Digital office data synchronization and sharing method and system based on cloud computing

Abstract

The invention provides a digital office data synchronization and sharing method and system based on cloud computing, and relates to the technical field of cloud computing, and the method comprises the steps: building a distributed data synchronization system on a cloud computing platform, and detecting the change of digital office data on each terminal device through a real-time monitoring mechanism; when data change is detected, triggering an incremental synchronization algorithm, and only synchronizing the changed data part; historical version information of each data change is recorded by using a block chain-based distributed account book technology, and data version tracking and backtracking are realized; constructing a role-based access control system, and allocating data access permissions to different users; a real-time collaborative editing function is realized, and multiple users are allowed to edit the same digital office data at the same time; establishing a data sharing log system, and recording data access and operation behaviors of each user; and a machine learning algorithm is utilized to analyze a user behavior mode, and related digital office data is automatically recommended.

Description

Technical Field

[0001] The present invention relates to the technical field of cloud computing, and in particular to a method and system for synchronizing and sharing digital office data based on cloud computing. Background Art

[0002] With the rapid development of information technology and the continuous evolution of office methods, digital office has become an important means for modern enterprises to improve efficiency and competitiveness. In a digital office environment, data synchronization and sharing are key factors to ensure collaboration efficiency and information consistency. Traditional data synchronization and sharing methods have many limitations and are difficult to meet the current needs of enterprises for efficient, secure, and flexible data management.

[0003] At present, common data synchronization and sharing methods mainly include the following: Local server storage: Enterprises deploy servers internally and store data in the local network. Although this method can ensure data security, it has problems such as high equipment maintenance costs, poor scalability, and inconvenient remote access. External hard disk or USB flash drive transfer: Data is transferred between different terminals through physical storage devices. This method is simple to operate, but it is easy to cause data version confusion, and there are security risks and low efficiency. Email attachments: Share files by sending email attachments. This method is widely used, but it has limitations for large file transfers and is difficult to perform real-time collaboration and version control. Traditional file sharing services: such as FTP servers or network shared folders. Although these methods can achieve file sharing, they lack advanced collaboration functions and access control mechanisms. Early cloud storage services: Provide basic file upload, download, and synchronization functions, but lack advanced features for enterprise-level needs, such as fine-grained permission control, audit logs, collaborative editing, etc.

[0004] These traditional methods have the following common problems in data synchronization and sharing: Data consistency is difficult to ensure: In a multi-user, multi-device environment, data version conflicts and inconsistencies are prone to occur. Insufficient security: The lack of a complete encryption mechanism and access control strategy can easily lead to data leakage or unauthorized access. Low collaboration efficiency: Insufficient real-time collaboration functions make it difficult to support multiple people editing the same document at the same time. Poor scalability: It is difficult to quickly expand storage capacity and computing resources according to business needs. Insufficient support for mobile office: It cannot meet the needs of accessing and editing documents anytime, anywhere. Low intelligence: The lack of advanced functions such as intelligent search and data analysis makes it difficult to fully tap the value of data. Low integration: Insufficient integration with other office systems, resulting in information islands.

[0005] Therefore, there is an urgent need for a new digital office data synchronization and sharing solution based on cloud computing technology to solve the above problems and improve the company's data management efficiency and collaboration level. Summary of the invention

[0006] The embodiments of the present invention provide a digital office data synchronization and sharing method and system based on cloud computing, which can solve the problems in the prior art.

[0007] According to a first aspect of the embodiments of the present invention, Provides a cloud computing-based digital office data synchronization and sharing method, including: Deploy a distributed data collection module with adaptive collection frequency to collect digital office data in real time from multiple heterogeneous terminal devices, including structured data, semi-structured data and unstructured data; use deep learning algorithms to intelligently classify and annotate the collected digital office data; upload the digital office data to the distributed cloud storage platform through a load balancing algorithm, and generate a unique identifier based on spatiotemporal characteristics and metadata information containing data lineage for each digital office data in the cloud storage platform; establish a data relationship index based on a graph database to achieve efficient data retrieval and association analysis; A distributed data synchronization system is established on the cloud computing platform to detect changes in digital office data on each terminal device through a real-time monitoring mechanism. When data changes are detected, an incremental synchronization algorithm is triggered to synchronize only the changed data to reduce the amount of data transmission. The blockchain-based distributed ledger technology is used to record the historical version information of each data change to achieve data version tracking and backtracking. Different synchronization strategies are adopted for different types of digital office data, using character-level difference comparison for text files and pixel-block-level difference comparison for image files. Build a role-based access control system to assign data access rights to different users; implement real-time collaborative editing capabilities to allow multiple users to edit the same digital office data at the same time, and resolve concurrent editing conflicts through operation conversion algorithms; establish a data sharing log system to record each user's data access and operation behavior; use machine learning algorithms to analyze user behavior patterns and automatically recommend relevant digital office data.

[0008] In an optional embodiment, The digital office data is uploaded to the distributed cloud storage platform through a load balancing algorithm, and a unique identifier based on spatiotemporal characteristics and metadata containing data lineage are generated for each digital office data in the cloud storage platform, including: The digital office data of multiple terminal devices are collected through a distributed data collection module; the digital office data is preprocessed, including data format standardization, redundant data elimination, compression and encryption; the preprocessed digital office data is uploaded to the distributed cloud storage platform using a dynamic weight-based load balancing algorithm, and the dynamic weight-based load balancing algorithm includes: real-time collection of the CPU usage, memory usage, network throughput and current number of connections of each upload node; using a weighted summation method to calculate the comprehensive performance score of each upload node; based on the comprehensive performance score, using a Softmax function to calculate the dynamic weight of each upload node; using a weighted polling algorithm to select a target upload node according to the dynamic weight; A Ceph-based distributed storage architecture is constructed in the distributed cloud storage platform, and the distributed storage architecture includes an object storage device, a monitor, a metadata server, and a RADOS gateway; a Redis-based distributed cache layer is introduced between the object storage device and the client; a data sharding strategy is dynamically adjusted according to data access patterns and relevance; and an asynchronous replication mechanism is used to replicate data; Generate a unique identifier based on spatiotemporal characteristics for each digital office data, the generation process of the unique identifier includes: generating a microsecond timestamp; converting the geographic location information of the data source into a Geohash code; obtaining a hash value of a device unique identifier; generating a random number; and concatenating the timestamp, the Geohash code, the hash value of the device unique identifier and the random number to generate a final unique identifier through a SHA-256 algorithm; Construct a metadata model based on a graph structure to represent data lineage relationships. Treat each data object as a node in the graph, and the relationship between nodes is represented by directed edges. Create a new node and record basic properties when data is created. Create a new node and add a derived relationship edge pointing to the original data when data is modified. Create a new node and add a merge relationship edge pointing to the source data when data is merged. Add a reference relationship edge between the referencing data and the referenced data when data is referenced.

