Document collaboration platform based on cloud computing

Abstract

The invention relates to the technical field of computers, and discloses a document collaboration platform based on cloud computing. Real-time semantic complementation, automatic typesetting optimization and content abstract generation are provided through the intelligent editing assistant module, the manual editing burden is remarkably reduced, and the efficiency is improved; a rule engine of an automatic operation module is combined to realize automatic management of document versions, intelligent resolution of conflicts and standardized review, and the problems of communication cost and conflicts in multi-person cooperation are thoroughly solved; the problem of information overload is solved by relying on the deep semantic processing capability, and accurate positioning of core content is realized; and meanwhile, based on the elastic expansion capability of the distributed cloud architecture, the stability and high availability of multi-terminal real-time cooperation are ensured. Finally, a document processing closed loop integrating intelligent assistance, an automatic process and efficient cooperation is formed, and the bottlenecks of a traditional platform in the aspects of editing efficiency, cooperation fluency, information processing and the automation level are comprehensively broken through.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a document collaboration platform based on cloud computing. Background Art

[0002] There are currently many cloud-based document collaboration platforms that allow multiple users to edit the same document online at the same time and provide collaboration features such as version management, comments, and chat.

[0003] However, existing cloud document collaboration platforms still have shortcomings in the following aspects:

[0004] Editing efficiency relies on manual labor: Document editing and formatting rely heavily on user experience and manual operations, which is inefficient.

[0005] Conflicts and communication costs during collaboration: When multiple people are editing simultaneously, editing conflicts may still arise. Resolving conflicts and reaching consensus requires a lot of manual communication, which is inefficient.

[0006] Information overload and difficulty extracting key information: As document content and collaboration history (comments, modification records) increase, it becomes difficult for users to quickly grasp the full picture of the document and key information.

[0007] Limited automation capabilities: Many repetitive or rule-based tasks (such as automatic typesetting and summary generation) still need to be completed manually;

[0008] Therefore, the present invention aims to overcome the shortcomings of existing capabilities and provide a more intelligent, efficient and convenient document collaboration solution. To this end, a document collaboration platform based on cloud computing is proposed. Summary of the Invention

[0009] In view of the deficiencies of the prior art, the present invention provides a document collaboration platform based on cloud computing to solve the background technical problems.

[0010] To achieve the above objectives, the present invention provides the following technical solutions: a document collaboration platform based on cloud computing, comprising:

[0011] Document storage and management module: uses distributed storage technology to implement document storage, version management, permission control, encrypted storage, and backup and recovery;

[0012] User interface module: supports multi-platform access via web browsers, desktop clients, and mobile applications, and provides an interface for document creation, editing, and collaboration;

[0013] Intelligent Editing Assistant Module: This module, triggered by the user, provides intelligent editing suggestions, intelligent typesetting, summary generation, keyword extraction, and intelligent search functions. It includes a natural language processing (NLP) engine, a machine learning model, and a knowledge graph.

[0014] Automation operation module: The system is passively triggered to provide automated task definition, scheduling, and execution;

[0015] Collaboration and communication module: Based on real-time communication technology, it realizes real-time synchronization, comments and message notifications of multiple users concurrent editing;

[0016] Cloud platform infrastructure components: Deployed in public, private, or hybrid cloud environments, they provide elastically scalable basic resources through virtualization, containerization, and load balancing technologies. They are responsible for providing cloud computing infrastructure and ensuring the stability and scalability of the platform.

[0017] Preferably, in the intelligent editing assistant module:

[0018] The natural language processing (NLP) engine is used to perform text analysis, semantic understanding, text generation and text classification;

[0019] The machine learning models include, but are not limited to, deep learning-based language models, sequence labeling models, and classification models, which are used to implement text completion, rewriting, style suggestions, summary generation, and keyword extraction tasks;

[0020] The knowledge graph is used to provide contextual knowledge and entity relationship information to assist intelligent editing and information understanding.

[0021] Preferably, the automated operation module includes a task scheduler and a rule engine:

[0022] The task scheduler manages and executes scheduled tasks and event-triggered tasks;

[0023] The rule engine triggers automatic typesetting based on document status, content characteristics or user behavior.

