Incremental storage synchronization method and device

Abstract

The invention provides an incremental storage synchronization method and device. The method comprises the following steps: acquiring modification requests of at least two terminals on the same logic fragment of a text in a target time slice; sorting the modification requests based on the timestamps to obtain a modification request sequence; extracting a modification intention feature of each modification request in the modification request sequence; on the basis of the modification intention features, calculating semantic similarity between each modification request in a modification request sequence and a previous modification request in sequence; whether the semantic similarity is smaller than a preset threshold value or not is judged, if yes, it is determined that the content of the modification request is an incremental copy of a logic fragment, and if not, the content of the two modification requests is combined based on a large language model. According to the method, for the logic fragments of the text, a collaborative generation mechanism for automatically merging / incremental copies is triggered based on semantic similarity, semantic-level conflict resolution is realized by utilizing AI, full-amount storage redundancy is eliminated, and resource waste caused by homogeneous storage is avoided.

Description

Technical Field

[0001] This application relates to the field of document management technology, and in particular to an incremental storage synchronization method and apparatus. Background Technology

[0002] Most current collaborative text editing storage synchronization technologies employ a full-data storage synchronization mechanism: Documents are stored as complete files, and users must upload the entire new version after making changes. Multi-device collaboration relies on Operation Transformation (OT) or Conflict-Free Data Types (CRDT) to achieve basic consistency. Version management uses linear snapshot backups, storing a complete copy of the document for each version. Typical technologies include OT-based collaborative editing engines and Git-based full version control systems.

[0003] The technical challenges of full-scale storage synchronization are manifested in the following aspects: Homogeneous storage wastes resources: modifying a single paragraph still requires storing the entire document, and whole file / binary difference synchronization results in the server storing multiple copies of the same content, causing a lot of redundant storage; Multi-device collaboration conflicts surge: When multiple users are editing, conflict detection is coarse-grained (file-level), and forced overwriting results in the loss of valid content. Real-world testing data shows that when the number of collaborating users exceeds 10, the OT algorithm's conflict resolution failure rate reaches as high as 35%. Version management is chaotic: historical versions are stored independently as complete files, resulting in low efficiency for retrieval and rollback; while the linear storage of full-text copies causes the complexity of the version tree to increase exponentially, and retrieving historical versions still requires a delay of minutes.

[0004] Semantic association missing: Binary differences cannot identify logical dependencies between paragraphs / sentences. Summary of the Invention

[0005] In view of this, this application provides an incremental storage synchronization method and apparatus to solve the above-mentioned technical problems.

[0006] In a first aspect, embodiments of this application provide an incremental storage synchronization method applied to a server, comprising: Obtain modification requests for the same logical segment of text from at least two terminals within the target time slice; The modification requests are sorted based on their timestamps to obtain a sequence of modification requests; Extract the modification intent features of each modification request in the modification request sequence; Based on the modification intent features, the semantic similarity between each modification request in the modification request sequence and its preceding modification requests is calculated sequentially. If the semantic similarity is less than a preset threshold, then the content of the modification request is determined to be an incremental copy of the logical segment; otherwise, the content of the two modification requests is merged based on the large language model.

[0007] In one possible implementation, the method further includes: The text is parsed using a segmenter to analyze its semantic structure. The logical boundaries of the text are identified through multi-weighting, and logical segments carrying semantic tags are generated. The semantic tags are the hash values ​​of the parent node and the semantic fingerprint. Multiple logical segments are published to different blockchain nodes, and each logical segment is stored in 3 copies, including a master copy and 2 verification copies.

[0008] In one possible implementation, the method further includes: encapsulating the logical fragment into a quadruple, including: a version number, a hash value, a semantic vector, and semantic dependencies.

[0009] In one possible implementation, the modification intent feature is a BERT embedding vector.

[0010] In one possible implementation, the semantic similarity between each modification request in the sequence of modification requests and its preceding modification requests is calculated, including: Calculate the cosine values ​​of the BERT embedding vectors for the two modification requests; Semantic similarity is 1 minus the cosine value.

[0011] In one possible implementation, the method further includes: automatically creating a version branch node and updating the global dependency graph after the logical segment is modified, maintaining the historical trajectory of the logical segment through a directed acyclic graph, and forming a version record chain.

