A modern web application for analysing and visualising UAV flight logs with an integrated LLM-powered chat assistant.
Upload a tlog or .bin, explore plots, and ask natural-language questions such as “What was the max climb rate?” or “Do you see any GPS anomalies?” — the assistant replies instantly and proposes follow-up questions.
- 📈 Interactive plots for attitude, GPS, battery, RC and custom messages
- 🔍 Semantic search over every telemetry field (Cohere embeddings)
- 🤖 Chat assistant (Groq Llama-3) that:
- Computes on-the-fly metrics (duration, distance, altitude, …)
- Detects anomalies via reasoning (wind, satellite drop-outs, attitude spikes)
- Suggests answerable follow-up questions
- 🏗 FastAPI backend, Vue 2 frontend (Webpack)
- 🐳 Single-command Docker build
- ⚡ GitHub Actions CI + rsync deploy workflow
UAVLogViewer
├─ backend # (ADDED)
│ ├─ run.py # local entry-point: uvicorn
│ ├─ requirements.txt # Python deps
│ ├─ setup.py # pip-installable package
│ └─ uav_log_viewer # Python package
│ ├─ __init__.py
│ ├─ core.py # FastAPI factory
│ ├─ analysis # telemetry analytics
│ │ ├─ __init__.py
│ │ ├─ data_extractor.py
│ │ ├─ telemetry.py
│ │ └─ anomalies.py
│ ├─ chat # LLM assistant
│ │ ├─ __init__.py
│ │ ├─ prompt.py
│ │ ├─ processor.py
│ │ └─ conversation.py
│ └─ routes # FastAPI routers
│ ├─ __init__.py
│ ├─ chat.py
│ └─ analysis.py
│
├─ src # Vue 2 front-end
│ ├─ components
│ │ └─ widgets
│ │ └─ ChatWidget.vue # (ADDED)
│ └─ … # views, router, store
│
├─ build/ # webpack configs
├─ config/ # extra webpack / jest config
├─ static/ # assets copied verbatim
│
├─ .github
│ └─ workflows
│ ├─ nodejs.yml # build / test
│ ├─ nodejsdeploy.yml # rsync deploy
│ ├─ nodejsdeploystable.yml
│ └─ release.yml # docker release
│
├─ Dockerfile
├─ package.json
├─ README.md
└─ preview.gif # animated demo
---
git clone https://github.com/rcaii/UAVLogViewer.git
cd UAVLogViewer
# ───────────── Front-end ─────────────
npm install # needs Node ≥ 18
# ───────────── Back-end ─────────────
python3 -m venv backend/venv
source backend/venv/bin/activate
pip install -r backend/requirements.txtCreate backend/.env and add your keys:
| Variable | Description |
|---|---|
GROQ_API_KEY |
API key for Groq Llama-3 (chat & anomalies) |
COHERE_API_KEY |
API key for Cohere embeddings / rerank |
GROQ_API_BASE |
(opt) Custom Groq endpoint |
GROQ_DEFAULT_MODEL |
(opt) Model name (defaults to llama3-70b-8192) |
Never commit keys – store them as GitHub Secrets for CI.
# terminal ① – FastAPI back-end
cd backend
python run.py # http://localhost:8000 (docs at /docs)
# terminal ② – Vue front-end
npm run dev # http://localhost:8080The frontend proxies API calls to localhost:8000 via CORS (already enabled).
npm run build # outputs static files to dist/Serve the dist folder behind Nginx, or build the Docker image:
docker build -t uavlogviewer .
docker run -p 8080:8080 -p 8000:8000 uavlogviewer| Method | Path | Description |
|---|---|---|
| GET | /health |
Health check (“ok”) |
| POST | /chat/ |
JSON {question, telemetry?} → {answer, suggested_questions} |
| POST | /analysis/metrics |
(example) compute full metric set |
Interactive docs available at http://localhost:8000/docs (Swagger UI).
# Vue unit & e2e
npm run unit # or npm test
# Python type check & lint (optional)
mypy backend/uav_log_viewer
pytest # once you add test_*.py- build.yml – Node build + unit tests
- nodejsdeploy.yml – rsync deploy (needs
DEPLOY_KEY,SERVER_IP,USERNAME,SERVER_DESTINATIONandSERVER_PORTGitHub Secrets) - release.yml – Docker build & push (optional)