[0024] Preferably, the intelligent editing suggestions of the intelligent editing assistant module specifically include: providing text completion, rewriting suggestions, and style suggestions based on the context to assist users in writing.

[0025] Preferably, the intelligent typesetting function of the intelligent editing assistant module specifically includes:

[0026] Automatically adjust the document's title, paragraph and picture formats;

[0027] Standardized typesetting is achieved according to preset templates or user-defined rules.

[0028] Preferably, the summary generation function of the intelligent editing assistant module specifically includes: automatically generating a document summary to help users quickly understand the document topic and key information.

[0029] Preferably, the keyword extraction function of the intelligent editing assistant module specifically includes: automatically extracting document keywords to facilitate user retrieval and classification.

[0030] Preferably, the intelligent search function of the intelligent editing assistant module supports semantic search and locates document content through natural language input.

[0031] Preferably, the automation tasks performed by the automation operation module include:

[0032] Automatic version management: automatically save document versions and provide version comparison and backtracking functions;

[0033] Automatic typesetting: Automatically typeset documents based on preset templates or user-defined rules;

[0034] Intelligent review: Identify potential issues with documents and provide improvement suggestions;

[0035] Automatic summary generation: Automatically generate document summaries.

[0036] Preferably, the distributed storage technology is object storage or distributed file system.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] The present invention provides real-time semantic completion, automatic typesetting optimization, and content summary generation through an intelligent editing assistant module, significantly reducing the burden of manual editing and improving efficiency. The rule engine of the automated operation module is combined to achieve automatic document version management, intelligent conflict resolution, and standardized review, thoroughly resolving communication costs and conflict issues in multi-person collaboration. Relying on deep semantic processing capabilities, the problem of information overload is overcome, enabling precise positioning of core content. At the same time, based on the elastic expansion capabilities of a distributed cloud architecture, the stability and high availability of multi-terminal real-time collaboration are guaranteed. Ultimately, a closed loop of document processing is formed that integrates intelligent assistance, automated processes, and efficient collaboration, comprehensively breaking through the bottlenecks of traditional platforms in editing efficiency, collaborative fluency, information processing, and automation levels.

[0039] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a flow chart of the document collaboration platform based on cloud computing of the present invention;

[0041] Figure 2 This is an exploded view of the intelligent editing assistant of the present invention;

[0042] Figure 3 Generate a flow chart for the invention editor suggestion. DETAILED DESCRIPTION

[0043] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this technical field without making any creative efforts shall fall within the scope of protection of the present invention.

[0044] Please refer to the figure, which shows a document collaboration platform based on cloud computing in the present invention.

[0045] 1. Overall Platform Architecture and Deployment

[0046] Technical implementation and function:

[0047] This platform uses a layered microservices architecture deployed in a hybrid cloud environment (a public cloud + a local private cloud), managing Docker container clusters through the Kubernetes container orchestration system. Its core role is to provide a highly available, elastically scalable infrastructure, ensuring stability for millions of concurrent users.

[0048] 1. Cloud platform infrastructure deployment:

[0049] A Kubernetes cluster (30 nodes) was created in a cloud VPC, while an OpenStack cluster (10 compute nodes) was deployed in a private cloud. Cross-cloud traffic management was achieved through the Istio service mesh, meeting the technical requirements of "deployment in public, private, or hybrid cloud environments." This hybrid deployment approach ensures the security of critical data (in the private cloud) while leveraging the elastic resources of the public cloud to handle traffic spikes.

[0050] 2. Module deployment:

[0051] Document Storage and Management Module: This module uses the MinIO object storage cluster (distributed storage technology) to implement document storage, version management, permission control, encrypted storage, and backup and recovery. This module uses the AES-256 encryption algorithm to protect data security and achieves high availability through erasure coding technology.

[0052] User interface module: Develop a responsive web interface that supports access from web browsers, desktop clients, and mobile applications to meet multi-platform access.

[0053] Intelligent editing assistant module: Deployed on a dedicated node equipped with NVIDIA A100 GPU, it loads the pre-trained BERT model and domain knowledge graph to implement intelligent functions.

[0054] Automated operation module: Integrates Apache Airflow as a task scheduler and Drools rule engine to implement automated tasks.