[0012] In one possible implementation, the method further includes: Based on the semantic dependencies of the logical segments being accessed by the target terminal, a long short-term memory network is used to predict the associated logical segments that the target terminal may access from all logical segments. The modified incremental data packet of the associated logical fragment is sent to the target terminal so that the target terminal can load the incremental data packet to update the associated logical fragment based on the topological dependency order of the logical fragment and realize incremental rendering.

[0013] Secondly, embodiments of this application provide an incremental storage synchronization device applied to a server, comprising: The acquisition unit is used to acquire modification requests for the same logical segment of text from at least two terminals in the target time slice; The sorting unit is used to sort the modification requests based on timestamps to obtain a sequence of modification requests; The feature extraction unit is used to extract the modification intent features of each modification request in the modification request sequence; The calculation unit is used to calculate the semantic similarity between each modification request in the modification request sequence and its previous modification requests, based on the modification intent features. The update unit is used to determine whether the semantic similarity is less than a preset threshold. If it is, the content of the modification request is determined to be an incremental copy of the logical segment. Otherwise, the content of the two modification requests is merged based on the large language model.

[0014] In one possible implementation, the apparatus further includes: a fragment generation unit, configured to perform semantic structure parsing on the text using a segmenter, identify the logical boundaries of the text through multi-weighting, and generate logical fragments carrying semantic tags; the semantic tags are hash values ​​of the parent node and the semantic fingerprint; and publish multiple logical fragments to different blockchain nodes, with each logical fragment storing 3 copies, including a primary copy and 2 verification copies.

[0015] In one possible implementation, the apparatus further includes an encapsulation unit for encapsulating the logical fragment into a quadruple, including a version number, a hash value, a semantic vector, and a semantic dependency.

[0016] In one possible implementation, the modification intent feature is a BERT embedding vector.

[0017] In one possible implementation, the semantic similarity between each modification request in the sequence of modification requests and its preceding modification requests is calculated, including: Calculate the cosine values ​​of the BERT embedding vectors for the two modification requests; Semantic similarity is 1 minus the cosine value.

[0018] In one possible implementation, the apparatus further includes: a version record modification unit, used to automatically create a version branch node and update the global dependency graph when the logical segment is modified, and maintain the historical trajectory of the logical segment through a directed acyclic graph to form a version record chain.

[0019] In one possible implementation, the apparatus further includes: a sending unit, configured to predict, based on the semantic dependencies of the logical segments being accessed by the target terminal, related logical segments that the target terminal may access from all logical segments using a long short-term memory network; and send incremental data packets of the modified related logical segments to the target terminal, so that the target terminal loads the incremental data packets to update the related logical segments based on the topological dependency order of the logical segments, and implements incremental rendering.

[0020] Thirdly, embodiments of this application provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method of embodiments of this application.

[0021] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer instructions that, when executed by a processor, implement the methods of embodiments of this application.

[0022] Fifthly, embodiments of this application provide a computer program product, including a computer program / instructions, which, when executed by a processor, implement the method of embodiments of this application.

[0023] This application targets logical fragments of text, triggering an automatic merging / incremental copy collaborative generation mechanism based on semantic similarity, and utilizes AI to resolve semantic-level conflicts, eliminating full-scale storage redundancy and avoiding resource waste from homogeneous storage. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the specific embodiments of this application or the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0025] Figure 1 A flowchart of the incremental storage synchronization method provided in the embodiments of this application; Figure 2 A functional structure diagram of the incremental storage synchronization device provided in the embodiments of this application; Figure 3 This is a structural diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0027] Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0028] First, a brief introduction to the design concept of the embodiments of this application will be given.

[0029] Existing cloud storage systems mainly employ two technical solutions: Full file synchronization mechanism: When a user modifies any content, the client needs to upload the entire file to the server, resulting in redundant transmission and storage; Incremental updates based on binary differences (such as the rsync algorithm): generate patch packages by comparing binary differences between files, and resolve conflicts using a full-text coverage mechanism of "last modification priority" or "manual selection", but still cannot identify semantically related segments.

