TTs whisper

This commit is contained in:
2026-06-18 13:45:32 +02:00
parent 0e7fbbfdca
commit 9a23863320
57 changed files with 10430 additions and 253 deletions

9
backend/.env.example Normal file
View File

@@ -0,0 +1,9 @@
# Raspberry Pi connection config
RASPBERRY_PI_HOST=http://raspberrypi.local
RASPBERRY_PI_PORT=8000
# Auth token for API endpoints
QUIBOT_TOKEN=MY_SECRET_TOKEN
# Backend server config
PORT=3000

6
backend/.gitignore vendored
View File

@@ -1,2 +1,4 @@
__pycache__/
venv/
node_modules/
dist/
.env
*.log

View File

@@ -1,210 +0,0 @@
from fastapi import FastAPI, File, Form, UploadFile, HTTPException, Query
from fastapi.middleware.cors import CORSMiddleware
import subprocess
import threading
import time
import os
import json
import uuid
import hashlib
from pathlib import Path
from pydantic import BaseModel
import RPi.GPIO as GPIO
app = FastAPI()
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_methods=["*"],
allow_headers=["*"],
)
INCOMING_DIR = Path("/tmp/quibot-audio/incoming")
LOCKED_DIR = Path("/tmp/quibot-audio/locked")
PROCESSED_DIR = Path("/tmp/quibot-audio/processed")
INCOMING_DIR.mkdir(parents=True, exist_ok=True)
LOCKED_DIR.mkdir(parents=True, exist_ok=True)
PROCESSED_DIR.mkdir(parents=True, exist_ok=True)
# -------------------------
# GPIO SETUP
# -------------------------
STEP = 23
DIR = 24
EN = 25
GPIO.setmode(GPIO.BCM)
GPIO.setup(STEP, GPIO.OUT)
GPIO.setup(DIR, GPIO.OUT)
GPIO.setup(EN, GPIO.OUT)
GPIO.output(EN, GPIO.LOW)
motor_thread = None
def step_motor(steps, direction, delay=0.001):
GPIO.output(DIR, direction)
for _ in range(steps):
GPIO.output(STEP, GPIO.HIGH)
time.sleep(delay)
GPIO.output(STEP, GPIO.LOW)
time.sleep(delay)
def motor_step(dir):
dir_pin = GPIO.HIGH if dir == "forward" else GPIO.LOW
time.sleep(0.02) # small delay before starting
print("Motor running...")
step_motor(200, dir_pin, 0.001)
# -------------------------
# SAFE COMMAND WHITELIST
# -------------------------
COMMANDS = {
"restart_nginx": ["sudo", "systemctl", "restart", "nginx"],
"uptime": ["uptime"],
"update": ["sudo", "apt", "update"]
}
# -------------------------
# API ENDPOINTS
# -------------------------
@app.post("/run")
def run_task(task: str, token: str):
if token != "MY_SECRET_TOKEN":
raise HTTPException(status_code=403, detail="Unauthorized")
if task not in COMMANDS:
raise HTTPException(status_code=400, detail="Invalid task")
try:
result = subprocess.check_output(COMMANDS[task], text=True)
return {"output": result}
except subprocess.CalledProcessError as e:
return {"error": e.output}
@app.post("/motor/step/forward")
def start_motor(token: str):
global motor_thread
if token != "MY_SECRET_TOKEN":
raise HTTPException(status_code=403, detail="Unauthorized")
motor_thread = threading.Thread(target=motor_step, args=("forward",), daemon=True)