[0009] In an optional embodiment, The real-time monitoring mechanism detects changes in digital office data on each terminal device. When data changes are detected, the incremental synchronization algorithm is triggered, and only the changed data is synchronized, including: Detect changes in digital office data on each terminal device in real time through the file system event monitoring mechanism; perform periodic scans to detect changes missed by the file system event monitoring mechanism by comparing file size, modification time, and MD5 hash value; predict the probability of file changes using an XGBoost-based machine learning model, which is trained based on the features of file type, file size, last modification time, whether it is a weekday, the hour of the day, and the user ID; The detected change information and predicted change probability are stored in a thread-safe change queue; when data changes are detected, the incremental synchronization process is triggered, including: Divide the changed file into fixed-size data blocks; calculate the weak checksum and strong checksum of each data block; send the data block signature to the receiver; the receiver compares the received signature with the local file signature and identifies the unmatched data blocks; the receiver requests the sender to transmit the unmatched data blocks; the sender transmits only the requested data blocks; the receiver reconstructs the file using the local existing data blocks and the newly received data blocks; The real-time monitoring mechanism and incremental synchronization algorithm are integrated through multi-threading. One thread is responsible for file system monitoring, another thread is responsible for periodic scanning, and another thread is responsible for obtaining change information from the change queue and performing incremental synchronization.

[0010] In an optional embodiment, Different synchronization strategies are adopted for different types of digital office data. Character-level difference comparison is used for text files, and pixel-block-level difference comparison is used for image files, including: Receiving digital office data to be synchronized; identifying the type of the digital office data, and classifying the digital office data into text files, image files, or other types of files; For digitized office data identified as text files, do the following: Convert the content of the text file into a string; use the Myers difference algorithm to perform character-level difference comparison on the string to generate an editing script; compress the editing script to obtain a compressed editing script; For digitized office data recognized as image files, do the following: Dividing the image file into a plurality of pixel blocks of fixed size; calculating a hash value of each pixel block; comparing the hash values ​​of corresponding pixel blocks in the new and old image files to determine changed pixel blocks; performing JPEG encoding on the changed pixel blocks; and compressing the encoded pixel block data to obtain compressed pixel block data; For digitized office data identified as other types of files, do the following: Reading all the contents of the other types of files; compressing all the contents to obtain compressed file contents; The compressed editing script, the compressed pixel block data or the compressed file content are encapsulated together with the corresponding file type information to form a synchronization data packet; and the synchronization data packet is transmitted to a receiving end.

[0011] In an optional embodiment, The method further comprises: At the receiving end, a corresponding decompression and reconstruction method is selected according to the file type information: for a text file, the compressed editing script is decompressed, and the editing script is applied to reconstruct the text content; for an image file, the compressed pixel block data is decompressed, the pixel block is decoded, and the decoded pixel block is updated to the original image; for other types of files, the compressed file content is decompressed and directly written into a new file; The reconstructed files are saved to the designated location of the receiving end to complete the differential synchronization of digital office data.

[0012] In an optional embodiment, Build a role-based access control system to assign data access rights to different users; implement real-time collaborative editing functions, allowing multiple users to edit the same digital office data at the same time, and resolve concurrent editing conflicts through operation conversion algorithms, including: Define roles and permissions in the system, where each role contains a set of permissions, and each permission corresponds to a specific type of operation; create a user management module, which is used to create user accounts, verify user identities, and assign roles to users; Implement resource access control logic, and control users' access to and operations on digital office data according to the permissions of their roles; establish collaborative editing sessions, allowing multiple users to join the same session at the same time and jointly edit the same digital office data; define basic operation types, including insert and delete operations, each of which contains the operation type, operation location, and operation content; Implementing an operation conversion function that receives two concurrent operations as input and outputs a converted operation to resolve concurrent editing conflicts; In a collaborative editing session, when an editing operation is received from a user, the following sub-steps are performed: Convert the received editing operation to all existing operations in the session one by one; if the converted operation is not empty, apply the converted operation to the current document content; add the converted operation to the operation history of the session; broadcast the converted operation to other users in the session; On the client side, when an operation broadcast is received from the server, the following sub-steps are performed: Convert the received operation to the locally unsynchronized operation; if the converted operation is not empty, apply the converted operation to the local document content; update the client's local view to display the latest document content; regularly synchronize the document status between the server and the client to ensure that the document content of all clients is consistent; when the user exits the collaborative editing session, save the final document content and update the document access history.

[0013] According to a second aspect of the embodiments of the present invention, Provides a digital office data synchronization and sharing system based on cloud computing, including: The first unit is used to deploy a distributed data collection module with an adaptive collection frequency, collect digital office data in real time from multiple heterogeneous terminal devices, and the digital office data includes structured data, semi-structured data and unstructured data; use a deep learning algorithm to intelligently classify and annotate the collected digital office data; upload the digital office data to a distributed cloud storage platform through a load balancing algorithm, and generate a unique identifier based on spatiotemporal features and metadata information containing data lineage for each digital office data in the cloud storage platform; establish a data relationship index based on a graph database to achieve efficient data retrieval and association analysis; The second unit is used to establish a distributed data synchronization system on the cloud computing platform, and detect changes in digital office data on each terminal device through a real-time monitoring mechanism; when data changes are detected, the incremental synchronization algorithm is triggered to synchronize only the changed data parts to reduce the amount of data transmission; the historical version information of each data change is recorded using blockchain-based distributed ledger technology to achieve data version tracking and backtracking; for different types of digital office data, differentiated synchronization strategies are adopted, using character-level difference comparison for text files and pixel-block-level difference comparison for image files; The third unit is used to build a role-based access control system to assign data access rights to different users; implement real-time collaborative editing functions, allowing multiple users to edit the same digital office data at the same time, and resolve concurrent editing conflicts through operation conversion algorithms; establish a data sharing log system to record each user's data access and operation behavior; use machine learning algorithms to analyze user behavior patterns and automatically recommend related digital office data.

[0014] According to a third aspect of the embodiments of the present invention, An electronic device is provided, comprising: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0015] According to a fourth aspect of the embodiments of the present invention, A computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the aforementioned method is implemented.

[0016] The present invention realizes efficient real-time collection of digital office data from multiple heterogeneous terminal devices by deploying a distributed data collection module with an adaptive collection frequency. The deep learning algorithm is used for intelligent classification and labeling, which improves the structured degree and availability of data. The unique identifier based on spatiotemporal characteristics and metadata information of data lineage, combined with the relational index of the graph database, greatly improves the efficiency of data retrieval and association analysis. The incremental synchronization algorithm and differentiated synchronization strategy significantly reduce the amount of data transmission and improve synchronization efficiency. The comprehensive application of these technologies makes the management of digital office data more efficient, accurate and intelligent.

[0017] The present invention constructs a role-based access control system, realizes fine-grained data access rights management, and effectively prevents unauthorized data access. The distributed ledger technology based on blockchain is used to record the data change history, ensuring the traceability and non-tamperability of the data version. The establishment of a data sharing log system enables each user's data access and operation behavior to be recorded and audited. These measures jointly build a secure, transparent, and traceable data management environment, greatly improving the security and reliability of digital office data.

