a growth fluctuation interval or a decline fluctuation interval. Finally, the start and end time, fluctuation amplitude and fluctuation characteristic data of each fluctuation interval are recorded by matrix storage, and the concurrent retrieval efficiency fluctuation interval is output. The implementation process of extracting the number of retrieval results needs to combine the pointer structure concurrent retrieval efficiency data with the pointer access log, and extract by analyzing the retrieval path of the pointer node operation and the corresponding number of returned results. Firstly, the statistical regression method is used to analyze the relationship between the number of accesses of each thread and the amount of data successfully retrieved, and the number of nodes returned by each retrieval is marked in the form of a scatter plot. Subsequently, a data structure based on a hash table is designed to store the operation frequency of each thread and the corresponding number of retrieval results. By enumerating the retrieval results, the maximum, minimum and median values ​​are recorded, and a frequency distribution histogram is generated to clarify the mainstream retrieval return volume range. Finally, the extracted data is normalized to facilitate the subsequent efficiency fluctuation and result number mapping and correlation analysis, and the retrieval result number value is obtained. According to the concurrent retrieval efficiency fluctuation interval, the retrieval result number value of the corresponding time period is associated and mapped. In this process, the efficiency fluctuation interval and the retrieval result number value are first aligned with a unified timestamp to ensure data consistency. Then, a two-dimensional matrix mapping method is used to construct a relationship matrix between the efficiency fluctuation and the number of results, with the fluctuation interval as the horizontal axis and the number of retrieval results as the vertical axis. The data of each unit in the relationship matrix is ​​normalized, and the significance of the mapping data is detected by the chi-square test method to filter out the influence of statistical noise. Through curve fitting technology, linear or nonlinear modeling is performed on the highly correlated fluctuation intervals and search result quantity points to extract the correlation pattern between efficiency fluctuations and the number of results, and form efficiency fluctuation-result quantity correlation mapping data. In the process of dynamic adjustment of heap space, according to the relationship between efficiency fluctuation-result quantity correlation mapping data, the memory distribution characteristics required for efficient retrieval and high search result return volume are analyzed. The dynamic partitioning algorithm is used to divide the heap space into multiple subspaces, each of which corresponds to a mainstream interval of the mapping data pattern. During the adjustment process, the real-time memory management technology is used to monitor the memory occupancy and access frequency of each subspace, and the distribution ratio of the heap space is dynamically adjusted through the memory allocator.The memory fragmentation recovery mechanism is introduced. The free space is sorted by the mark-and-sweep method, and the scattered memory blocks are re-merged for efficient use. Finally, the dynamic adjustment data of the heap space is output, which includes the allocation ratio, occupancy status and adjustment records of each subspace. The release switching fragment optimization processing requires a detailed analysis of the fragmentation in the dynamic adjustment data of the heap space. First, the distribution of memory fragments is calculated by the block algorithm, and the fragment size, location and frequency characteristics are recorded. The segmentation strategy is adopted to separate large fragments from small fragments. For large fragments, a bidirectional adaptation algorithm is applied to allocate them to the high-frequency access pointer area first; for small fragments, multiple small areas are reassembled into large memory blocks through memory merging technology. The adaptive release algorithm is adopted to release the temporarily occupied memory area in real time without affecting the main storage task. In order to optimize the switching performance, the memory release and allocation process is synchronously recorded based on the page table management technology, and finally the release switching fragment optimization data is generated. The implementation of the adaptive allocation control of the heap space first requires the integration of the dynamic adjustment data of the heap space and the release switching fragment optimization data to construct a dynamic adjustment strategy. Based on the usage of partitions, the heap space is adaptively optimized through a heuristic allocation algorithm. During the control process, the linear programming method is used to optimize the memory usage efficiency, and the dynamic range of the minimum allocation block and the maximum allocation block is set to ensure that the memory requirements of high-frequency operations are met. The memory usage curve is monitored in real time, and the problem of sudden demand during the memory allocation process is solved by using a fast compensation allocation mechanism. At the same time, the memory is used to prevent over-allocation algorithms to avoid the overall performance degradation caused by uneven heap space allocation. Finally, the heap space adaptive allocation control data is generated, which contains real-time records of the control process and the adjusted allocation strategy.