[0030] To address this, this application establishes a fragmented version management protocol to achieve incremental semantic collaboration, eliminate full storage redundancy, and resolve fragment-level conflicts automatically in high-concurrency scenarios. Simultaneously, it reduces network bandwidth and storage resource consumption in multi-terminal collaboration scenarios, enabling automatic merging of fragment-level conflicts and avoiding manual intervention. Furthermore, by constructing a traceable version tree structure, it supports precise semantic rollback, maintaining the integrity of textual logical units.

[0031] The design concept of the technical solution in this application is as follows: 1. Fragment the text to generate multiple logical fragments with unique identifiers; The BERT-Logic segmenter is used to perform semantic structure parsing on the document (such as NLP paragraph segmentation and Markdown heading tree), generating logical fragments (LF) with unique IDs. Each fragment's unique identifier, Fragment_ID, is calculated as a hash of its parent node ID and semantic fingerprint, ensuring its uniqueness. The specific technical solution is as follows: (1) Fragmentation mechanism By identifying textual logical boundaries (such as paragraph transitions and semantic objects) through multiple weights, logical fragments (LFs) with semantic tags are generated. Each LF is encapsulated as a four-tuple object: LF_ID〈Version, Hash, SemanticVector, Dependency〉 Among them, Dependency records semantic dependencies across fragments (such as causal chains and referential associations).

[0032] (2) Versioned storage topology Build a fragment-level version tree instead of file-level snapshots. Each time a new LF is generated, a version branch node is automatically created and the global dependency graph is updated. The fragment history trajectory is maintained through a directed acyclic graph (DAG), forming a chain of version records.

[0033] (3) Distributed storage strategy A multi-chain sharding storage model is adopted, distributing LFs to different blockchain nodes according to their hash values. Each LF stores 3 copies (primary copy + 2 verification copies), and shard-level consistency is guaranteed through the RAFT protocol.

[0034] After the splitting is complete, the engine automatically creates logical fragment tuples based on the semantic unit splitting results; and in terms of storage strategy, incremental synchronization is strictly performed only when the LF that has been changed is triggered.

[0035] 2. AI-driven conflict resolution protocol The engine will extract modification intent features (such as BERT embedding vectors) based on fragment vector clock comparison, then calculate semantic similarity Sim, and decide whether to merge complementary content or generate a smart copy based on the semantic similarity threshold.

[0036] Where Sim = 1 - Cosine(Frag_userA, Frag_userB) (1) Conflict detection layer When multiple users modify the same fragment within a preset time slice, a Fragment Vector Clock (FVC) algorithm is introduced. Each LF carries a vector clock stamp [T1, T2, ..., T]. n (n is the number of terminals), when max(T) m )>min(T n If the semantic vector cosine similarity is less than 0.85, it is defined as a normal incremental copy (and the conflict area to be reviewed is marked); otherwise, it is defined as a triggered conflict and submitted for the next step of merging.

[0037] (2) Semantic Differential Engine Three-level semantic diffing is performed using LLM (GPT-4 Turbo): Automatic merging is performed when the intent score is greater than 0.85; otherwise, an incremental copy Δ-Copy is generated and the conflict domain is marked.

[0038] 3. Cross-terminal real-time rendering optimization layer The Delta Streaming Protocol was designed to solve the problem of fragmented rendering latency; the terminal only receives lightweight data packets with modified LF, and the bandwidth usage can be reduced to 1 / 100 of the traditional OT protocol (20Kbps vs 2Mbps).

[0039] Simultaneously, a dynamic document tree reorganization algorithm is constructed to asynchronously load dependent fragments (concurrent requests ≤ 5 LFs), prioritizing the rendering of placeholders for changed LFs; fragment spatial location is performed based on B+ tree indexes. The key to reorganization latency < 50ms lies in the prefetch prediction model: using LSTM to predict the associated fragments that the user may access next and preloading them into the local cache.

[0040] In summary, by using a differential streaming protocol in the real-time rendering engine, the server only needs to send incremental changes, and the terminal only needs to receive the change LF data packets to obtain the data required for reloading. The client loads the incremental change fragments according to the given topological dependency order, dynamically constructs the Document Object Model (DOM) tree, and finally achieves the cross-terminal minimum incremental rendering process by redrawing only the specified changed areas.