[0055] Collaborative communication module: Develop real-time communication services based on Socket.IO (encapsulating the WebSocket protocol) to meet the real-time synchronization of concurrent editing by multiple users.

[0056] 3. Passive telescopic mechanism:

[0057] Kubernetes automatically detects CPU load > 80%, triggers horizontal pod autoscaling (HPA), adds new pod instances to handle user requests, and the load balancer routes new user sessions to the newly added nodes.

[0058] 2. Intelligent Editing Assistant Module - Active Trigger Mechanism

[0059] The Smart Editing Assistant module consists of three core components:

[0060] A natural language processing (NLP) engine, fine-tuned based on the BERT-Large model, to perform tasks such as text analysis and semantic understanding;

[0061] Machine learning model cluster: includes a variety of specialized models, such as deep learning language models, sequence labeling models, classification models, etc.

[0062] Knowledge graph: stores contextual knowledge and entity relationship information.

[0063] Scenario 1: Real-time editing suggestions

[0064] Function: Proactively provide intelligent suggestions during the user's editing process to improve writing efficiency;

[0065] Trigger mechanism: keyboard input pauses for more than 500ms or a shortcut key is pressed (active trigger);

[0066] 1. Context Capture:

[0067] The front end captures the text context of 200 characters before and after the cursor position. The pseudo code is as follows:

[0068]

[0069]

[0070] 2. Semantic analysis:

[0071] The NLP engine performs dependency parsing and constructs a grammatical tree. For example, if we input the sentence "Artificial intelligence has extensive applications in the medical field...", the analysis result will indicate that the object is missing.

[0072] 3. Suggestion generation:

[0073] Generate sentence completion suggestions based on the language model: "Applications, such as imaging diagnosis and drug development"; implement text completion and rewriting suggestions.

[0074] 4. Style adaptation:

[0075] The classification model (RoBERTa) analyzes document types and provides style suggestions. For example, in academic or business contexts, "improve efficiency" can be changed to "significantly improve" (formal style) or "substantially optimize" (concise style).

[0076] 5. Front-end rendering:

[0077] Displays a list of suggestions near the cursor in a floating panel that users can apply with a single click.

[0078] Scenario 2: Smart Layout

[0079] Automatically standardize document formats to reduce the burden of typesetting;

[0080] Trigger condition: User inserts an image / table or modifies the title style (active trigger).

[0081] Typesetting rules engine:

[0082] Adopting an event-driven architecture, when a document element change is detected, it automatically matches the preset rule set to perform formatting operations, including the following operations:

[0083] Rule matching mechanism: The engine continuously monitors document DOM tree change events;

[0084] Element type identification: select a rule set based on the label type of the changed node (img / table / h1, etc.);

[0085] Conditional validation: Check context such as document status (draft / final), user role (author / reviewer), etc.

[0086] Rule execution: Applying a predefined sequence of layout instructions.

[0087] Specific implementation example:

[0088] When a user uploads a CT scan image:

[0089] The system automatically adds a numbered legend: Figure 3 .2-1 Chest CT scan results";

[0090] Set the image width to 70% of the page;

[0091] Increase the top and bottom spacing by 12pt;

[0092] Insert at the reference location.

[0093] It can realize automatic adjustment of document format and standardized typesetting according to preset templates.

[0094] This intelligent typesetting system achieves intelligent and standardized processing of document formats through the deep integration of NLP engine and rule engine. Compared with traditional manual typesetting methods, this system has three core advantages:

[0095] Dynamic adaptability: Automatically select the optimal layout template based on the document's semantic content (e.g., medical, legal, or financial domain characteristics). For example, medical papers automatically match the AMA format, and legal documents automatically apply the Bluebook standard.

[0096] Context-aware: Intelligently adjusts the layout structure by analyzing the logical relationships between paragraphs. For example, when a three-paragraph structure of "Experimental Methods-Results-Discussion" is detected, it automatically creates a section separator page.

[0097] Cross-platform consistency: Using a responsive typesetting engine to ensure a unified and professional style across different media such as web / mobile / print preview.

