33 lines
1.1 KiB
Markdown
33 lines
1.1 KiB
Markdown
# Qwen Context: alexander_smart-speaker
|
|
|
|
## Goal
|
|
Voice assistant for Linux with wake word, STT/TTS, AI dialogue, weather, timer/alarm/stopwatch and volume control.
|
|
|
|
## Architecture
|
|
- Entry: `run.py` -> `app/main.py`
|
|
- Audio layer: `app/audio/` (`wakeword.py`, `stt.py`, `tts.py`, `sound_level.py`)
|
|
- Core logic: `app/core/` (`commands.py`, `ai.py`, `config.py`, `cleaner.py`)
|
|
- Features: `app/features/` (weather, timer, stopwatch, alarm, music, cities game)
|
|
- State: `data/*.json`
|
|
|
|
## High-Value Files
|
|
- `app/core/commands.py` for intent routing
|
|
- `app/main.py` for event loop and orchestration
|
|
- `app/core/config.py` for env configuration
|
|
|
|
## How To Work In This Repo
|
|
1. Keep edits minimal and local.
|
|
2. Prefer fixes with clear fallback behavior (microphone/API failures).
|
|
3. Do not hardcode secrets; use `.env` and `.env.example`.
|
|
4. Update README when behavior/commands change.
|
|
|
|
## Quick Checks
|
|
```bash
|
|
./scripts/qwen-check.sh
|
|
```
|
|
|
|
## Notes For Agent
|
|
- If touching audio code, keep Linux compatibility first.
|
|
- For command parsing, add/adjust tests when test infra exists.
|
|
- Preserve Russian command phrases compatibility.
|