[0144] The process of dynamically adjusting the data in the heap space to release the switching fragmentation optimization process includes the following steps:

[0145] According to the efficiency fluctuation-result quantity association mapping data, the heap space dynamic adjustment data is evaluated for the space switching release fragment intensity to obtain the space release fragment intensity data;

[0146] Marking the high-frequency address area of ​​the spatial release fragmentation intensity data to obtain the fragmentation intensity high-frequency address area;

[0147] Estimate the fragment release overhead for the high-frequency address area of ​​fragment intensity to obtain fragment release overhead data;

[0148] Perform adjacent region fragment merging processing on the fragment intensity high frequency address region according to the fragment release overhead data to obtain adjacent region fragment merging data;

[0149] Perform multi-threaded fragment compression processing on the adjacent region fragment merged data to obtain multi-threaded fragment compressed data;

[0150] The release switching fragment optimization processing is performed according to the adjacent area fragment merging data and the multi-thread fragment compression data to obtain the release switching fragment optimization data.

[0151] In an embodiment of the present invention, the implementation process of the spatial switching release fragmentation strength evaluation starts from the efficiency fluctuation-result quantity association mapping data, and firstly extracts the frequently used memory areas by analyzing the efficient search intervals in the mapping data. For these frequently used areas, the fragmentation strength is calculated using the weighted average method according to the number of accesses and the degree of fragmentation of each address. The calculation of the fragmentation strength takes into account two factors: one is the frequency of use of the memory area, and the other is the distribution density of the memory block. By comparing the free space of adjacent memory blocks, a fragmentation strength scoring data is generated, and the fragmentation strength score is matched with the spatial switching release demand. Using space optimization technology, combined with the average fragmentation degree and memory usage fluctuation data, the switching release potential of each memory area is calculated to obtain the spatial release fragmentation strength data. The data includes the fragmentation strength value of each memory area, the priority of the switching release and the resource occupancy ratio. The process of marking the high-frequency address area evaluates the memory address area through the fragmentation strength value in the spatial release fragmentation strength data, and divides the memory area according to the high and low fragmentation strength values. First, a fragmentation strength threshold is set, and the memory area above the threshold is marked as a high-frequency address area. The determination of this threshold is based on the optimization goal of the memory allocator and the memory utilization requirement. Usually, a fragmentation intensity of more than 80% is selected as the high-frequency marking standard. Then, the memory address is mapped to a two-dimensional coordinate system, and the memory space is divided into multiple sub-regions using the region segmentation algorithm. The fragmentation intensity of each sub-region is combined with the access frequency data to form a region marking matrix. Through this method, those memory address regions with high fragmentation intensity and frequent access can be effectively marked. These regions will be given priority in the subsequent fragmentation processing steps, and finally the fragmentation intensity high-frequency address regions are obtained. The implementation method of fragment release overhead estimation is based on the high-frequency address region marking results, and the time, space and computational cost required for the memory release operation of each high-frequency address region are used as the cost indicators. First, the memory management algorithm is used to calculate the memory cleanup time required for the release of each high-frequency address region, which includes the time spent on free space arrangement and fragmentation removal. Secondly, the CPU load of each release operation is evaluated by analyzing the memory page table operations and memory reallocation calculations involved in the memory release process. Furthermore, based on the continuity analysis of the heap space, the additional memory space required after the fragments are merged is calculated for reallocation. All these overhead data are weighted and integrated by statistical methods to obtain the final fragment release overhead data. This data includes the release time overhead, CPU load, and additional space overhead of memory reallocation for each high-frequency address area, and finally forms the fragment release overhead data. The process of merging adjacent region fragments first identifies adjacent high-frequency address areas in the memory based on the fragment release overhead data. These areas usually have continuous or close memory blocks. By merging adjacent region fragments, the