[0041] After introducing the application scenarios and design concepts of the embodiments of this application, the technical solutions provided by the embodiments of this application will be described below.

[0042] like Figure 1 As shown in the figure, this application provides an incremental storage synchronization method applied to a server, including: Step 101: Obtain modification requests for the same logical segment of text from at least two terminals within the target time slice; Step 102: Sort the modification requests based on timestamps to obtain a sequence of modification requests; Step 103: Extract the modification intent features of each modification request in the modification request sequence; Step 104: Based on the modification intent features, calculate the semantic similarity between each modification request in the modification request sequence and its preceding modification requests in turn; Step 105: Determine whether the semantic similarity is less than a preset threshold. If yes, determine that the content of the modification request is an incremental copy of the logical segment. Otherwise, merge the content of the two modification requests based on the large language model.

[0043] For example, the preset threshold is 0.85.

[0044] This embodiment targets logical segments of text, triggering an automatic merging / incremental copy collaborative generation mechanism based on semantic similarity, and using AI to resolve semantic-level conflicts, eliminating full-scale storage redundancy and avoiding the waste of resources from homogeneous storage: modifying a single paragraph still requires storing the entire document, and the synchronization of whole files / binary differences leads to the server storing multiple copies of the same content, resulting in a large amount of redundant storage.

[0045] In some embodiments, the method further includes: The text is parsed using a segmenter to analyze its semantic structure. The logical boundaries of the text are identified through multi-weighting, and logical segments carrying semantic tags are generated. The semantic tags are the hash values ​​of the parent node and the semantic fingerprint. Multiple logical segments are published to different blockchain nodes, and each logical segment is stored in 3 copies, including a master copy and 2 verification copies.

[0046] This embodiment divides text into multiple logical segments by constructing a fragment-level version tree, thereby pinpointing modifications to the text by the terminal to modifications of logical segments, effectively reducing conflicts in multi-terminal collaboration.

[0047] In some embodiments, the method further includes: encapsulating the logical fragment into a quadruple, including: version number Version, hash value Hash, semantic vector SemanticVector, and semantic dependency Dependency.

[0048] In some embodiments, the modification intent feature is a BERT embedding vector.

[0049] In some embodiments, calculating the semantic similarity between each modification request in the modification request sequence and its preceding modification requests includes: Calculate the cosine values ​​of the BERT embedding vectors for the two modification requests; Semantic similarity is 1 minus the cosine value.

[0050] In some embodiments, the method further includes: automatically creating a version branch node and updating the global dependency graph after the logical segment is modified, maintaining the historical trajectory of the logical segment through a directed acyclic graph, and forming a version record chain.

[0051] This embodiment manages the historical versions of logical segments through a multi-branch version tree, which can improve retrieval and rollback efficiency.

[0052] In some embodiments, the method further includes: Based on the semantic dependencies of the logical segments being accessed by the target terminal, a long short-term memory network is used to predict the associated logical segments that the target terminal may access from all logical segments. The modified incremental data packet of the associated logical fragment is sent to the target terminal so that the target terminal can load the incremental data packet to update the associated logical fragment based on the topological dependency order of the logical fragment and realize incremental rendering.

[0053] This embodiment uses a differential stream transmission data packet structure to implement data compression format and reassembly process in cross-terminal real-time rendering, thus solving the terminal rendering latency problem.

[0054] Based on the same inventive concept, this application provides an incremental storage synchronization device, see reference. Figure 2 As shown, the incremental storage synchronization device 200 provided in this application embodiment includes at least: The acquisition unit 201 is used to acquire modification requests for the same logical segment of text from at least two terminals in the target time slice; The sorting unit 202 is used to sort the modification requests based on timestamps to obtain a sequence of modification requests; Feature extraction unit 203 is used to extract the modification intent features of each modification request in the modification request sequence; The calculation unit 204 is used to calculate the semantic similarity between each modification request in the modification request sequence and its previous modification requests, based on the modification intent features. The update unit 205 is used to determine whether the semantic similarity is less than a preset threshold. If it is, the content of the modification request is determined to be an incremental copy of the logical segment. Otherwise, the content of the two modification requests is merged based on the large language model.