[0018] The real-time collaborative editing function and operation conversion algorithm implemented by the present invention enable multiple users to edit the same digital office data at the same time, effectively resolve concurrent editing conflicts, and greatly improve team collaboration efficiency. The function of machine learning algorithms to analyze user behavior patterns and automatically recommend relevant data not only optimizes the user experience, but also promotes the full utilization and value mining of data. The application of distributed cloud storage platforms ensures high availability and scalability of data, meeting the growing data storage and processing needs of enterprises. The combination of these innovative functions and technologies has significantly improved the collaborative experience in digital office environments and provided technical support for deep mining of data value. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A schematic diagram of a process of synchronizing and sharing digital office data based on cloud computing according to an embodiment of the present invention; Figure 2 It is a structural diagram of a digital office data synchronization and sharing system based on cloud computing according to an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The technical solution of the present invention is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0021] Figure 1 FIG. 1 is a flow chart of a method for synchronizing and sharing digital office data based on cloud computing according to an embodiment of the present invention. Figure 1 As shown, the method includes: S101. Deploy a distributed data collection module with an adaptive collection frequency to collect digital office data in real time from multiple heterogeneous terminal devices, wherein the digital office data includes structured data, semi-structured data, and unstructured data; use a deep learning algorithm to intelligently classify and annotate the collected digital office data; upload the digital office data to a distributed cloud storage platform through a load balancing algorithm, and generate a unique identifier based on spatiotemporal features and metadata information containing data lineage for each digital office data in the cloud storage platform; establish a data relationship index based on a graph database to achieve efficient data retrieval and association analysis; S102. Establish a distributed data synchronization system on the cloud computing platform, and detect changes in digital office data on each terminal device through a real-time monitoring mechanism; when data changes are detected, trigger an incremental synchronization algorithm to synchronize only the changed data to reduce the amount of data transmission; use blockchain-based distributed ledger technology to record the historical version information of each data change to achieve data version tracking and backtracking; use differentiated synchronization strategies for different types of digital office data, using character-level difference comparison for text files and pixel-block-level difference comparison for image files; S103. Build a role-based access control system to assign data access rights to different users; implement real-time collaborative editing capabilities to allow multiple users to edit the same digital office data at the same time, and resolve concurrent editing conflicts through operation conversion algorithms; establish a data sharing log system to record each user's data access and operation behavior; use machine learning algorithms to analyze user behavior patterns and automatically recommend relevant digital office data.

[0022] The specific implementation of the digital office data synchronization and sharing method based on cloud computing is as follows: First, deploy a distributed data acquisition module with adaptive acquisition frequency. This module includes data acquisition clients installed on different terminal devices (such as PCs, mobile phones, tablets, etc.) and data acquisition servers deployed in the cloud. The data acquisition client can monitor the changes of digital office data on terminal devices in real time, including documents, pictures, audio and video, etc. The acquisition frequency will be automatically adjusted according to the frequency of data changes. For frequently changing data (such as documents being edited), high-frequency acquisition is used, and for less changing data (such as archived files), low-frequency acquisition is used to balance the real-time acquisition and system resource consumption.

[0023] The data collection client transmits the collected data to the data collection server in the cloud. The collection server receives data from various terminals, performs preliminary data cleaning and format conversion, and unifies heterogeneous data from different sources into a standard format. For example, Office files such as Word documents, Excel tables, and PPT presentations are uniformly converted into XML format, and image files such as JPG and PNG are uniformly converted into lossless compressed TIFF format.

[0024] Next, a deep learning algorithm is used to intelligently classify and annotate the collected digital office data. Specifically, a multi-label classification model based on a convolutional neural network (CNN) is used. The model is trained with a large amount of annotated office data and can automatically identify data types, topics, keywords and other features. For example, for a sales report document, the model can automatically annotate "sales", "report", "monthly", "data analysis" and other labels. For a product design drawing, labels such as "product", "design", "3D model", and "industrial design" can be annotated. The results of these intelligent annotations will be stored as metadata together with the original data.

[0025] Then, the processed digital office data is uploaded to the distributed cloud storage platform through the load balancing algorithm. The consistent hashing algorithm is used to store the data in multiple storage nodes to ensure the high availability and scalability of data access. In the cloud storage platform, a unique identifier based on spatiotemporal characteristics is generated for each digital office data, including the data creation time, location, device ID and other information. At the same time, metadata information containing data lineage is generated to record the source, processing process, version history, etc. of the data.

[0026] Based on the stored data and metadata, a data relationship index based on the graph database is established. Using the Neo4j graph database, each digital office data is used as a node in the graph, and the relationship between the data is used as an edge. Through the graph traversal algorithm, related data can be quickly found to achieve efficient data retrieval and association analysis. For example, all historical versions related to a sales report, the referenced data sources, and other reports generated using the report can be quickly found.

[0027] A distributed data synchronization system is established on the cloud computing platform, including synchronization clients deployed on each terminal and synchronization servers in the cloud. The synchronization client detects changes in local digital office data in real time through file system monitoring (such as the inotify mechanism). When data changes are detected, the hash value of the changed data is calculated and compared with the hash value of the previous version of the data stored in the cloud to determine whether synchronization is required.

[0028] Trigger the incremental synchronization algorithm to synchronize only the changed data. For text files, use a character-level difference comparison algorithm (such as the Myers difference algorithm) to calculate the specific characters and positions of the changes. For binary files (such as images, audio and video), use a rolling hash algorithm to divide the files into blocks and only transfer the changed data blocks. This incremental synchronization method can significantly reduce the amount of data transmission and improve synchronization efficiency.

[0029] For example, for a 100KB text document, if only 10 characters are modified, the traditional method requires retransmitting the entire file, while incremental synchronization only requires transmitting the 10 changed characters and their location information, and the data volume may be only about 100 bytes.

[0030] The distributed ledger technology based on blockchain is used to record the historical version information of each data change. Each data change is written into the blockchain as a transaction record, including the change time, the operating user, the hash value of the change content, etc. The tamper-proof nature of the blockchain ensures the credibility of the data change history and enables data version tracking and backtracking.

[0031] Different synchronization strategies are adopted for different types of digital office data: - For text files (such as documents, codes, etc.), character-level difference comparison is used. The longest common subsequence (LCS) algorithm is used to calculate the differences of files and only the changed characters are synchronized.

[0032] - For image files, pixel block level difference comparison is used. The image is divided into multiple pixel blocks, and the characteristic value (such as average color) of each pixel block is calculated, and only the pixel blocks that have changed are synchronized.

[0033] - For audio and video files, use frame-level difference comparison. Compare the differences of key frames and only synchronize the changed frame data.

