Week #2 #

Detailed Requirements Elaboration #

User stories for MVP #

1. Frelancer:
   • Motivation: Protecting professional interests and minimizing legal risks when working with clients.
   • Pain: Lack of legal knowledge, disputes with clients, payment delays, unclear project boundaries, fear of fraud.
   • Needs: Fast contract review, identification of dangerous clauses, clear recommendations, copyright protection.
   • Expectations: Intuitive interface, fast analysis, specific recommendations in plain language, support for different types of contracts.
   • Features: Limited budget, high workload, irregular income, working with different types of clients.
   • Key areas of analysis: Payment terms, intellectual property rights, scope of work, deadlines and penalties, contract termination, confidentiality.
2. Small business owner
   • Motivation: Protection of business when working with clients and contractors.
   • Disadvantages: Lack of legal education, fear of unscrupulous partners, lengthy negotiation of terms.
   • Needs: Quick check, identification of weaknesses, recommendations on how to improve the contract.
   • Expectations: The service should be simple, provide clear recommendations, help avoid mistakes and protect business interests.
   • Features: Limited budget, high workload, need to save time.
3. Tenants/Landlords
   • Motivation: Minimization of risks when renting/leasing housing or office.
   • Pain: Lack of legal literacy, fear of fraud, lack of accessible and cheap legal advice.
   • Needs: Simplicity, speed, affordable cost, clear explanations.
   • Expectations: Service should be intuitive, provide clear advice, highlight risks and explain complex points.
   • Features: Not ready to pay a lot of money for services, value convenience and accessibility.
4. Procurement and Tender Departments
   • Motivation: Optimization of work with suppliers, reduction of risks when concluding contracts, acceleration of procurement processes.
   • Pain: Passage of disputable or unfavorable conditions due to lack of legal expertise at all stages, the need to quickly check a large number of documents.
   • Needs: Automatic audit of contract terms, quick search for non-standard wording, screening of potential risks.
   • Expectations: The service should provide a summary of key differences, support work with large amounts of data, be easy to use.
   • Features: Working with documents from different suppliers, the need to comply with internal standards, high responsibility for each decision.

Acceptance criteria #

1. Document Upload

• Given: user is in space
• When: user uploads a PDF or DOC file up to 10 pages
• Then: the file is successfully uploaded in 5 seconds

2. Analyzing the document

• Given: document is uploaded
• When: user begins chat with the system
• Then: the system already has the document in the database and can answer the user’s question

3. Displaying the results

• Given: analysis is completed
• When: user opens analyzed document
• Then: he/she can see:
  • List of problems found
  • Risk level for each problem (high/medium/low)
  • Simple explanation of each problem.

4. Basic checks

• Given: user is in space
• When: user asks the system to check the document
• Then: the system checks:
  • Payment terms
  • Due dates
  • Termination terms

5. Errors

• Given: user uploads an unsupported file
• When: the system processes the request
• Then: a clear error message is displayed.

Prioritized backlog #

Sprint analysis #

Tasks for the next week #

Project specific progress #

User flow diagram #

Design #

The platform provides a complete web interface with three main screens:

1. Upload Screen: Drag-and-drop interface for document upload.
2. Processing Screen: Real-time processing status with progress indicators.
3. Results Screen: Comprehensive analysis results with risks and recommendations.

Frontend #

• Technology: Vue.js 3 with modern UI components and routing.
• User Interface: A full web application with a complete UI workflow for document analysis, featuring:
  • A smart upload interface with drag-and-drop and progress tracking.
  • A real-time processing screen with status updates.
  • A comprehensive results screen with detailed analysis.
• Architecture: The frontend is a Vue.js Single Page Application (SPA) running on port 8080.

Backend #

• Technology: Spring Boot, PostgreSQL with pgvector, and Docker Compose.
• Architecture:
  • Backend API (port 8000): A Spring Boot application that handles document uploads and orchestrates the analysis pipeline.
  • Database (port 5432): PostgreSQL with the pgvector extension for vector similarity search.
• Key Features:
  • REST API: Comprehensive endpoints with Swagger documentation.
  • Docker Deployment: Fully containerized for easy setup.
  • Error Handling: Includes robust validation and user feedback.

ML #

• Technology: FastAPI, LangChain, OpenRouter, and BAAI/bge-m3 embeddings.
• Architecture:
  • Analyzer Microservice (port 8001): A FastAPI service providing endpoints for analysis, retrieval, and embedding. It contains the RAG pipeline.
• AI Capabilities:
  • AI-Powered Analysis: Leverages LLM for in-depth legal document review.
  • Legal Knowledge Base: A vector database of over 4,400 Russian Civil Code articles.
  • Semantic Search: Vector-based search with configurable distance functions (cosine, L2, inner product).

Weekly commitments #

Individual contribution of each participant #

Alexander Malyy:

• Define API endpoints for MVP (issue)
• Initialize database schema (issue)
• Implement RAG functionality (issue)
• Implement analysis of the document point by point (issue)
• Fix the issues with the frontend (pull request)

Arthur Babkin:

• Create a parser of Russian Legal Database (issue)
• Defined database schema

Vladimir Zhidkov:

• Setup Java backend service for routing between ML microservices and Frontend (issue)
• Сonfigure docker (issue)
• Implement swagger documentation (issue)
• Analyse user stories in more details (issue)
• Analyze acceptance criteria (issue)
• Wrote report to week2

Ilsaf Abdulkhakov:

• Set up the frontend project structure (issue)
• Create a user-flow diagram for MVP (issue)
• Link frontend and backend (pull request)

Nikita Tsukanov:

• Inspect and analyze the thesis provided by Albert (issue)
• Defined the initial RAG architecture

Plan for Next Week #

Frontend #

• Implement final version of frontend
• Link frontend and backend

Backend #

• Implement message broker
• Deploy mvp version of project

ML #

• Experiment and provide report on different document chunking techniques
• Create an assessment data and criteria
• Research better retrieve techniques and analyze the results
• Research and analyze different approaches for better query construction

Confirmation of the code’s operability #

We confirm that the code in the main branch:

• [+] In working condition.
• [+] Run via docker-compose (or another alternative described in the README.md).