degree of memory fragmentation can be effectively reduced.First, a merge rule algorithm is used to set a merge threshold to determine whether the fragmentation strength of adjacent memory regions is sufficient for merging. If the fragmentation strength of both is low and the memory usage demand is large, they can be merged. Then, these adjacent regions are merged using the region merging algorithm. The merging process includes updating the memory page table information and adjusting the space allocation strategy of the memory allocator. In order to ensure the efficiency of memory allocation and that the merged region will not cause additional fragmentation in future retrieval processes, internal fragmentation detection is performed after merging, and the allocation scheme of the memory region is optimized by adjusting the memory layout. Finally, the adjacent region fragment merging data is obtained, which records the memory region of the merge operation, the fragmentation strength changes before and after the merge, and the memory adjustment data during the merge process. Implementation of multi-threaded fragment compression processing First, the adjacent region fragment merging data is used to process the fragments of multiple memory regions in parallel. A multi-threaded parallel compression algorithm is used to process memory fragments, in which each thread is responsible for processing different memory regions. Each thread reduces the number of memory fragments and optimizes the memory storage structure through the compression algorithm. First, for each memory region, a greedy algorithm is used to try to reduce the number of memory blocks during the compression process and merge adjacent free blocks into larger free regions. In each thread, for the compressed memory space, an appropriate compression algorithm (such as Run-Length Encoding, RLE) is used to further compress the fragments and optimize the use of storage space. All threads are executed in parallel, and the compressed fragment data will be reintegrated into the global memory space for another fragmentation detection. Finally, the multi-threaded fragment compression data is output, which contains the memory area processed by each thread, the compressed memory size, the remaining fragments and the optimized memory layout. The process of releasing and switching fragment optimization processing first optimizes the fragmentation of the memory as a whole by comprehensively analyzing the memory layout information in the adjacent area fragment merging data and the multi-threaded fragment compression data. First, based on the merged memory area information, the memory scheduling algorithm is used to evaluate which areas need to be further switched and released. For areas that still have fragments after merging, the heuristic algorithm is used to determine the priority and order of fragment release. Then, the optimization results in the multi-threaded parallel compression data are used to determine which memory areas can achieve more efficient memory utilization by adjusting the access frequency and release strategy. Combined with the memory access pattern and fragmentation intensity, the threshold of optimization switching is set, and the release order of memory areas is adjusted through the resource allocator, so as to achieve the purpose of reducing memory fragmentation. Finally, release-switching fragment optimization data is generated, which contains the release status, switching strategy and optimized memory layout information of each optimized area.

[0152] The present invention also provides a file management system for executing the file management method as described above, the file management system comprising:

[0153] The subject level analysis module is used to obtain archive text data; the subject level analysis of different subject contents is performed on the text subject content data to obtain the subject level data of the archive text;

[0154] The memory pointer structure design module is used to process the context perception between different subject content levels of the archive text data according to the archive text subject level data to obtain subject content context level perception data; perform associated expansion word retrieval on the subject content context level perception data to obtain semantic associated expansion word data; perform memory pointer structure design based on the semantic associated expansion word data to obtain associated expansion memory pointer structure data;

[0155] A virtual memory allocation module is used to perform heap space adaptive allocation control according to the associated extended memory pointer structure data to obtain heap space adaptive allocation control data; perform virtual memory temporary allocation scheduling based on the heap space adaptive allocation control data to obtain virtual memory temporary allocation scheduling data;

[0156] The archive management execution module is used to formulate an archive scheduling memory strategy based on the associated extended memory pointer structure data and the virtual memory temporary allocation scheduling data, obtain the archive scheduling memory strategy, and send the archive scheduling memory strategy to the server underlying system to execute archive management.