[0034] Build a role-based access control system (RBAC) to assign data access permissions to different users. Define multiple roles (such as administrator, ordinary user, guest, etc.), each with different data operation permissions. Users are assigned to the corresponding role and inherit the permissions of the role. For example, ordinary users can read and edit documents within a specified range, but cannot delete them; while administrators have all operation permissions.

[0035] Real-time collaborative editing is achieved, allowing multiple users to edit the same digital office data at the same time. The operation conversion algorithm (OT) is used to resolve concurrent editing conflicts. When multiple users edit the same document at the same time, each user's editing operation will be sent to the central server. The server converts each operation according to the order in which the operations arrive, ensuring that all users ultimately see consistent document content.

[0036] For example, user A inserts "Hello" at the beginning of the document, and user B inserts "World" at the end of the document at the same time. After receiving these two operations, the server applies them to the document and broadcasts them to all users, and eventually all users will see "HelloWorld".

[0037] Establish a data sharing log system to record each user's data access and operation behavior. The log contains information such as user ID, operation time, operation type (such as viewing, editing, downloading, etc.), operation object, etc. These log data are stored in a distributed log system (such as ELK Stack) for subsequent analysis and auditing.

[0038] Finally, we use machine learning algorithms to analyze user behavior patterns and automatically recommend relevant digital office data. We use collaborative filtering algorithms to calculate the similarity between users based on their historical operation records and recommend data commonly used by similar users. At the same time, we use content-based recommendation algorithms to recommend other data on related topics based on the data content of the user's current operation. For example, when a user is viewing a sales report for a product, the system may recommend related documents such as inventory reports and customer feedback for the product.

[0039] Through the above steps, efficient synchronization and intelligent sharing of digital office data based on cloud computing are achieved, improving office efficiency and data utilization.

[0040] In an optional implementation, the digital office data is uploaded to the distributed cloud storage platform through a load balancing algorithm, and a unique identifier based on spatiotemporal characteristics and metadata information containing data lineage are generated for each digital office data in the cloud storage platform, including: The digital office data of multiple terminal devices are collected through a distributed data collection module; the digital office data is preprocessed, including data format standardization, redundant data elimination, compression and encryption; the preprocessed digital office data is uploaded to the distributed cloud storage platform using a dynamic weight-based load balancing algorithm, and the dynamic weight-based load balancing algorithm includes: real-time collection of the CPU usage, memory usage, network throughput and current number of connections of each upload node; using a weighted summation method to calculate the comprehensive performance score of each upload node; based on the comprehensive performance score, using a Softmax function to calculate the dynamic weight of each upload node; using a weighted polling algorithm to select a target upload node according to the dynamic weight; A Ceph-based distributed storage architecture is constructed in the distributed cloud storage platform, and the distributed storage architecture includes an object storage device, a monitor, a metadata server, and a RADOS gateway; a Redis-based distributed cache layer is introduced between the object storage device and the client; a data sharding strategy is dynamically adjusted according to data access patterns and relevance; and an asynchronous replication mechanism is used to replicate data; Generate a unique identifier based on spatiotemporal characteristics for each digital office data, the generation process of the unique identifier includes: generating a microsecond timestamp; converting the geographic location information of the data source into a Geohash code; obtaining a hash value of a device unique identifier; generating a random number; and concatenating the timestamp, the Geohash code, the hash value of the device unique identifier and the random number to generate a final unique identifier through a SHA-256 algorithm; Construct a metadata model based on a graph structure to represent data lineage relationships. Treat each data object as a node in the graph, and the relationship between nodes is represented by directed edges. Create a new node and record basic properties when data is created. Create a new node and add a derived relationship edge pointing to the original data when data is modified. Create a new node and add a merge relationship edge pointing to the source data when data is merged. Add a reference relationship edge between the referencing data and the referenced data when data is referenced.

[0041] In a specific implementation, digital office data is first collected from multiple terminal devices through a distributed data collection module. These terminal devices may include personal computers, mobile devices, smart office devices, etc. The collected data types include but are not limited to documents, tables, pictures, audio and video, etc. The data collection module adopts a distributed architecture, deploys a collection agent on each terminal device, and collects the generated office data in real time.

[0042] Preprocess the collected digital office data. The preprocessing steps include data format standardization, redundant data elimination, compression and encryption. Data format standardization converts heterogeneous data from different sources into a standard format, such as converting different versions of Word documents into docx format. Redundant data elimination identifies and deletes duplicate data through file fingerprint comparison and other methods. Data compression uses compression algorithms suitable for different data types, such as GZIP algorithm for text data and WebP format for pictures. Data encryption uses AES-256 encryption algorithm, and the key is securely distributed through the key management system.

[0043] The preprocessed data is uploaded to the distributed cloud storage platform using a dynamic weight-based load balancing algorithm. The algorithm first collects the performance indicators of each upload node in real time, including CPU usage, memory occupancy, network throughput, and current number of connections. Taking a certain upload node as an example, the collected indicators may be: CPU usage 30%, memory occupancy 50%, network throughput 100Mbps, and current number of connections 50. Then the weighted summation method is used to calculate the comprehensive performance score of each node. Continuing with this node as an example, assuming that the weights of each indicator are 0.3, 0.3, 0.2, and 0.2 respectively, the comprehensive score of the node is: 30*0.3 + 50*0.3 + 100*0.2 + 50*0.2 = 55. Next, the Softmax function is used to convert the score of each node into a dynamic weight between 0 and 1. Finally, a weighted polling algorithm is used to select the target upload node according to the dynamic weight of each node. The higher the weight, the greater the probability of the node being selected, thereby achieving load balancing.

[0044] Construct a distributed storage architecture based on Ceph in a distributed cloud storage platform. The architecture includes object storage devices (OSD), monitors, metadata servers (MDS), and RADOS gateways (RGW). OSD is responsible for actual data storage, Monitor maintains cluster status, MDS manages file system metadata, and RGW provides object storage interfaces. A distributed cache layer based on Redis is introduced between OSD and clients to improve the access performance of hot data. The data sharding strategy is dynamically adjusted according to data access patterns and relevance, such as storing data that is frequently accessed together in the same shard. An asynchronous replication mechanism is used for data replication to improve write performance while ensuring data reliability.

[0045] Generate a unique identifier based on spatiotemporal characteristics for each digital office data. The generation process includes the following steps: First, generate a microsecond timestamp, such as "1620123456789123". Then convert the geographic location information of the data source into Geohash code, such as the Geohash code of a certain location in Beijing is "wx4g0ec". Then obtain the hash value of the device's unique identifier (such as MAC address), assuming it is "5a7f6c". Generate a random number, such as "9876". Finally, concatenate these elements to get "1620123456789123wx4g0ec5a7f6c9876", and generate the final 64-bit hexadecimal unique identifier through the SHA-256 algorithm.

[0046] Build a metadata model based on a graph structure to represent data lineage. Treat each data object as a node in the graph, and the relationship between nodes is represented by directed edges. When data is created, create a new node and record basic attributes, such as creation time, creator, file type, etc. When data is modified, create a new node to represent the modified version, and add a derived relationship edge from the original data to the new version. When data is merged, create a new node to represent the merge result, and add a merge relationship edge from the source data to the merge result. When one data references another data, add a reference relationship edge between the referencing data and the referenced data.