motor_thread.start()
return {"status": "motor started"}
@app.post("/motor/step/backwards")
def start_motor(token: str):
global motor_thread
if token != "MY_SECRET_TOKEN":
raise HTTPException(status_code=403, detail="Unauthorized")
motor_thread = threading.Thread(target=motor_step, args=("backwards",), daemon=True)
motor_thread.start()
return {"status": "motor started"}
@app.post("/motor/stop")
def stop_motor(token: str):
if token != "MY_SECRET_TOKEN":
raise HTTPException(status_code=403, detail="Unauthorized")
GPIO.output(EN, GPIO.HIGH) # disable driver
return {"status": "motor stopped"}
@app.post("/audio/upload")
async def upload_audio(file: UploadFile = File(...), format: str = "wav"):
raw_content = await file.read()
checksum = hashlib.sha256(raw_content).hexdigest()[:16]
filename = f"{checksum[:10]}-{uuid.uuid4().hex[:8]}.wav"
filepath = INCOMING_DIR / filename
filepath.write_bytes(raw_content)
return {"status": "received", "filename": str(filepath), "lock_url": f"/audio/lock/{filepath.name}"}
@app.get("/audio/incoming")
def list_incoming():
files = []
for f in sorted(INCOMING_DIR.iterdir()):
meta = f.stat()
files.append({
"filename": f.name,
"size_bytes": meta.st_size,
"modified_iso": time.ctime(meta.st_mtime),
})
return {"count": len(files), "files": files}
@app.post("/audio/lock/{filename}")
def lock_audio(filename: str):
src = INCOMING_DIR / filename
dst = LOCKED_DIR / filename
if not src.exists():
raise HTTPException(status_code=404, detail=f"File {filename} not found")
if dst.exists():
return {"status": "already_locked", "filename": filename}
os.rename(str(src), str(dst))
return {"status": "locked", "filename": filename}
@app.post("/audio/unlock/{filename}")
def unlock_audio(filename: str):
src = LOCKED_DIR / filename
dst = INCOMING_DIR / filename
if not src.exists():
raise HTTPException(status_code=404, detail=f"File {filename} not found")
os.rename(str(src), str(dst))
return {"status": "unlocked", "filename": filename}
@app.post("/audio/cancel/{filename}")
def cancel_audio(filename: str):
src = LOCKED_DIR / filename
dst = INCOMING_DIR / filename
if not src.exists():
raise HTTPException(status_code=404, detail=f"File {filename} not found")
os.rename(str(src), str(dst))
return {"status": "cancelled", "filename": filename}
@app.post("/audio/process/{filename}")
def process_audio(filename: str):
locked = LOCKED_DIR / filename
processed = PROCESSED_DIR / filename
if not locked.exists():
raise HTTPException(status_code=404, detail=f"File {filename} not found")
os.rename(str(locked), str(processed))
return {"status": "processed", "filename": filename}
@app.on_event("shutdown")
def shutdown():
global motor_running
motor_running = False
GPIO.output(EN, GPIO.HIGH)
GPIO.cleanup()