[0157] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0158] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A file management method, characterized in that: The following steps are involved: Step S1: Acquire archive text data; perform archive text theme level analysis between different theme contents on the text theme content data to obtain archive text theme level data; Step S2: performing context perception processing between different subject content levels on the archive text data according to the archive text subject level data to obtain subject content context level perception data; performing associated extension word retrieval on the subject content context level perception data to obtain semantic associated extension word data; performing memory pointer structure design based on the semantic associated extension word data to obtain associated extension memory pointer structure data; Step S3: performing heap space adaptive allocation control according to the associated extended memory pointer structure data to obtain heap space adaptive allocation control data; Perform virtual memory temporary allocation scheduling based on heap space adaptive allocation control data to obtain virtual memory temporary allocation scheduling data; Step S4: Formulate an archive scheduling memory strategy based on the associated extended memory pointer structure data and the virtual memory temporary allocation scheduling data to obtain the archive scheduling memory strategy, and send the archive scheduling memory strategy to the underlying system of the server to execute archive management; Step S2 includes the following steps: Step S21: performing context perception processing between different subject content levels on the archive text data according to the archive text subject level data to obtain subject content context level perception data; Step S22: performing semantic enhancement segment analysis on the subject content context level perception data to obtain content level semantic key segments; Step S23: performing a search for associated expansion words on the subject content context level perception data according to the content level semantic key fragments to obtain semantic associated expansion word data; Step S24: Designing a memory pointer structure based on the semantically associated extended word data to obtain associated extended memory pointer structure data; Step S24 includes the following steps: Step S241: performing context relevance evaluation on the subject content context level perception data based on the semantically associated extended word data to obtain extended word-subject content relevance data; Step S242: Designing an expansion word chain pointer rule for the semantically associated expansion word data according to the expansion word-subject content correlation data to obtain an expansion word chain pointer design rule; Step S243: matching the pointer allocation synchronization mechanism with the extended word chain pointer design rule to obtain the extended word chain pointer design synchronization rule; Step S244: performing chain parallel partitioning on the extended word chain pointer design synchronization rule to obtain chain parallel partition pointer design data; Step S245: Designing a memory pointer structure according to the chained parallel partition pointer design data to obtain associated extended memory pointer structure data; Designing extension word chain pointer rules for semantically associated extension word data based on extension word-subject content relevance data includes the following steps: Performing extended word context position difference analysis on semantically related extended word data to obtain extended word context position difference data; Based on the extended word context position difference data, the extended word-topic content relevance data is quantified with respect to the content relevance weight, thereby obtaining the extended word-topic content relevance weight quantification data; According to the weight quantification data of the correlation between the extended word and the subject content, a bidirectional linked list pointer structure is constructed for the semantically associated extended word data to obtain an extended word associated linked list pointer structure; Based on the extended word associated linked list pointer structure, a pointer reconfigurable linked list is constructed to obtain a pointer reconfigurable linked list; According to the pointer structure of the extension word association chain list and the pointer reconfigurable chain list, the extension word chain pointer rule is designed for the semantically associated extension word data, and the extension word chain pointer design rule is obtained.

2. The file management method according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: Obtaining archive text data; Step S12: Fill missing values ​​in the archive text data to obtain archive text filling data; Step S13: classifying the archive text filling data by text subject content to obtain text subject content data; Step S14: performing archive text theme level analysis between different theme contents on the text theme content data according to the text theme content data to obtain archive text theme level data.

3. The file management method according to claim 1, characterized in that: Designing synchronization rules for chain pointers of extended words to perform chain parallel partitioning includes the following steps: The pointer traversal retrieval time complexity is simulated for the synchronization rule of the extended word chain pointer design, and the pointer traversal retrieval time complexity is obtained; According to the pointer traversal retrieval time complexity, the synchronization rule of the extended word chain pointer design is matched with the limit chain length to obtain the limit chain length matching data; Based on the extreme chain length matching data and the expansion word-topic content relevance data, the expansion word chain pointer design synchronization rules are used to construct contextual parallel subchains, and the expansion word contextual parallel subchains are obtained; Perform subchain incremental constraint analysis on the parallel subchains in the context of the extended word to obtain subchain incremental constraint data; According to the extended word context parallel sub-chains and sub-chain incremental constraint data, the extended word chain pointer design synchronization rule is chained and partitioned to obtain chained parallel partition pointer design data.