[0047] Through the above steps, the efficient collection, preprocessing, load-balanced uploading, distributed storage, unique identification and lineage tracking of digital office data are realized, laying the foundation for the full life cycle management of data. This method can significantly improve the efficiency and reliability of data management in large-scale office environments, and at the same time provide strong support for data traceability, version control and collaborative office.

[0048] In an optional implementation, a real-time monitoring mechanism is used to detect changes in digital office data on each terminal device; when a data change is detected, an incremental synchronization algorithm is triggered, and only the changed data is synchronized, including: Detect changes in digital office data on each terminal device in real time through the file system event monitoring mechanism; perform periodic scans to detect changes missed by the file system event monitoring mechanism by comparing file size, modification time, and MD5 hash value; predict the probability of file changes using an XGBoost-based machine learning model, which is trained based on the features of file type, file size, last modification time, whether it is a weekday, the hour of the day, and the user ID; The detected change information and predicted change probability are stored in a thread-safe change queue; when data changes are detected, the incremental synchronization process is triggered, including: Divide the changed file into fixed-size data blocks; calculate the weak checksum and strong checksum of each data block; send the data block signature to the receiver; the receiver compares the received signature with the local file signature and identifies the unmatched data blocks; the receiver requests the sender to transmit the unmatched data blocks; the sender transmits only the requested data blocks; the receiver reconstructs the file using the local existing data blocks and the newly received data blocks; The real-time monitoring mechanism and incremental synchronization algorithm are integrated through multi-threading. One thread is responsible for file system monitoring, another thread is responsible for periodic scanning, and another thread is responsible for obtaining change information from the change queue and performing incremental synchronization.

[0049] This embodiment provides a method for real-time monitoring and incremental synchronization of digital office data. The method detects changes in data on terminal devices through multiple mechanisms and uses an incremental synchronization algorithm to efficiently synchronize the changed data.

[0050] First, deploy a file system event monitoring mechanism on each terminal device. This mechanism can capture file creation, modification, deletion and other operations in real time to detect data changes. In specific implementation, you can use the API provided by the operating system, such as ReadDirectoryChangesW of Windows or inotify of Linux. The monitoring program will register the file system events of interest and trigger the callback function immediately when the event occurs. In the callback function, record the path of the changed file, the change type and other information, and add it to the change queue.

[0051] In order to make up for the possible vulnerabilities of file system event monitoring, periodic scanning is also required. The scanner regularly traverses the specified directory and compares the size, modification time and MD5 hash value of the file. For example, it scans once every 5 minutes and compares the difference between the current file attributes and the last scan result. If changes are found, the change information is also added to the change queue. The calculation of the MD5 hash value can be done in blocks to avoid full hashing of large files.

[0052] In addition, this method also introduces a machine learning model based on XGBoost to predict the probability of file changes. The model uses file type, file size, last modification time, whether it is a weekday, the hour of the day, and the user ID as features. The training data can be obtained from historical synchronization records. The model regularly scores files and considers files above the threshold as likely to have changed and adds them to the change queue for verification. This prediction mechanism helps to proactively discover potential data changes.

[0053] All detected change information and predicted change probabilities are stored in a thread-safe change queue. The queue uses a read-write lock mechanism to ensure the security of multi-threaded access. Each element in the queue contains fields such as file path, change type, timestamp, and predicted probability.

[0054] When there is data in the change queue, the incremental synchronization process is triggered. First, the change file is divided into fixed-size data blocks, usually 32KB per block. Then the weak checksum and strong checksum of each data block are calculated. The weak checksum can use a 32-bit rolling checksum algorithm with fast calculation speed; the strong checksum uses algorithms such as MD5 to ensure accuracy.

[0055] Next, the signature information of the data block is sent to the receiver. The signature includes the block's serial number, weak checksum, and strong checksum. The receiver compares the received signature with the corresponding signature of the local file. If a mismatch is found, the serial number of the data block is recorded.

[0056] After the comparison is completed, the receiver requests the sender to transmit the unmatched data blocks. The request contains the sequence numbers of all the data blocks that need to be transmitted. The sender only transmits the contents of these specific data blocks according to the request.

[0057] Finally, the receiver reconstructs the file using the local existing data blocks and the newly received data blocks. During reconstruction, the data blocks are spliced ​​in the order of their serial numbers to ensure the integrity and consistency of the file.

[0058] To improve performance, this method integrates various components in a multi-threaded manner. One thread is responsible for file system event monitoring, running continuously and responding to file system events. Another thread performs periodic scanning and traverses the file system at predetermined time intervals. Another thread is responsible for obtaining change information from the change queue and executing the incremental synchronization algorithm. These threads communicate and collaborate through a shared change queue.

[0059] In actual applications, the number of threads and parameters can be adjusted according to the system's hardware resources and performance requirements. For example, on a multi-core processor, the number of scanning threads can be increased, or the frequency and range of scanning can be adjusted. At the same time, a priority mechanism can be set to ensure that changes to important files are synchronized first.

[0060] Through the above method, real-time monitoring and efficient synchronization of digital office data can be achieved. This method makes comprehensive use of multiple detection mechanisms to capture data changes to the maximum extent. The incremental synchronization algorithm ensures that only necessary data is transmitted, greatly reducing the network bandwidth usage. This method is particularly suitable for large-scale distributed office environments, and can maintain data consistency between multiple terminals and improve collaboration efficiency.

[0061] In an optional implementation, for different types of digital office data, a differentiated synchronization strategy is adopted, character-level difference comparison is adopted for text files, and pixel-block-level difference comparison is adopted for image files, including: Receiving digital office data to be synchronized; identifying the type of the digital office data, and classifying the digital office data into text files, image files, or other types of files; For digitized office data identified as text files, do the following: Convert the content of the text file into a string; use the Myers difference algorithm to perform character-level difference comparison on the string to generate an editing script; compress the editing script to obtain a compressed editing script; For digitized office data recognized as image files, do the following: Dividing the image file into a plurality of pixel blocks of fixed size; calculating a hash value of each pixel block; comparing the hash values ​​of corresponding pixel blocks in the new and old image files to determine changed pixel blocks; performing JPEG encoding on the changed pixel blocks; and compressing the encoded pixel block data to obtain compressed pixel block data; For digitized office data identified as other types of files, do the following: Reading all the contents of the other types of files; compressing all the contents to obtain compressed file contents; The compressed editing script, the compressed pixel block data or the compressed file content are encapsulated together with the corresponding file type information to form a synchronization data packet; and the synchronization data packet is transmitted to a receiving end.

[0062] In a specific implementation, for different types of digital office data, a differentiated synchronization strategy is adopted, and a character-level difference comparison is adopted for text files, and a pixel-block-level difference comparison is adopted for image files as follows: First, the digital office data to be synchronized is received, and the data type is identified through the file extension or file header information, and the data is classified into text files, image files or other types of files.

