Update yml/matnbesada.py

yml/matnbesada.py

@@ -1,52 +1,116 @@
-import os, sys, requests
+import os, sys, requests, time, json
 from gradio_client import Client, handle_file
 
-# دریافت اطلاعات از گیت‌هاب اکشن
-# در متد api/edit فایل در متغیر IMAGE_URL قرار می‌گیرد
-file_url = os.environ.get('IMAGE_URL', '') 
-run_id = os.environ.get('RUN_ID')
-space_url = os.environ.get('SPACE_URL')
-github_run_id = os.environ.get('GITHUB_RUN_ID')
-hf_token = os.environ.get('HF_TOKEN', '')
+raw_prompt = os.environ.get('PROMPT', '')
+run_id = os.environ.get('RUN_ID', '')
+space_url = os.environ.get('SPACE_URL', '')
+github_run_id = os.environ.get('GITHUB_RUN_ID', '')
 
-print(f"شروع پردازش ویسپر برای شناسه: {run_id}")
+def report_failure(error_msg):
+    try:
+        requests.post(
+            f"{space_url}/api/webhook/fail",
+            json={
+                "run_id": run_id,
+                "error": error_msg,
+                "event_type": "matnbesada",
+                "client_payload": {
+                    "prompt": raw_prompt,
+                    "run_id": run_id,
+                    "space_url": space_url
+                },
+                "github_run_id": github_run_id
+            },
+            timeout=15
+        )
+    except Exception as e:
+        print(f"Failed to report failure: {e}")
 
+print('1. Parsing parameters (Masquerade handling)...')
 try:
-    # ۱. دانلود فایل صوتی از هاست شما
-    r = requests.get(file_url, timeout=120)
-    local_filename = "audio_input.mp3"
-    with open(local_filename, 'wb') as f:
-        f.write(r.content)
-
-    # ۲. اتصال به موتور اختصاصی ویسپر
-    client = Client("Opera8/sadabematn", token=hf_token if hf_token else None)
+    # استخراج دیتای جاسازی شده در پرامپت
+    parts = raw_prompt.split("___")
+    actual_run_id = parts[0]
+    lang_val = parts[1] if len(parts) > 1 else "Auto-Detect"
+    opt_val = parts[2].lower() == 'true' if len(parts) > 2 else True
+except Exception as e:
+    actual_run_id = run_id
+    lang_val = "Auto-Detect"
+    opt_val = True
+
+audio_download_url = f"{space_url}/static/images/{actual_run_id}.audio"
+local_audio_path = f"input_audio_{actual_run_id}.tmp"
+
+print(f'2. Downloading audio source from Docker: {audio_download_url}')
+try:
+    req = requests.get(audio_download_url, timeout=120)
+    if req.status_code != 200:
+        raise Exception(f"Audio not found. HTTP Status: {req.status_code}")
     
-    # ۳. اجرای تبدیل (خروجی: متن، SRT، JSON)
+    with open(local_audio_path, 'wb') as f:
+        f.write(req.content)
+    print("   -> Audio downloaded successfully.")
+except Exception as e:
+    err_str = f"Download failed: {str(e)}"
+    print(err_str)
+    report_failure(err_str)
+    sys.exit(1)
+
+print('3. Connecting to WhisperX AI Space (Opera8/sadabematn)...')
+try:
+    client = Client("Opera8/sadabematn")
+    
+    print(f'4. Transcribing audio... (Lang: {lang_val}, Optimize: {opt_val})')
     result = client.predict(
-        handle_file(local_filename),
-        "Persian / فارسی", 
-        True,
+        handle_file(local_audio_path),
+        lang_val,
+        opt_val,
         api_name="/transcribe"
     )
+    
+    # تفکیک خروجی‌های برگشتی
+    full_text = result[0] if isinstance(result, (list, tuple)) else result
+    srt_content = result[1] if isinstance(result, (list, tuple)) and len(result) > 1 else ""
+    word_json = result[2] if isinstance(result, (list, tuple)) and len(result) > 2 else "{}"
+    
+    if isinstance(full_text, dict):
+        full_text = str(full_text)
 
-    full_text = result[0]
-    srt_content = result[1]
-
-    # ۴. ذخیره نتایج در فایل‌های مجزا
-    with open("final_text.txt", "w", encoding="utf-8") as f: f.write(full_text)
-    with open("final_sub.srt", "w", encoding="utf-8") as f: f.write(srt_content)
+    print('5. Uploading Multi-format results back to Docker Space...')
+    
+    # آپلود فایل SRT
+    with open("result.srt", "w", encoding="utf-8") as f:
+        f.write(str(srt_content))
+    with open("result.srt", 'rb') as f:
+        requests.post(f'{space_url}/api/webhook/upload', 
+                     data={'run_id': f"{actual_run_id}_srt", 'github_run_id': '', 'ext': 'srt'}, 
+                     files={'file': f})
+                     
+    # آپلود فایل JSON خام زمان‌بندی
+    with open("result.json", "w", encoding="utf-8") as f:
+        f.write(str(word_json))
+    with open("result.json", 'rb') as f:
+        requests.post(f'{space_url}/api/webhook/upload', 
+                     data={'run_id': f"{actual_run_id}_json", 'github_run_id': '', 'ext': 'json'}, 
+                     files={'file': f})
 
-    # ۵. آپلود به داکر (متن اصلی)
-    requests.post(f"{space_url}/api/webhook/upload", 
-                 data={"run_id": run_id, "github_run_id": github_run_id, "ext": "txt"},
-                 files={"file": open("final_text.txt", "rb")})
+    # آپلود متن کامل (این مرحله سیستم بررسی وضعیت (Status Check) را فعال میکند)
+    with open("result.txt", "w", encoding="utf-8") as f:
+        f.write(str(full_text))
+    with open("result.txt", 'rb') as f:
+        res = requests.post(f'{space_url}/api/webhook/upload', 
+                         data={'run_id': actual_run_id, 'github_run_id': github_run_id, 'ext': 'txt'}, 
+                         files={'file': f})
 
-    # ۶. آپلود به داکر (زیرنویس با پسوند srt)
-    requests.post(f"{space_url}/api/webhook/upload", 
-                 data={"run_id": f"{run_id}_srt", "github_run_id": github_run_id, "ext": "srt"},
-                 files={"file": open("final_sub.srt", "rb")})
+    if res.status_code == 200:
+        print('6. SUCCESS! Lifecycle completed.')
+    else:
+        sys.exit(1)
 
-    print("عملیات با موفقیت پایان یافت.")
 except Exception as e:
-    print(f"خطا: {str(e)}")
+    err_str = str(e)
+    print(f'CRITICAL GENERATION ERROR: {err_str}')
+    if 'quota' in err_str.lower() or 'zerogpu' in err_str.lower():
+        err_str = "سهمیه سرور گیت‌هاب اکشن یا هاگینگ‌فیس پر شده است. لطفاً کمی بعد تلاش کنید."
+    report_failure(err_str)
     sys.exit(1)