[0098] Scenario 3: Summary Generation and Keyword Extraction

[0099] Dual-mode trigger mechanism:

[0100]

[0101]

[0102] Keyword extraction algorithm:

[0103] Using the keyword scoring formula:

[0104] Score(w)=0.4×TFIDF(w)+0.3×Position(w)+0.3×DomainRelevance(w)

[0105] Where TFIDF(w) is the term frequency minus the inverse document frequency, Position(w) is the position weight (title = 1.0, first paragraph of the text = 0.8), and DomainRelevance(w) is the domain relevance calculated by the knowledge graph.

[0106] Processing example:

[0107] Input text: "Deep learning models demonstrate superior performance in medical image analysis...";

[0108] Output keywords: [“deep learning”, “medical image analysis”, “tumor detection”];

[0109] Realize the function of automatically extracting document keywords.

[0110] Scenario 4: Semantic Search

[0111] The system uses natural language to accurately locate content and deep learning models to understand the semantic intent of queries, enabling accurate cross-document content retrieval. The system adopts a three-tier processing architecture:

[0112] Semantic parsing layer: converts natural language queries into structured intents;

[0113] Vector retrieval layer: calculates the semantic similarity between the query and the document;

[0114] Result optimization layer: Sort the final results based on business rules.

[0115] Specific content:

[0116] Query analysis:

[0117] The user enters a natural language query (e.g., "latest lung cancer screening guidelines").

[0118] 1. Semantic parsing engine execution:

[0119] Entity recognition: [“lung cancer”, “screening guidelines”];

[0120] Intent classification: Identified as "obtaining regulatory documents";

[0121] Query expansion: add synonyms ["lung malignancy", "screening program"] based on the knowledge graph;

[0122] 2. Vectorized processing:

[0123] Use the BERT model to convert the query text into a 384-dimensional vector;

[0124] The formula is:

[0125]

[0126] in, represents the query vector, Represents 384-dimensional real number space.

[0127] 3. Nearest neighbor search:

[0128] Perform approximate nearest neighbor searches in the Faiss index library;

[0129] Retrieve the top 100 relevant document fragments;

[0130] The distance metric used is cosine similarity:

[0131]

[0132] in, is the document vector, and ||·|| represents the vector modulus.

[0133] 4. Reorder the results:

[0134] Combined with the traditional BM25 algorithm score (30%);

[0135] Add business weights: document type (Guidelines > Research > Comments), timeliness (preferred within the past three years);

[0136] Final sorting formula:

[0137] finalScore=0.7×cosine+0.2×BM25+0.1×recency

[0138] Among them, finalScore is the final relevance score of the document, which is used for result sorting; cosine is the cosine similarity between the query vector and the document vector; BM25 is the relevance score of the traditional keyword retrieval algorithm; recency is the document timeliness weight.

[0139] 3. Automated Operation Module - Passive Trigger Mechanism

[0140] The automated operation module is one of the core innovations of this invention. It implements a passive triggering mechanism through an event-driven architecture and a rule engine, significantly reducing manual intervention. This module consists of two core components:

[0141] Task scheduler (Apache Airflow): responsible for the execution management of scheduled tasks and event-triggered tasks;

[0142] Rules engine (Drools): automatically triggers actions based on predefined rules;

[0143] Ensure operational consistency through standardized rules; automatically handle repetitive tasks (such as version management and batch typesetting); support dynamic addition of new rules without affecting system operation.

[0144] Scenario 5: Automatic version management

[0145] Realize fully automatic tracking and backtracking of historical versions of documents;

[0146] Trigger condition: document save event (passive trigger);

[0147] 1. Event capture:

[0148] When the user clicks the Save button, the system generates a DOC_SAVED event, which carries metadata such as the document ID, modifier, and timestamp. For example:

[0149]

[0150] 2. Difference Analysis:

[0151] Calculate the difference between the current version and the previous version using the longest common subsequence (LCS) algorithm. Example:

[0152]

[0153] 3. Version snapshot generation:

[0154] Stores complete difference data and generates readable summary of changes. Example:

[0155]

[0156]

[0157] 4. Version tree construction:

[0158] An improved Merkle tree structure is used to manage version relationships, supporting:

[0159] Linear version: v1.0→v1.1→v1.2;

[0160] Branch version: feature / ai-review → merge to main

[0161] 5. Quick backtracking:

[0162] The user can select any version through the timeline interface, and the system completes the recovery within 200ms. Example:

[0163]

[0164] Scenario 6: Automated Typesetting

[0165] Automatically apply typographical rules based on document status;

[0166] Trigger mechanism: document status change events (such as submission for review, publication);

[0167] Rules engine workflow:

[0168] 1. Rule definition (XML format):

[0169]

[0170]

[0171] 2. Rule matching:

[0172] When the document status changes to "approved", the rule engine: checks that the document type is "research_paper";

[0173] Load the associated APA format template;

[0174] Execute automatic numbering of level 3 headings.