1919
backend/package-lock.json generated Normal file

File diff suppressed because it is too large Load Diff

28
backend/package.json Normal file
View File

@@ -0,0 +1,28 @@
{
"name": "quibot-backend",
"version": "1.0.0",
"description": "QuiBot robot controller backend - runs on local laptop",
"type": "module",
"main": "dist/index.js",
"scripts": {
"dev": "tsx watch src/index.ts",
"build": "tsc",
"start": "node dist/index.js"
},
"dependencies": {
"axios": "^1.7.0",
"cors": "^2.8.5",
"dotenv": "^16.4.5",
"express": "^4.21.0",
"multer": "^1.4.5-lts.1",
"openai": "^6.44.0"
},
"devDependencies": {
"@types/cors": "^2.8.17",
"@types/express": "^4.17.21",
"@types/multer": "^1.4.12",
"@types/node": "^22.19.21",
"tsx": "^4.19.0",
"typescript": "^5.6.0"
}
}

View File

@@ -0,0 +1 @@
Col·la, pítalo, la ola, ola.

View File

@@ -0,0 +1 @@
Hola, què tal, hola, hola, hola, hola...

View File

@@ -0,0 +1 @@
Hola, que tal, bon dia.

34
backend/src/config.ts Normal file
View File

@@ -0,0 +1,34 @@
import dotenv from 'dotenv';
dotenv.config();
let _raspberryHost = process.env.RASPBERRY_PI_HOST ?? 'http://raspberrypi.local';
let _raspberryPort = Number(process.env.RASPBERRY_PI_PORT) || 8000;
let _token = process.env.QUIBOT_TOKEN ?? 'MY_SECRET_TOKEN';
const APP_PORT = Number(process.env.PORT) || 5000;
export const getRaspberryHost = () => _raspberryHost;
export const getRaspberryPort = () => _raspberryPort;
export const getToken = () => _token;
export function setRaspberryHost(host: string) {
_raspberryHost = host;
}
export function setRaspberryPort(port: number) {
_raspberryPort = port;
}
export function setToken(token: string) {
_token = token;
}
export const getConfig = () => ({
raspberryPi: {
host: getRaspberryHost(),
port: getRaspberryPort(),
},
token: getToken(),
});
export const getAppPort = () => APP_PORT;

View File

@@ -0,0 +1,119 @@
import { Router } from 'express';
import multer from 'multer';
import { execFile } from 'child_process';
import { tmpdir } from 'os';
import { join } from 'path';
import { promisify } from 'util';
import { writeFile, unlink } from 'fs';
import { raspiService } from '../services/raspi.service.js';
const execFileAsync = promisify(execFile);
const writeFileAsync = promisify(writeFile);
const unlinkAsync = promisify(unlink);
const router = Router();
const upload = multer({ storage: multer.memoryStorage() });
router.get('/incoming', async (_req, res) => {
try {
const result = await raspiService.listIncomingAudio();
res.json(result);
} catch (err: unknown) {
const message = err instanceof Error ? err.message : 'Unknown error';
res.status(500).json({ error: `List incoming failed: ${message}` });
}
});
router.post('/lock/:filename', async (req, res) => {
try {
const { filename } = req.params;
const result = await raspiService.lockAudio({ filename });
res.json(result);
} catch (err: unknown) {
const message = err instanceof Error ? err.message : 'Unknown error';
res.status(500).json({ error: `Lock audio failed: ${message}` });
}
});
router.post('/unlock/:filename', async (req, res) => {
try {
const { filename } = req.params;
const result = await raspiService.unlockAudio({ filename });
res.json(result);
} catch (err: unknown) {
const message = err instanceof Error ? err.message : 'Unknown error';
res.status(500).json({ error: `Unlock audio failed: ${message}` });
}
});
router.post('/cancel/:filename', async (req, res) => {
try {
const { filename } = req.params;
const result = await raspiService.cancelAudio({ filename });
res.json(result);
} catch (err: unknown) {
const message = err instanceof Error ? err.message : 'Unknown error';
res.status(500).json({ error: `Cancel audio failed: ${message}` });
}
});
router.post('/process/:filename', async (req, res) => {
try {
const { filename } = req.params;
const result = await raspiService.processAudio({ filename });
res.json(result);
} catch (err: unknown) {
const message = err instanceof Error ? err.message : 'Unknown error';
res.status(500).json({ error: `Process audio failed: ${message}` });
}
});
const whisperModel = process.env.WHISPER_MODEL ?? 'base';
const whisperLanguage = process.env.WHISPER_LANGUAGE ?? 'ca';
router.post('/upload', upload.single('file'), async (req, res) => {
let tmpFile: string | undefined;
try {
if (!req.file) {
return res.status(400).json({ error: 'No audio file provided' });
}
const ext = req.file.originalname.split('.').pop()?.toLowerCase() || 'wav';
tmpFile = join(tmpdir(), `quibot-audio-${Date.now()}.${ext}`);
await writeFileAsync(tmpFile, req.file.buffer);
console.log(`[whisper] Model: ${whisperModel}, Language: ${whisperLanguage}, File: ${tmpFile}`);
const { stdout, stderr } = await execFileAsync('whisper', [
tmpFile,
'--model', whisperModel,
'--language', whisperLanguage,
'--output_format', 'txt',
], { maxBuffer: 50 * 1024 * 1024 });
if (stderr) {
console.log(`[whisper] stderr: ${stderr}`);
}
const transcription = stdout.trim();
res.json({
transcription,
originalFilename: req.file.originalname,
});
} catch (err: unknown) {
const message = err instanceof Error ? err.message : 'Unknown error';
res.status(500).json({ error: `Audio transcription failed: ${message}` });
} finally {
if (tmpFile) {
try {
await unlinkAsync(tmpFile);
} catch {
// ignore cleanup errors
}
}
}
});
export default router;