[0063] For data identified as text files, do the following: Read the contents of a text file and convert it to a string. For example, convert "Hello World" to the string "Hello World". Use the Myers difference algorithm to compare the character-level differences of the string. The specific steps are: build an edit graph for the string, find the shortest edit path in the edit graph, and generate an edit script based on the shortest path. Taking "HelloWorld" and "Hello Java" as examples, the generated edit script may be "keep 5 characters, delete 5 characters, and insert 4 characters". Compress the edit script and use compression algorithms such as run-length encoding to obtain a compressed edit script, such as "R5D5I4".

[0064] For data identified as image files, do the following: Divide the image file into multiple pixel blocks of fixed size, such as 32x32 pixels. Calculate the hash value of each pixel block using the perceptual hash algorithm. Compare the hash values ​​of the corresponding pixel blocks in the new and old image files to determine the changed pixel blocks. Perform JPEG encoding on the changed pixel blocks and set the appropriate compression quality. Compress the encoded pixel block data using compression algorithms such as DEFLATE to obtain compressed pixel block data.

[0065] For data identified as other types of files, do the following: Read the entire content of the file. Compress the entire content and use a general compression algorithm such as DEFLATE to obtain the compressed file content.

[0066] Finally, the compressed editing script, compressed pixel block data or compressed file content are encapsulated together with the corresponding file type information to form a synchronization data packet. The file type information includes the file name, file type identifier, etc. The synchronization data packet is transmitted to the receiving end through the network.

[0067] At the receiving end, the synchronization data packet is parsed according to the file type information to restore the corresponding file content. For text files, the original file is modified according to the editing script. For image files, the changed pixel blocks are replaced with the original image. For other types of files, the original file is directly replaced with the new file content.

[0068] By adopting differentiated synchronization strategies, the amount of data that needs to be transmitted can be greatly reduced, and the synchronization efficiency can be improved. At the same time, the synchronization accuracy of different types of files is guaranteed, so that office data such as text and images can be efficiently and accurately remotely collaborated.

[0069] In an optional implementation, the method further includes: At the receiving end, a corresponding decompression and reconstruction method is selected according to the file type information: for a text file, the compressed editing script is decompressed, and the editing script is applied to reconstruct the text content; for an image file, the compressed pixel block data is decompressed, the pixel block is decoded, and the decoded pixel block is updated to the original image; for other types of files, the compressed file content is decompressed and directly written into a new file; The reconstructed files are saved to the designated location of the receiving end to complete the differential synchronization of digital office data.

[0070] At the receiving end, the corresponding decompression and reconstruction method is selected according to the file type information. The specific process is as follows: First, the receiving end receives the compressed data packet transmitted by the sending end, which contains file type information, compressed difference data, etc. The receiving end parses the data packet and extracts the file type information.

[0071] Then, according to the file type information, select the corresponding decompression and reconstruction method: For text files, use the following steps: 1) Use a general compression algorithm such as zlib to decompress the compressed editing script to obtain the original editing script.

[0072] 2) Read the original content of the text file saved locally.

[0073] 3) Follow the operating instructions in the editing script to insert, delete, replace, and other operations on the original content in turn to generate updated text content.

[0074] For example, if the edit script is "INSERT 10 Hello", "Hello" will be inserted at the 10th character position of the original text.

[0075] For image files, use the following steps: 1) Use image compression algorithms such as JPEG to decompress the compressed pixel block data to obtain the encoded pixel block.

[0076] 2) Decode the encoded pixel block and restore it to the original pixel data.

[0077] 3) Read the original image saved locally.

[0078] 4) Update the decoded pixel block to the corresponding position of the original image to generate an updated image.

[0079] For example, update the 100x100 pixel block at position (100,100).

[0080] For other file types, follow these steps: 1) Use a general compression algorithm such as zlib to decompress the compressed file content to obtain the complete new file content.

[0081] 2) Directly write the decompressed file contents to a new file.

[0082] Next, save the reconstructed file to the specified location of the receiving end. The specific process is: 1) Determine the save path based on the file type, such as saving text files to the " / documents / " directory and saving pictures to the " / images / " directory.

[0083] 2) Generate a file name by adding a timestamp to the original file name.

[0084] 3) Write the reconstructed file content into a new file under the target path.

[0085] Finally, update the file synchronization status, record the version number, modification time and other information of the latest file in the database, and complete the differential synchronization of digital office data.

[0086] Through the above steps, the receiving end can select the appropriate decompression and reconstruction method according to different file types, efficiently complete the update of file content, and save the updated file to the specified location, realizing the differentiated synchronization of digital office data. This method can greatly reduce the amount of data transmission, improve synchronization efficiency, and is suitable for real-time collaborative editing scenarios of various types of files.

[0087] In an optional implementation, a role-based access control system is constructed to assign data access rights to different users; a real-time collaborative editing function is implemented to allow multiple users to edit the same digital office data at the same time, and concurrent editing conflicts are resolved through an operation conversion algorithm, including: Define roles and permissions in the system, where each role contains a set of permissions, and each permission corresponds to a specific type of operation; create a user management module, which is used to create user accounts, verify user identities, and assign roles to users; Implement resource access control logic, and control users' access to and operations on digital office data according to the permissions of their roles; establish collaborative editing sessions, allowing multiple users to join the same session at the same time and jointly edit the same digital office data; define basic operation types, including insert and delete operations, each of which contains the operation type, operation location, and operation content; Implementing an operation conversion function that receives two concurrent operations as input and outputs a converted operation to resolve concurrent editing conflicts; In a collaborative editing session, when an editing operation is received from a user, the following sub-steps are performed: Convert the received editing operation to all existing operations in the session one by one; if the converted operation is not empty, apply the converted operation to the current document content; add the converted operation to the operation history of the session; broadcast the converted operation to other users in the session; On the client side, when an operation broadcast is received from the server, the following sub-steps are performed: Convert the received operation to the locally unsynchronized operation; if the converted operation is not empty, apply the converted operation to the local document content; update the client's local view to display the latest document content; regularly synchronize the document status between the server and the client to ensure that the document content of all clients is consistent; when the user exits the collaborative editing session, save the final document content and update the document access history.

[0088] When implementing a role-based access control system and real-time collaborative editing, first define the roles and permissions in the system. For example, you can set up three roles: "Administrator", "Editor" and "Viewer". Administrators have all permissions, including creating and deleting documents, editing document content, inviting other users, etc. Editors can edit document content but cannot delete documents. Viewers can only view documents and cannot perform any editing operations.

[0089] Next, create a user management module to implement user registration, login, and role assignment. When registering, users need to provide basic information such as username and password, and the system automatically assigns the "Viewer" role to new users. Administrators can modify user roles and promote specific users to "Editor" or "Administrator". When a user logs in, the system verifies the username and password and obtains the user's role information.