[0175] 3. Typesetting execution:

[0176] Title processing:

[0177]

[0178] Reference sorting: Rearrange the reference list in alphabetical order of the author's last name. Layout comparison:

[0179]

[0180] Scenario 7: Intelligent review automatically detects document issues and provides correction suggestions;

[0181] Trigger conditions:

[0182] Immediate trigger: document lock event;

[0183] Scheduled trigger: Batch review at 02:00 every day.

[0184] Three-level review system:

[0185] 1. Syntax layer check:

[0186] Spell checking (PySpellChecker library);

[0187] Passive voice detection:

[0188]

[0189] 2. Fact-checking:

[0190] Connect to the medical knowledge graph to verify the authenticity of the statement. Example:

[0191]

[0192] 3. Logical integrity analysis:

[0193] Use graph neural networks to build a document argument relationship graph and detect missing arguments (isolated nodes) and circular arguments (circular paths).

[0194] Review report example:

[0195] Key issues

[0196] 1. Factual Errors (High Risk)

[0197] Position: Section X;

[0198] Content: "RNA vaccines are absolutely safe and have no side effects";

[0199] Basis: Knowledge graph association 12 studies reporting mild side effects;

[0200] Suggestion: Modify to "RNA vaccines are relatively safe, and common side effects are..."

[0201] 2. Format issues (medium risk)

[0202] All tables lack unit descriptions;

[0203] Recommendation: Add “(unit: mg / dL)” to the table title.

[0204] 4. Multi-module Collaboration Scenario

[0205] 1. Collaborative editing core process

[0206] Technical interaction logic:

[0207] Triggering stage: When the user initiates an editing operation through the interface module, the intelligent editing assistant module analyzes the context in real time (active triggering);

[0208] Conflict resolution phase: The collaborative communication module detects concurrent modifications and calls the operation conversion algorithm to automatically resolve the conflict;

[0209] Persistence phase: The document storage and management module performs versioned encrypted storage and notifies the automated operation module.

[0210] Feedback stage: The rule engine triggers automated tasks (such as format verification / summary update) based on the modified content.

[0211] Key technical features achieved:

[0212] Dual-channel communication architecture:

[0213] Control channel (metadata transmission): handles permission verification and version number synchronization;

[0214] Data channel (content transfer): Only the text changeset is transferred via a differencing algorithm.

[0215] Real-time status synchronization mechanism:

[0216] Use CRDT (Conflict-Free Replicated Data Type) to ensure eventual consistency;

[0217] Collaborative cursor position mapping algorithm: Position = BaseOffset + LocalOffset.

[0218] 2. Document lifecycle automation

[0219] Creation phase:

[0220] The user interface module receives the creation request → the storage module generates a unique document ID and initializes the version tree;

[0221] Editing stage:

[0222] User actions trigger the intelligent editing assistant (active suggestions);

[0223] The automated operation module is executed passively each time a save is made:

[0224] Version snapshot (difference storage);

[0225] Trigger association rules (e.g., start summary generation when word count > 5000);

[0226] Review stage:

[0227] Status mark changes (such as "Submit for review") → the rule engine automatically matches the strategy;

[0228] Call the smart module to perform terminology checking / format validation;

[0229] The collaboration module notifies relevant personnel;

[0230] Release phase:

[0231] The automated operation module performs final version processing:

[0232] Solidify the layout (apply enterprise templates);

[0233] Generate an archive copy (PDF / A format);

[0234] The storage module starts cross-region backup;

[0235] Archiving stage:

[0236] Migration is triggered by scheduled tasks (e.g., no modification for 6 months);

[0237] The storage module moves the data to a lower-cost storage tier.