4. The file management method according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: performing concurrent search efficiency simulation on the associated extended memory pointer structure data to obtain pointer structure concurrent search efficiency data; Step S32: performing heap space adaptive allocation control according to the pointer structure concurrent retrieval efficiency data to obtain heap space adaptive allocation control data; Step S33: Perform virtual memory temporary allocation scheduling based on the heap space adaptive allocation control data and the pointer structure concurrent retrieval efficiency data to obtain virtual memory temporary allocation scheduling data.

5. The file management method according to claim 4, characterized in that: Step S32 includes the following steps: Step S321: performing efficiency fluctuation range analysis on the pointer structure concurrent retrieval efficiency data to obtain the concurrent retrieval efficiency fluctuation range; Step S322: extracting the number of search results from the pointer structure concurrent search efficiency data to obtain the number of search results; Step S323: mapping the search result quantity value according to the concurrent search efficiency fluctuation range to obtain efficiency fluctuation-result quantity association mapping data; Step S324: dynamically adjust the heap space according to the efficiency fluctuation-result quantity association mapping data to obtain heap space dynamic adjustment data; Step S325: performing release switching fragment optimization processing on the heap space dynamic adjustment data to obtain release switching fragment optimization data; Step S326: performing heap space adaptive allocation control according to the heap space dynamic adjustment data and the release switching fragment optimization data to obtain heap space adaptive allocation control data.

6. The file management method according to claim 5, characterized in that: The process of dynamically adjusting the data in the heap space to release the switching fragmentation optimization process includes the following steps: According to the efficiency fluctuation-result quantity association mapping data, the heap space dynamic adjustment data is evaluated for the space switching release fragment intensity to obtain the space release fragment intensity data; Marking the high-frequency address area of ​​the spatial release fragmentation intensity data to obtain the fragmentation intensity high-frequency address area; Estimate the fragment release overhead for the high-frequency address area of ​​fragment intensity to obtain fragment release overhead data; Perform adjacent region fragment merging processing on the fragment intensity high frequency address region according to the fragment release overhead data to obtain adjacent region fragment merging data; Perform multi-threaded fragment compression processing on the adjacent region fragment merged data to obtain multi-threaded fragment compressed data; The release switching fragment optimization processing is performed according to the adjacent area fragment merging data and the multi-thread fragment compression data to obtain the release switching fragment optimization data.

7. A file management system, characterized in that: Used to execute the archive management method as claimed in claim 1, the archive management system comprises: The subject level analysis module is used to obtain archive text data; the subject level analysis of different subject contents is performed on the text subject content data to obtain the subject level data of the archive text; The memory pointer structure design module is used to process the context perception between different subject content levels of the archive text data according to the archive text subject level data to obtain subject content context level perception data; perform associated expansion word retrieval on the subject content context level perception data to obtain semantic associated expansion word data; perform memory pointer structure design based on the semantic associated expansion word data to obtain associated expansion memory pointer structure data; A virtual memory allocation module is used to perform heap space adaptive allocation control according to the associated extended memory pointer structure data to obtain heap space adaptive allocation control data; perform virtual memory temporary allocation scheduling based on the heap space adaptive allocation control data to obtain virtual memory temporary allocation scheduling data; The archive management execution module is used to formulate an archive scheduling memory strategy based on the associated extended memory pointer structure data and the virtual memory temporary allocation scheduling data, obtain the archive scheduling memory strategy, and send the archive scheduling memory strategy to the server underlying system to execute archive management.

Images