How daetoya608/examassistent works
ExamAssistant is an exam-preparation assistant: users upload lecture PDFs, then ask questions that can draw on those documents. The documented upload path parses and chunks PDF pages, stores the file, and indexes chunks; the documented question path may retrieve relevant chunks before generating an answer. The graph includes the API, use cases, domain services, and principal adapters. Some internal chat/agent behavior is represented from README descriptions and source-index references; unsampled implementation details are not asserted as verified calls.
Main entry points: upload_document.py Upload use case · chat.py Chat use case · service.py Vector service · session.py Postgres session
Components 18
API and workflows
- Chat endpoint
app/api/v1/chat.py - Document endpoint
app/api/v1/documents.py - Chat use case
app/application/use_cases/chat.py - Upload use case
app/application/use_cases/upload_document.py
Domain services
- Chat service
app/domains/chats/service.py - Document service
app/domains/documents/service.py - Vector service
app/domains/vector_db/service.py - User service
app/domains/users/service.py
AI and file integrations
- PDF parser
app/infrastructure/parsers/pdf_parser/pdf_parser.py - Backblaze storage
app/infrastructure/file_storage/s3/backblaze_storage.py - LangGraph agent
app/infrastructure/langgraph_agent/agent.py - OpenAI adapter
app/infrastructure/openai_llm/langchain_openai_repo.py - Qdrant chunk index
app/infrastructure/vector_db/qdrant/docs_repository.py
Persistence
- Chat repository
app/infrastructure/persistence/postgres/modules/chats/repository.py - Document repository
app/infrastructure/persistence/postgres/modules/documents/repository.py - User repository
app/infrastructure/persistence/postgres/modules/users/repository.py - Postgres session
app/infrastructure/persistence/postgres/connection/session.py
Other components
- Student
Connections 19
- Student → Document endpoint: uploads PDF
- Student → Chat endpoint: asks question
- Document endpoint → Upload use case: invokes
app/api/v1/documents.py - Chat endpoint → Chat use case: invokes
app/api/v1/chat.py - Upload use case → PDF parser: parses PDF
app/application/use_cases/upload_document.py - Upload use case → Document service: chunks pages
app/application/use_cases/upload_document.py - Upload use case → Backblaze storage: saves file
app/application/use_cases/upload_document.py - Upload use case → Vector service: indexes chunks
app/application/use_cases/upload_document.py - Upload use case → Document repository: saves metadata
app/application/use_cases/upload_document.py - Vector service → Qdrant chunk index: delegates storage
- Chat use case → Chat service: uses history
app/application/use_cases/chat.py - Chat use case → Vector service: retrieves context
app/application/use_cases/chat.py - Chat use case → LangGraph agent: coordinates response
- Chat use case → OpenAI adapter: generates answer
- Chat service → Chat repository: persists messages
- Upload use case → User service: uses user service
app/application/use_cases/upload_document.py - Chat repository → Postgres session: uses session
- Document repository → Postgres session: uses session
- User repository → Postgres session: uses session