[0055] In some embodiments, the apparatus further includes: a fragment generation unit, configured to perform semantic structure parsing on the text using a segmenter, identify the logical boundaries of the text through multi-weighting, and generate logical fragments carrying semantic tags; the semantic tags are hash values ​​of the parent node and the semantic fingerprint; and publish multiple logical fragments to different blockchain nodes, with each logical fragment storing 3 copies, including a primary copy and 2 verification copies.

[0056] In some embodiments, the apparatus further includes: an encapsulation unit for encapsulating the logical fragment into a quadruple, including: a version number (Version), a hash value (Hash), a semantic vector (SemanticVector), and a semantic dependency (Dependency).

[0057] In some embodiments, the modification intent feature is a BERT embedding vector.

[0058] In some embodiments, calculating the semantic similarity between each modification request in the modification request sequence and its preceding modification requests includes: Calculate the cosine values ​​of the BERT embedding vectors for the two modification requests; Semantic similarity is 1 minus the cosine value.

[0059] In some embodiments, the apparatus further includes: a version record modification unit, configured to automatically create version branch nodes and update the global dependency graph when the logical segment is modified, maintain the historical trajectory of the logical segment through a directed acyclic graph, and form a version record chain.

[0060] In some embodiments, the apparatus further includes: a sending unit, configured to predict, based on the logical segment currently being accessed by the target terminal, an associated logical segment that the target terminal may access; and send an incremental data packet of the modified associated logical segment to the target terminal, so that the target terminal loads the incremental data packet to update the associated logical segment based on the topological dependency order of the logical segments, and implements incremental rendering.

[0061] In some embodiments, based on the logical segment that the target terminal is currently accessing, it predicts the associated logical segments that the target terminal may access; including: Based on the semantic dependencies of the logical segments being accessed by the target terminal, a long short-term memory network is used to predict the associated logical segments that the target terminal may access from all logical segments.

[0062] Based on the same inventive concept, embodiments of this application also provide an electronic device, such as... Figure 3 As shown, it includes a memory and a processor. The memory stores an executable program, and the processor executes the executable program to implement the steps of the incremental storage synchronization method provided in the above embodiments.

[0063] The aforementioned processor can be a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The aforementioned PLD can be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof. The general-purpose processor can be a microprocessor or any conventional processor, etc.

[0064] Since the electronic device described in this application embodiment is an electronic device equipped with a memory that implements the incremental storage synchronization method disclosed in this application embodiment, those skilled in the art can understand the structure and variations of the electronic device described in this application embodiment based on the incremental storage synchronization method described in this application embodiment, and therefore will not be described again here.

[0065] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the incremental storage synchronization method provided in the above embodiments.

[0066] The storage medium in this embodiment may be included in an electronic device; or it may exist independently and not be assembled into an electronic device. The storage medium carries one or more computer programs, which, when executed, implement the steps of the incremental storage synchronization method provided in the above embodiment.

[0067] It should be understood that the various solutions in this embodiment have the same technical effects as those in the above method embodiments, and will not be repeated here.

[0068] According to embodiments of this application, the computer-readable storage medium can be a non-volatile computer-readable storage medium, such as including but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. Optionally, specific examples in this embodiment can refer to the examples described in any embodiment of this application, which will not be repeated here. Obviously, those skilled in the art should understand that the various modules or steps of this application described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those presented here, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, this application is not limited to any particular hardware and software combination.

[0069] This application also provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the incremental storage synchronization method provided in the above embodiments.

[0070] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions targeted in the blocks may occur in a different order than those targeted in the drawings. For example, two consecutively represented blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0071] Furthermore, while the operations are described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or in a sequential order. Multitasking and parallel processing may be advantageous in certain environments. Similarly, while several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this application. Certain features described in the context of individual embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments.

Claims

1. An incremental storage synchronization method, applied to a server, characterized in that, include: Obtain modification requests for the same logical segment of text from at least two terminals within the target time slice; The modification requests are sorted based on their timestamps to obtain a sequence of modification requests; Extract the modification intent features of each modification request in the modification request sequence; Based on the modification intent features, the semantic similarity between each modification request in the modification request sequence and its preceding modification requests is calculated sequentially. If the semantic similarity is less than a preset threshold, then the content of the modification request is determined to be an incremental copy of the logical segment; otherwise, the content of the two modification requests is merged based on the large language model.