[0090] Then implement the resource access control logic. When a user tries to access or operate a document, the system first checks the user's role. If the user is a viewer, the system only allows them to view the document content; if they are an editor, the system allows them to edit the document; if they are an administrator, the system allows them to perform all operations. For unauthorized operations, the system will reject them and return an error message.

[0091] To realize the collaborative editing function, the system establishes a collaborative editing session. When a user opens a document for editing, the system creates a new session or adds the user to an existing session. Each session contains information such as the content of the current document, a list of users participating in the editing, and operation history.

[0092] The system defines two basic operation types: insert operation and delete operation. Insert operation includes the insertion position and content, such as inserting "Hello" after the 10th character in the document. Delete operation includes the starting position and length of the deletion, such as deleting 3 characters starting from the 5th character.

[0093] To resolve concurrent editing conflicts, the system implements an operation conversion function. This function receives two concurrent operations, converts them according to their type and position relationship, and outputs the converted operation. For example, user A inserts "Hello" at position 10, and user B deletes 3 characters at position 5. The operation conversion function adjusts A's insertion position to 7 to accommodate B's deletion operation.

[0094] In a collaborative editing session, when receiving an editing operation from a user, the system first converts the operation with all existing operations in the session. For example, if there are already 10 operations in the session, the newly received operation needs to be converted with these 10 operations one by one. After conversion, if the operation is still valid (for example, the inserted content is not empty), it is applied to the current document content. Then, the system adds the converted operation to the operation history of the session and broadcasts it to other users in the session.

[0095] On the client side, when receiving an operation broadcast by the server, the client will also perform a similar conversion process. The client converts the received operation with the local unsynchronized operation, and then applies it to the local document content. For example, if there is an unsynchronized insert operation locally, and the server broadcast is a delete operation, the client will first convert the operation and then update the local document content. After the update, the client refreshes the user interface to display the latest document content.

[0096] To ensure that the document content of all clients is consistent, the system will perform synchronization operations regularly. For example, every 30 seconds, the server sends the complete document content and the latest operation history to all clients. After receiving the synchronization information, the client updates the local document content to the content sent by the server and reapplies the local unsynchronized operations.

[0097] When the user exits the collaborative editing session, the system saves the final document content. At the same time, the system updates the document access history, including the users who participated in the editing, editing time, major changes, etc. These records can be used for subsequent document version management and audit tracking.

[0098] Through the above steps, the system implements role-based access control and real-time collaborative editing functions, enabling multiple users to jointly edit digital office documents safely and efficiently.

[0099] Figure 2FIG. 1 is a schematic diagram of the structure of a digital office data synchronization and sharing system based on cloud computing according to an embodiment of the present invention. Figure 2 As shown, the system comprises: The first unit is used to deploy a distributed data collection module with an adaptive collection frequency, collect digital office data in real time from multiple heterogeneous terminal devices, and the digital office data includes structured data, semi-structured data and unstructured data; use a deep learning algorithm to intelligently classify and annotate the collected digital office data; upload the digital office data to a distributed cloud storage platform through a load balancing algorithm, and generate a unique identifier based on spatiotemporal features and metadata information containing data lineage for each digital office data in the cloud storage platform; establish a data relationship index based on a graph database to achieve efficient data retrieval and association analysis; The second unit is used to establish a distributed data synchronization system on the cloud computing platform, and detect changes in digital office data on each terminal device through a real-time monitoring mechanism; when data changes are detected, the incremental synchronization algorithm is triggered to synchronize only the changed data parts to reduce the amount of data transmission; the historical version information of each data change is recorded using blockchain-based distributed ledger technology to achieve data version tracking and backtracking; for different types of digital office data, differentiated synchronization strategies are adopted, using character-level difference comparison for text files and pixel-block-level difference comparison for image files; The third unit is used to build a role-based access control system to assign data access rights to different users; implement real-time collaborative editing functions, allowing multiple users to edit the same digital office data at the same time, and resolve concurrent editing conflicts through operation conversion algorithms; establish a data sharing log system to record each user's data access and operation behavior; use machine learning algorithms to analyze user behavior patterns and automatically recommend related digital office data.

[0100] According to a third aspect of the embodiments of the present invention, An electronic device is provided, comprising: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0101] According to a fourth aspect of the embodiments of the present invention, A computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the aforementioned method is implemented.

[0102] The present invention may be a method, an apparatus, a system and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for executing various aspects of the present invention.

[0103] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A digital office data synchronization and sharing method based on cloud computing, characterized in that: include: Deploy a distributed data collection module with adaptive collection frequency to collect digital office data in real time from multiple heterogeneous terminal devices, including structured data, semi-structured data and unstructured data; use deep learning algorithms to intelligently classify and annotate the collected digital office data; upload the digital office data to the distributed cloud storage platform through a load balancing algorithm, and generate a unique identifier based on spatiotemporal characteristics and metadata information containing data lineage for each digital office data in the cloud storage platform; establish a data relationship index based on a graph database to achieve efficient data retrieval and association analysis; A distributed data synchronization system is established on the cloud computing platform to detect changes in digital office data on each terminal device through a real-time monitoring mechanism. When data changes are detected, an incremental synchronization algorithm is triggered to synchronize only the changed data to reduce the amount of data transmission. The blockchain-based distributed ledger technology is used to record the historical version information of each data change to achieve data version tracking and backtracking. Different synchronization strategies are adopted for different types of digital office data, using character-level difference comparison for text files and pixel-block-level difference comparison for image files. Build a role-based access control system to assign data access rights to different users; implement real-time collaborative editing functions, allowing multiple users to edit the same digital office data at the same time, and resolve concurrent editing conflicts through operation conversion algorithms; establish a data sharing log system to record each user's data access and operation behavior; Use machine learning algorithms to analyze user behavior patterns and automatically recommend relevant digital office data.

2. The method according to claim 1, characterized in that The digital office data is uploaded to the distributed cloud storage platform through a load balancing algorithm, and a unique identifier based on spatiotemporal characteristics and metadata containing data lineage are generated for each digital office data in the cloud storage platform, including: The digital office data of multiple terminal devices are collected through a distributed data collection module; the digital office data is preprocessed, including data format standardization, redundant data elimination, compression and encryption; the preprocessed digital office data is uploaded to the distributed cloud storage platform using a dynamic weight-based load balancing algorithm, and the dynamic weight-based load balancing algorithm includes: real-time collection of the CPU usage, memory usage, network throughput and current number of connections of each upload node; using a weighted summation method to calculate the comprehensive performance score of each upload node; based on the comprehensive performance score, using a Softmax function to calculate the dynamic weight of each upload node; using a weighted polling algorithm to select a target upload node according to the dynamic weight; A Ceph-based distributed storage architecture is constructed in the distributed cloud storage platform, and the distributed storage architecture includes an object storage device, a monitor, a metadata server, and a RADOS gateway; a Redis-based distributed cache layer is introduced between the object storage device and the client; a data sharding strategy is dynamically adjusted according to data access patterns and relevance; and an asynchronous replication mechanism is used to replicate data; Generate a unique identifier based on spatiotemporal characteristics for each digital office data, the generation process of the unique identifier includes: generating a microsecond timestamp; converting the geographic location information of the data source into a Geohash code; obtaining a hash value of a device unique identifier; generating a random number; and concatenating the timestamp, the Geohash code, the hash value of the device unique identifier and the random number to generate a final unique identifier through a SHA-256 algorithm; Construct a metadata model based on a graph structure to represent data lineage relationships. Treat each data object as a node in the graph, and the relationship between nodes is represented by directed edges. Create a new node and record basic properties when data is created. Create a new node and add a derived relationship edge pointing to the original data when data is modified. Create a new node and add a merge relationship edge pointing to the source data when data is merged. Add a reference relationship edge between the referencing data and the referenced data when data is referenced.

