Update app.py
Browse files
app.py
CHANGED
|
@@ -253,107 +253,140 @@ async def api_transcribe_sync(request: Request):
|
|
| 253 |
)
|
| 254 |
|
| 255 |
# ====== Gradio 介面 ======
|
| 256 |
-
with gr.Blocks(
|
| 257 |
-
gr.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 258 |
|
| 259 |
with gr.Tab("🌐 Web Upload"):
|
| 260 |
-
gr.Markdown("### Upload audio file
|
| 261 |
-
with gr.Row():
|
| 262 |
-
with gr.Column(scale=1):
|
| 263 |
-
pw_ui = gr.Textbox(label="Password", type="password")
|
| 264 |
-
file_ui = gr.File(label="Upload Audio File", file_types=["audio"])
|
| 265 |
-
btn_ui = gr.Button("Start Transcription 🚀", variant="primary", size="lg")
|
| 266 |
-
with gr.Column(scale=2):
|
| 267 |
-
status_ui = gr.Textbox(label="Status", interactive=False)
|
| 268 |
-
transcript_ui = gr.Textbox(label="Transcription Result", lines=10)
|
| 269 |
-
summary_ui = gr.Textbox(label="AI Summary", lines=6)
|
| 270 |
|
| 271 |
-
|
| 272 |
-
|
| 273 |
-
|
| 274 |
-
|
| 275 |
-
|
| 276 |
-
|
| 277 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 278 |
|
| 279 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 280 |
|
| 281 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 282 |
|
| 283 |
-
|
|
|
|
|
|
|
|
|
|
| 284 |
|
| 285 |
-
|
| 286 |
```json
|
| 287 |
{
|
| 288 |
"password": "your_password",
|
| 289 |
-
"file_data": "data:audio/m4a;base64
|
| 290 |
"file_name": "recording.m4a"
|
| 291 |
}
|
| 292 |
```
|
| 293 |
|
| 294 |
-
|
| 295 |
```json
|
| 296 |
{
|
| 297 |
"status": "success",
|
| 298 |
-
"transcription": "
|
| 299 |
-
"summary": "
|
| 300 |
}
|
| 301 |
```
|
| 302 |
|
| 303 |
---
|
| 304 |
|
| 305 |
-
###
|
| 306 |
-
|
| 307 |
-
**動作流程:**
|
| 308 |
|
| 309 |
-
|
| 310 |
-
2. **Base64 編碼**
|
| 311 |
-
3. **文字** (組合 data URL):
|
| 312 |
-
```
|
| 313 |
-
data:audio/m4a;base64,Base64編碼結果
|
| 314 |
-
```
|
| 315 |
-
4. **字典** (請求本文):
|
| 316 |
-
- 鍵: `password`, 值: `chou`
|
| 317 |
-
- 鍵: `file_data`, 值: 上一步的文字
|
| 318 |
-
- 鍵: `file_name`, 值: `recording.m4a`
|
| 319 |
-
5. **取得 URL 內容**:
|
| 320 |
-
- URL: `https://你的網址/api/transcribe`
|
| 321 |
-
- 方法: `POST`
|
| 322 |
-
- 標頭: `Content-Type` = `application/json`
|
| 323 |
-
- 請求本文: 上一步的字典
|
| 324 |
-
- 請求本文類型: `JSON`
|
| 325 |
-
6. **從字典取得值**:
|
| 326 |
-
- 鍵: `transcription` → 轉錄結果
|
| 327 |
-
- 鍵: `summary` → 摘要
|
| 328 |
|
| 329 |
-
|
| 330 |
|
| 331 |
-
|
| 332 |
|
| 333 |
-
|
| 334 |
-
- ✅ 無論音檔多長,都會自動處理完成
|
| 335 |
-
- ✅ 不需要設定等待時間或輪詢機制
|
| 336 |
-
- ✅ 直接取得最終結果,不會有 `event_id`
|
| 337 |
|
| 338 |
-
###
|
| 339 |
|
| 340 |
-
|
| 341 |
-
|
| 342 |
-
|
| 343 |
-
|
| 344 |
-
|
| 345 |
-
|
| 346 |
-
|
| 347 |
-
|
| 348 |
-
|
| 349 |
-
|
|
|
|
|
|
|
|
|
|
| 350 |
""")
|
| 351 |
|
| 352 |
gr.Markdown("""
|
| 353 |
---
|
| 354 |
-
💡 **
|
| 355 |
-
📦 **Max File Size**: 25MB per chunk (auto-split)
|
| 356 |
-
🔒 **Security**: Password-protected
|
| 357 |
""")
|
| 358 |
|
| 359 |
# ====== 掛載 Gradio 到 FastAPI ======
|
|
|
|
| 253 |
)
|
| 254 |
|
| 255 |
# ====== Gradio 介面 ======
|
| 256 |
+
with gr.Blocks(
|
| 257 |
+
theme=gr.themes.Soft(),
|
| 258 |
+
title="LINE Audio Transcription",
|
| 259 |
+
css="""
|
| 260 |
+
/* 手機優化樣式 */
|
| 261 |
+
@media (max-width: 768px) {
|
| 262 |
+
.gradio-container {
|
| 263 |
+
padding: 10px !important;
|
| 264 |
+
}
|
| 265 |
+
/* 強制單欄布局 */
|
| 266 |
+
.contain {
|
| 267 |
+
flex-direction: column !important;
|
| 268 |
+
}
|
| 269 |
+
/* 調整按鈕大小 */
|
| 270 |
+
button {
|
| 271 |
+
font-size: 16px !important;
|
| 272 |
+
padding: 12px !important;
|
| 273 |
+
}
|
| 274 |
+
/* 調整輸入框 */