2. The method according to claim 1, characterized in that, The method further includes: The text is parsed using a segmenter to analyze its semantic structure. The logical boundaries of the text are identified through multi-weighting, and logical segments carrying semantic tags are generated. The semantic tags are the hash values ​​of the parent node and the semantic fingerprint. Multiple logical segments are published to different blockchain nodes, and each logical segment is stored in 3 copies, including a master copy and 2 verification copies.

3. The method according to claim 2, characterized in that, The method further includes: encapsulating the logical fragment into a quadruple, including: version number, hash value, semantic vector, and semantic dependency.

4. The method according to claim 1, characterized in that, The modification intent feature is a BERT embedding vector.

5. The method according to claim 4, characterized in that, Calculate the semantic similarity between each modification request in the modification request sequence and its preceding modification requests, including: Calculate the cosine values ​​of the BERT embedding vectors for the two modification requests; The semantic similarity is 1 minus the cosine value.

6. The method according to claim 1, characterized in that, The method further includes: automatically creating version branch nodes and updating the global dependency graph when the logical segment is modified, maintaining the historical trajectory of the logical segment through a directed acyclic graph, and forming a version record chain.

7. The method according to claim 3, characterized in that, The method further includes: Based on the semantic dependencies of the logical segments being accessed by the target terminal, a long short-term memory network is used to predict the associated logical segments that the target terminal may access from all logical segments. The modified incremental data packet of the associated logical fragment is sent to the target terminal so that the target terminal can load the incremental data packet to update the associated logical fragment based on the topological dependency order of the logical fragment and realize incremental rendering.

8. An incremental storage synchronization device, applied to a server, characterized in that, include: The acquisition unit is used to acquire modification requests for the same logical segment of text from at least two terminals in the target time slice; The sorting unit is used to sort the modification requests based on timestamps to obtain a sequence of modification requests; The feature extraction unit is used to extract the modification intent features of each modification request in the modification request sequence; The calculation unit is used to calculate the semantic similarity between each modification request in the modification request sequence and its previous modification requests, based on the modification intent features. The update unit is used to determine whether the semantic similarity is less than a preset threshold. If it is, the content of the modification request is determined to be an incremental copy of the logical segment. Otherwise, the content of the two modification requests is merged based on the large language model.

9. The apparatus according to claim 8, characterized in that, The device further includes: a fragment generation unit, used to perform semantic structure parsing on the text using a segmenter, identify the logical boundaries of the text through multi-weighting, and generate logical fragments carrying semantic tags; the semantic tags are the hash values ​​of the parent node and the semantic fingerprint; multiple logical fragments are published to different blockchain nodes, and each logical fragment stores 3 copies, including a primary copy and 2 verification copies.

10. The apparatus according to claim 9, characterized in that, The device further includes an encapsulation unit for encapsulating the logical fragment into a quadruple, including a version number, a hash value, a semantic vector, and a semantic dependency.

11. The apparatus according to claim 8, characterized in that, The modification intent feature is a BERT embedding vector.

12. The apparatus according to claim 11, characterized in that, Calculate the semantic similarity between each modification request in the modification request sequence and its preceding modification requests, including: Calculate the cosine value of the modification intent feature of the two modification requests; The semantic similarity is 1 minus the cosine value.

13. The apparatus according to claim 8, characterized in that, The device further includes: a version record modification unit, used to automatically create version branch nodes and update the global dependency graph when the logical segment is modified, and maintain the historical trajectory of the logical segment through a directed acyclic graph to form a version record chain.

14. The apparatus according to claim 10, characterized in that, The device further includes: a sending unit, used to predict the associated logical segments that the target terminal may access from all logical segments based on the semantic dependencies of the logical segments being accessed by the target terminal using a long short-term memory network; and sending incremental data packets of the modified associated logical segments to the target terminal, so that the target terminal loads the incremental data packets to update the associated logical segments based on the topological dependency order of the logical segments, and realizes incremental rendering.

15. An electronic device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the method as claimed in any one of claims 1-7.

16. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the method as described in any one of claims 1-7.

17. A computer program product, characterized in that, Includes a computer program / instruction that, when executed by a processor, implements the method as described in any one of claims 1-7.

Images