ticket system

Rest api for a support ticket system including submit, query, process tickets.
It leverages LLM model to categorize, priority and create initial response to tickets.

features

• create ticket, ticket will be added to redis queue for LLM to process.
• query ticket by ticket id.
• assign ticket priority, category and initial response by AI providers automatically.
• filter tickets by status, priority and category.
• manually trigger ticket processing in queue.

tech stack

• fastapi for webservice
• rq for task queue
• sqlalchemy and pydantic for ORM & data validation
• sqlite as database
• langchain for LLM

setup

• install redis using docker by docker run --name my-redis -p 6379:6379 -d redis.
• install rq(redis queue) by pip install rq.
• setup env variables, please refer to the env file. AI proxy urls could be empty, set them if u access the api through a proxy.

build & run

• build: poetry install && poetry build.
• lint: poetry run flake8.
• run:
  • poetry run uvicorn src.main:app, server will start at localhost:8000 by default.
  • start rq worker: poetry run rq worker -u redis://localhost:6379
• dev mode: poetry run uvicorn src.main:app --reload, reload when codes changed.
• test: poetry run pytest.
• docs: visit localhost:8000/docs for Swagger UI, localhost:8000/redoc for ReDoc.