|
| 275 |
+
input, textarea {
|
| 276 |
+
font-size: 16px !important;
|
| 277 |
+
}
|
| 278 |
+
/* Tab 標籤更大更好點擊 */
|
| 279 |
+
.tabs button {
|
| 280 |
+
padding: 12px 16px !important;
|
| 281 |
+
font-size: 15px !important;
|
| 282 |
+
}
|
| 283 |
+
}
|
| 284 |
+
"""
|
| 285 |
+
) as demo:
|
| 286 |
+
gr.Markdown("# 🎧 LINE Audio Transcription")
|
| 287 |
|
| 288 |
with gr.Tab("🌐 Web Upload"):
|
| 289 |
+
gr.Markdown("### Upload audio file from browser")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 290 |
|
| 291 |
+
# 手機版:改用單欄布局
|
| 292 |
+
pw_ui = gr.Textbox(
|
| 293 |
+
label="Password",
|
| 294 |
+
type="password",
|
| 295 |
+
placeholder="Enter password"
|
| 296 |
+
)
|
| 297 |
+
file_ui = gr.File(
|
| 298 |
+
label="Upload Audio File",
|
| 299 |
+
file_types=["audio"]
|
| 300 |
+
)
|
| 301 |
+
btn_ui = gr.Button(
|
| 302 |
+
"Start Transcription 🚀",
|
| 303 |
+
variant="primary",
|
| 304 |
+
size="lg"
|
| 305 |
+
)
|
| 306 |
|
| 307 |
+
status_ui = gr.Textbox(
|
| 308 |
+
label="Status",
|
| 309 |
+
interactive=False,
|
| 310 |
+
show_label=True
|
| 311 |
+
)
|
| 312 |
+
transcript_ui = gr.Textbox(
|
| 313 |
+
label="Transcription Result",
|
| 314 |
+
lines=8,
|
| 315 |
+
placeholder="Transcription will appear here...",
|
| 316 |
+
show_copy_button=True
|
| 317 |
+
)
|
| 318 |
+
summary_ui = gr.Textbox(
|
| 319 |
+
label="AI Summary",
|
| 320 |
+
lines=5,
|
| 321 |
+
placeholder="Summary will appear here...",
|
| 322 |
+
show_copy_button=True
|
| 323 |
+
)
|
| 324 |
|
| 325 |
+
btn_ui.click(
|
| 326 |
+
transcribe_ui,
|
| 327 |
+
inputs=[pw_ui, file_ui],
|
| 328 |
+
outputs=[status_ui, transcript_ui, summary_ui]
|
| 329 |
+
)
|
| 330 |
+
|
| 331 |
+
with gr.Tab("📱 API Info"):
|
| 332 |
+
gr.Markdown("""
|
| 333 |
+
### iPhone Shortcuts Integration
|
| 334 |
|
| 335 |
+
**Endpoint:**
|
| 336 |
+
```
|
| 337 |
+
POST /api/transcribe
|
| 338 |
+
```
|
| 339 |
|
| 340 |
+
**Request (JSON):**
|
| 341 |
```json
|
| 342 |
{
|
| 343 |
"password": "your_password",
|
| 344 |
+
"file_data": "data:audio/m4a;base64,...",
|
| 345 |
"file_name": "recording.m4a"
|
| 346 |
}
|
| 347 |
```
|
| 348 |
|
| 349 |
+
**Response:**
|
| 350 |
```json
|
| 351 |
{
|
| 352 |
"status": "success",
|
| 353 |
+
"transcription": "...",
|
| 354 |
+
"summary": "..."
|
| 355 |
}
|
| 356 |
```
|
| 357 |
|
| 358 |
---
|
| 359 |
|
| 360 |
+
### Key Points
|
|
|
|
|
|
|
| 361 |
|
| 362 |
+
✅ **Fully synchronous** - Returns result directly
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 363 |
|
| 364 |
+
✅ **No polling needed** - Waits until completion
|
| 365 |
|
| 366 |
+
✅ **Works with any audio length** - Auto-handles long files
|
| 367 |
|
| 368 |
+
---
|
|
|
|
|
|
|
|
|
|
| 369 |
|
| 370 |
+
### Shortcuts Setup
|
| 371 |
|
| 372 |
+
1. Get file → Your audio
|
| 373 |
+
2. Base64 encode
|
| 374 |
+
3. Text: `data:audio/m4a;base64,[encoded]`
|
| 375 |
+
4. Dictionary:
|
| 376 |
+
- `password`: `chou`
|
| 377 |
+
- `file_data`: Step 3 text
|
| 378 |
+
- `file_name`: `recording.m4a`
|
| 379 |
+
5. Get URL contents:
|
| 380 |
+
- URL: `/api/transcribe`
|
| 381 |
+
- Method: POST
|
| 382 |
+
- Header: `Content-Type: application/json`
|
| 383 |
+
- Body: Step 4 dictionary (JSON)
|
| 384 |
+
6. Get `transcription` and `summary` from response
|
| 385 |
""")
|
| 386 |
|
| 387 |
gr.Markdown("""
|
| 388 |
---
|
| 389 |
+
💡 **Formats**: MP4, M4A, MP3, WAV, OGG, WEBM | **Max**: 25MB/chunk | 🔒 **Password-protected**
|
|
|
|
|
|
|
| 390 |
""")
|
| 391 |
|
| 392 |
# ====== 掛載 Gradio 到 FastAPI ======
|