[0238] Implement a non-intervention workflow of "status change - automatic response - closed-loop processing", covering the full-cycle automated management of documents from creation to destruction.

[0239] An embodiment of the present invention provides a document collaboration platform based on cloud computing. Its core is the deep collaboration of the intelligent editing assistant module (integrating a natural language processing engine, a machine learning model and a knowledge graph to achieve real-time interactive assistance such as semantic completion, dynamic typesetting optimization and content summary generation) and the automated operation module (based on a rule engine and a task scheduler to perform document version management, intelligent review and standardized typesetting and other system-level process control), combined with the multi-terminal real-time collaboration capabilities guaranteed by a distributed cloud architecture, to form an intelligent management closed loop covering the entire life cycle of documents, significantly improving team collaboration efficiency and professional document quality.

Claims

1. A document collaboration platform based on cloud computing, characterized by: include: Document storage and management module: uses distributed storage technology to implement document storage, version management, permission control, encrypted storage, and backup and recovery; User interface module: supports multi-platform access via web browsers, desktop clients, and mobile applications, and provides an interface for document creation, editing, and collaboration; Intelligent Editing Assistant Module: This module, triggered by the user, provides intelligent editing suggestions, intelligent typesetting, summary generation, keyword extraction, and intelligent search functions. It includes a natural language processing (NLP) engine, a machine learning model, and a knowledge graph. Automation operation module: The system is passively triggered to provide automated task definition, scheduling, and execution; Collaboration and communication module: Based on real-time communication technology, it realizes real-time synchronization, comments and message notifications of multiple users concurrent editing; Cloud platform infrastructure components: Deployed in public, private, or hybrid cloud environments, they provide elastically scalable basic resources through virtualization, containerization, and load balancing technologies. They are responsible for providing cloud computing infrastructure and ensuring the stability and scalability of the platform.

2. A document collaboration platform based on cloud computing according to claim 1, characterized in that: In the intelligent editing assistant module: The natural language processing (NLP) engine is used to perform text analysis, semantic understanding, text generation and text classification; The machine learning models include, but are not limited to, deep learning-based language models, sequence labeling models, and classification models, which are used to implement text completion, rewriting, style suggestions, summary generation, and keyword extraction tasks; The knowledge graph is used to provide contextual knowledge and entity relationship information to assist intelligent editing and information understanding.

3. The document collaboration platform based on cloud computing according to claim 1, characterized in that: The automation operation module includes a task scheduler and a rule engine: The task scheduler manages and executes scheduled tasks and event-triggered tasks; The rule engine triggers automatic typesetting based on document status, content characteristics or user behavior.

4. The document collaboration platform based on cloud computing according to claim 1, characterized in that: The intelligent editing suggestions of the intelligent editing assistant module specifically include: providing text completion, rewriting suggestions, and style suggestions based on the context to assist users in writing.

5. The document collaboration platform based on cloud computing according to claim 1, characterized in that: The intelligent typesetting function of the intelligent editing assistant module specifically includes: Automatically adjust the document's title, paragraph and picture formats; Standardized typesetting is achieved according to preset templates or user-defined rules.

6. The document collaboration platform based on cloud computing according to claim 1, characterized in that: The summary generation function of the intelligent editing assistant module specifically includes: automatically generating a document summary to help users quickly understand the document topic and key information.

7. The document collaboration platform based on cloud computing according to claim 1, characterized in that: The keyword extraction function of the intelligent editing assistant module specifically includes: automatically extracting document keywords to facilitate user retrieval and classification.

8. The document collaboration platform based on cloud computing according to claim 1, characterized in that: The intelligent search function of the intelligent editing assistant module supports semantic search and locates document content through natural language input.

9. The document collaboration platform based on cloud computing according to claim 1, characterized in that: The automation tasks performed by the automation operation module include: Automatic version management: automatically save document versions and provide version comparison and backtracking functions; Automatic typesetting: Automatically typeset documents based on preset templates or user-defined rules; Intelligent review: Identify potential issues with documents and provide improvement suggestions; Automatic summary generation: Automatically generate document summaries.

10. The document collaboration platform based on cloud computing according to claim 1, characterized in that: The distributed storage technology is object storage or distributed file system.

Images