View File

@@ -0,0 +1,20 @@
import { Router } from 'express';
import { raspiService } from '../services/raspi.service.js';
const router = Router();
router.post('/', async (req, res) => {
const { task } = req.body;
if (!task) {
return res.status(400).json({ error: 'Task name is required' });
}
try {
const result = await raspiService.runTask({ task });
res.json(result);
} catch (err: unknown) {
const message = err instanceof Error ? err.message : 'Unknown error';
res.status(500).json({ error: `Run task failed: ${message}` });
}
});
export default router;

View File

@@ -0,0 +1,53 @@
import { Router } from 'express';
import multer from 'multer';
import { raspiService } from '../services/raspi.service.js';
const router = Router();
// Multer config - store in memory for proxying to RasPi
const upload = multer({ storage: multer.memoryStorage() });
router.post('/step/forward', async (_req, res) => {
try {
const result = await raspiService.motorStepForward();
res.json(result);
} catch (err: unknown) {
const message = err instanceof Error ? err.message : 'Unknown error';
res.status(500).json({ error: `Motor step forward failed: ${message}` });
}
});
router.post('/step/backward', async (_req, res) => {
try {
const result = await raspiService.motorStepBackward();
res.json(result);
} catch (err: unknown) {
const message = err instanceof Error ? err.message : 'Unknown error';
res.status(500).json({ error: `Motor step backward failed: ${message}` });
}
});
router.post('/stop', async (_req, res) => {
try {
const result = await raspiService.motorStop();
res.json(result);
} catch (err: unknown) {
const message = err instanceof Error ? err.message : 'Unknown error';
res.status(500).json({ error: `Motor stop failed: ${message}` });
}
});
router.post('/upload', upload.single('file'), async (req, res) => {
try {
if (!req.file) {
return res.status(400).json({ error: 'No audio file provided' });
}
const result = await raspiService.uploadAudio(req.file.buffer, req.file.originalname);
res.json(result);
} catch (err: unknown) {
const message = err instanceof Error ? err.message : 'Unknown error';
res.status(500).json({ error: `Audio upload failed: ${message}` });
}
});
export default router;

View File

@@ -0,0 +1,25 @@
import { Router } from 'express';
import { getConfig, setRaspberryHost, setRaspberryPort, setToken } from '../config.js';
const router = Router();
router.get('/', (_req, res) => {
const settings = getConfig();
res.json(settings);
});
router.put('/', (req, res) => {
const { raspberryPi, token } = req.body;
if (raspberryPi?.host !== undefined) {
setRaspberryHost(raspberryPi.host);
}
if (raspberryPi?.port !== undefined) {
setRaspberryPort(Number(raspberryPi.port));
}
if (token !== undefined) {
setToken(token);
}
res.json(getConfig());
});
export default router;