3. The method according to claim 1, characterized in that The real-time monitoring mechanism detects changes in digital office data on each terminal device. When data changes are detected, the incremental synchronization algorithm is triggered, and only the changed data is synchronized, including: Detect changes in digital office data on each terminal device in real time through the file system event monitoring mechanism; perform periodic scans to detect changes missed by the file system event monitoring mechanism by comparing file size, modification time, and MD5 hash value; predict the probability of file changes using an XGBoost-based machine learning model, which is trained based on the features of file type, file size, last modification time, whether it is a weekday, the hour of the day, and the user ID; The detected change information and predicted change probability are stored in a thread-safe change queue; when data changes are detected, the incremental synchronization process is triggered, including: Divide the changed file into fixed-size data blocks; calculate the weak checksum and strong checksum of each data block; send the data block signature to the receiver; the receiver compares the received signature with the local file signature and identifies the unmatched data blocks; the receiver requests the sender to transmit the unmatched data blocks; the sender transmits only the requested data blocks; the receiver reconstructs the file using the local existing data blocks and the newly received data blocks; The real-time monitoring mechanism and incremental synchronization algorithm are integrated through multi-threading. One thread is responsible for file system monitoring, another thread is responsible for periodic scanning, and another thread is responsible for obtaining change information from the change queue and performing incremental synchronization.

4. The method according to claim 1, characterized in that: Different synchronization strategies are adopted for different types of digital office data. Character-level difference comparison is used for text files, and pixel-block-level difference comparison is used for image files, including: Receiving digital office data to be synchronized; identifying the type of the digital office data, and classifying the digital office data into text files, image files, or other types of files; For digitized office data identified as text files, do the following: Convert the content of the text file into a string; use the Myers difference algorithm to perform character-level difference comparison on the string to generate an editing script; compress the editing script to obtain a compressed editing script; For digitized office data recognized as image files, do the following: Dividing the image file into a plurality of pixel blocks of fixed size; calculating a hash value of each pixel block; comparing the hash values ​​of corresponding pixel blocks in the new and old image files to determine changed pixel blocks; performing JPEG encoding on the changed pixel blocks; and compressing the encoded pixel block data to obtain compressed pixel block data; For digitized office data identified as other types of files, do the following: Reading all the contents of the other types of files; compressing all the contents to obtain compressed file contents; The compressed editing script, the compressed pixel block data or the compressed file content are encapsulated together with the corresponding file type information to form a synchronization data packet; and the synchronization data packet is transmitted to a receiving end.

5. The method according to claim 4, characterized in that The method further comprises: At the receiving end, a corresponding decompression and reconstruction method is selected according to the file type information: for a text file, the compressed editing script is decompressed, and the editing script is applied to reconstruct the text content; for an image file, the compressed pixel block data is decompressed, the pixel block is decoded, and the decoded pixel block is updated to the original image; for other types of files, the compressed file content is decompressed and directly written into a new file; The reconstructed files are saved to the designated location of the receiving end to complete the differential synchronization of digital office data.

6. The method according to claim 1, characterized in that Build a role-based access control system to assign data access rights to different users; Realize real-time collaborative editing function, allowing multiple users to edit the same digital office data at the same time, and resolve concurrent editing conflicts through operation conversion algorithms, including: Define roles and permissions in the system, where each role contains a set of permissions, and each permission corresponds to a specific type of operation; create a user management module, which is used to create user accounts, verify user identities, and assign roles to users; Implement resource access control logic, and control users' access to and operations on digital office data according to the permissions of their roles; establish collaborative editing sessions, allowing multiple users to join the same session at the same time and jointly edit the same digital office data; define basic operation types, including insert and delete operations, each of which contains the operation type, operation location, and operation content; Implementing an operation conversion function that receives two concurrent operations as input and outputs a converted operation to resolve concurrent editing conflicts; In a collaborative editing session, when an editing operation is received from a user, the following sub-steps are performed: Convert the received editing operation to all existing operations in the session one by one; if the converted operation is not empty, apply the converted operation to the current document content; add the converted operation to the operation history of the session; broadcast the converted operation to other users in the session; On the client side, when an operation broadcast is received from the server, the following sub-steps are performed: Convert the received operation to the locally unsynchronized operation; if the converted operation is not empty, apply the converted operation to the local document content; update the client's local view to display the latest document content; regularly synchronize the document status between the server and the client to ensure that the document content of all clients is consistent; when the user exits the collaborative editing session, save the final document content and update the document access history.

7. A digital office data synchronization and sharing system based on cloud computing, used to implement the method according to any one of claims 1 to 6, characterized in that: include: The first unit is used to deploy a distributed data collection module with an adaptive collection frequency, collect digital office data in real time from multiple heterogeneous terminal devices, and the digital office data includes structured data, semi-structured data and unstructured data; use a deep learning algorithm to intelligently classify and annotate the collected digital office data; upload the digital office data to a distributed cloud storage platform through a load balancing algorithm, and generate a unique identifier based on spatiotemporal characteristics and metadata information containing data lineage for each digital office data in the cloud storage platform; Establish data relationship index based on graph database to achieve efficient data retrieval and association analysis; The second unit is used to establish a distributed data synchronization system on the cloud computing platform, and detect changes in digital office data on each terminal device through a real-time monitoring mechanism; when data changes are detected, the incremental synchronization algorithm is triggered to synchronize only the changed data parts to reduce the amount of data transmission; the historical version information of each data change is recorded using blockchain-based distributed ledger technology to achieve data version tracking and backtracking; for different types of digital office data, differentiated synchronization strategies are adopted, using character-level difference comparison for text files and pixel-block-level difference comparison for image files; Unit 3 is used to build a role-based access control system to assign data access rights to different users; Realize real-time collaborative editing function, allowing multiple users to edit the same digital office data at the same time, and resolve concurrent editing conflicts through operation conversion algorithms; establish a data sharing log system to record each user's data access and operation behavior; Use machine learning algorithms to analyze user behavior patterns and automatically recommend relevant digital office data.

8. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images