Nice-2-Meet-U Feedback Microservice

API service for collecting and querying both profile-to-profile and app-level feedback within the Nice-2-Meet-U platform. Built with FastAPI, backed by MariaDB/MySQL JSON capabilities, and designed for easy consumption by other services.

Getting Started

• Install deps: python -m venv venv && source venv/bin/activate && pip install -r requirements.txt
• Configure env: copy .env.example (or create .env) and provide DB_HOST, DB_PORT, DB_USER, DB_PASSWORD, DB_NAME, and optional FASTAPIPORT.
• Run locally: uvicorn main:app --reload --port ${FASTAPIPORT:-8000}
• Health check: curl http://localhost:8000/health

On startup the service will auto-create the feedback_profile and feedback_app tables if they do not exist.

Docker

• Build: docker build -t feedback-service .
• Run (port 8001): docker run --env-file .env -e FASTAPIPORT=8001 -p 8001:8001 feedback-service
• Override FASTAPIPORT to use a different container/host port (update both -p mapping and env var).
• Compose (API + DB): FASTAPIPORT=8001 DB_PORT=3310 docker compose up --build
• App connects to DB via service name db on port 3306 inside the network; host port is DB_PORT (default 3310). If you already have MySQL on the host, pick an unused DB_PORT.
• Compose waits for the DB healthcheck before starting the app (first start may take a few seconds).

Data Model Highlights

• feedback_profile
  • Captures reviewer→reviewee meeting feedback.
  • tags persists as JSON (properly dumped on write, coerced to list on read).
  • Uniqueness guard: one feedback record per (match_id, reviewer_profile_id).
• feedback_app
  • Stores overall impressions of the Nice-2-Meet-U application.
  • Also uses JSON-backed tags.

Endpoints

Health

GET /health
GET /health/{path_echo}
Returns uptime metadata, host IP, UTC timestamp, and echoes (query/path) for connectivity checks.

Profile Feedback

• POST /feedback/profile
  • Body: ProfileFeedbackCreate (IDs, ratings, headline/comment, optional tags: list[str]).
  • Inserts a new record, returns ProfileFeedbackOut with generated id and timestamps.
• GET /feedback/profile/{id}
  • Path param: feedback UUID.
  • Returns the stored record or 404.
• PATCH /feedback/profile/{id}
  • Body: ProfileFeedbackUpdate (any subset of fields).
  • Updates mutable fields plus updated_at, returns the refreshed record.
• DELETE /feedback/profile/{id}
  • Removes the record (204 on success).
• GET /feedback/profile
  • Query params: filter by reviewee_profile_id, reviewer_profile_id, match_id, tags (comma-separated OR), min_overall, max_overall, since, search (headline/comment contains), plus cursor-based pagination (limit, cursor, sort, order).
  • Response: {"items": [...ProfileFeedbackOut], "next_cursor": "...", "count": N}.
• GET /feedback/profile/stats
  • Query params: reviewee_profile_id (required), optional tags, since.
  • Aggregates totals, averages, distributions, and top tags.

App Feedback

• POST /feedback/app
  • Body: AppFeedbackCreate (overall/facet scores, headline/comment, optional tags).
  • Returns AppFeedbackOut with generated metadata.
• GET /feedback/app/{id}
  • Fetches a single record.
• PATCH /feedback/app/{id}
  • Partial updates via AppFeedbackUpdate.
• DELETE /feedback/app/{id}
  • Removes the entry (204).
• GET /feedback/app
  • Query params: author_profile_id, tags, ratings filters, since, search, sorting, and pagination controls (either cursor or explicit offset with limit).
  • Response mirrors the profile list shape but also includes a pagination object with limit, offset, total, and navigation helpers.
• GET /feedback/app/stats
  • Aggregates totals, rating distribution, facet averages, and tag counts (all optional filters except none required).

Development Notes

• All UUIDs are stored as CHAR(36) strings; convert to/from uuid.UUID when instantiating Pydantic models.
• tags insert/update paths use json.dumps and _coerce_tags ensures outbound data is always list[str].
• Collection pagination defaults to base64 cursoring for backwards compatibility, while the app feedback listing also supports explicit limit/offset metadata for page-based clients.
• Tests can be written with pytest (see requirements.txt).