25
backend/src/index.ts Normal file
View File

@@ -0,0 +1,25 @@
import express from 'express';
import cors from 'cors';
import router from './routes/router.js';
import { getAppPort, getConfig } from './config.js';
const app = express();
app.use(cors());
app.use(express.json());
// Handle multipart in motor controller separately
app.use('/audio', express.json());
app.use('/motor', express.json());
app.use('/commands', express.json());
app.use(router);
app.get('/health', (_req, res) => {
const settings = getConfig();
res.json({ status: 'ok', settings });
});
app.listen(getAppPort(), () => {
console.log(`QuiBot backend listening on port ${getAppPort()}`);
});

View File

@@ -0,0 +1,14 @@
import { Router } from 'express';
import motorController from '../controllers/motor.controller.js';
import audioController from '../controllers/audio.controller.js';
import commandController from '../controllers/command.controller.js';
import settingsController from '../controllers/settings.controller.js';
const router = Router();
router.use('/motor', motorController);
router.use('/audio', audioController);
router.use('/commands', commandController);
router.use('/settings', settingsController);
export default router;

View File

@@ -0,0 +1,77 @@
import axios from 'axios';
import { getRaspberryHost, getRaspberryPort, getToken } from '../config.js';
interface RunTaskParams {
task: string;
}
interface AudioLockParams {
filename: string;
}
export const raspiService = {
async runTask(params: RunTaskParams) {
const res = await axios.post(`${getRaspberryHost()}:${getRaspberryPort()}/run`, null, {
params: { task: params.task, token: getToken() },
});
return res.data;
},
async motorStepForward() {
const res = await axios.post(`${getRaspberryHost()}:${getRaspberryPort()}/motor/step/forward`, null, {
params: { token: getToken() },
});
return res.data;
},
async motorStepBackward() {
const res = await axios.post(`${getRaspberryHost()}:${getRaspberryPort()}/motor/step/backwards`, null, {
params: { token: getToken() },
});
return res.data;
},
async motorStop() {
const res = await axios.post(`${getRaspberryHost()}:${getRaspberryPort()}/motor/stop`, null, {
params: { token: getToken() },
});
return res.data;
},
async uploadAudio(buffer: Buffer, filename?: string) {
const fname = filename || 'audio.wav';
const ext = fname.split('.').pop()?.toLowerCase() || 'wav';
const formData = new FormData();
formData.append('file', new Blob([buffer]), fname);
const res = await axios.post(`${getRaspberryHost()}:${getRaspberryPort()}/audio/upload`, formData, {
params: { format: ext },
headers: { 'Content-Type': 'multipart/form-data' },
});
return res.data;
},
async listIncomingAudio() {
const res = await axios.get(`${getRaspberryHost()}:${getRaspberryPort()}/audio/incoming`);
return res.data;
},
async lockAudio(params: AudioLockParams) {
const res = await axios.post(`${getRaspberryHost()}:${getRaspberryPort()}/audio/lock/${params.filename}`);
return res.data;
},
async unlockAudio(params: AudioLockParams) {
const res = await axios.post(`${getRaspberryHost()}:${getRaspberryPort()}/audio/unlock/${params.filename}`);
return res.data;
},
async cancelAudio(params: AudioLockParams) {
const res = await axios.post(`${getRaspberryHost()}:${getRaspberryPort()}/audio/cancel/${params.filename}`);
return res.data;
},
async processAudio(params: AudioLockParams) {
const res = await axios.post(`${getRaspberryHost()}:${getRaspberryPort()}/audio/process/${params.filename}`);
return res.data;
},
};

20
backend/tsconfig.json Normal file
View File

@@ -0,0 +1,20 @@
{
"compilerOptions": {
"target": "ES2022",
"module": "Node16",
"moduleResolution": "Node16",
"lib": ["ES2022"],
"outDir": "./dist",
"rootDir": "./src",
"strict": true,
"esModuleInterop": true,
"skipLibCheck": true,
"forceConsistentCasingInFileNames": true,
"resolveJsonModule": true,
"declaration": true,
"declarationMap": true,
"sourceMap": true
},
"include": ["src/**/*"],
"exclude": ["node_modules", "dist"]
}