feat(transcription): v0.4.0 with structured json, inline editing and deletion

- Backend: Switched prompt to JSON output for structured data
- Backend: Added PUT /chunks/{id} endpoint for persistence
- Backend: Fixed app.py imports and initialization logic
- Frontend: Complete rewrite for Unified View (flattened chunks)
- Frontend: Added Inline Editing (Text/Speaker) and Row Deletion
- Docs: Updated TRANSCRIPTION_TOOL.md with v0.4 features

transcription-tool/backend/services/transcription_service.py

@@ -19,8 +19,8 @@ class TranscriptionService:
         """
         logger.info(f"Uploading chunk {file_path} to Gemini...")
         
-        # 1. Upload file
-        media_file = self.client.files.upload(path=file_path)
+        # 1. Upload file (positional argument)
+        media_file = self.client.files.upload(file=file_path)
         
         # 2. Wait for processing (usually fast for audio)
         while media_file.state == "PROCESSING":
@@ -32,12 +32,18 @@ class TranscriptionService:
 
         # 3. Transcribe with Diarization and Timestamps
         prompt = """
-        Transkribiere dieses Audio wortgetreu. 
-        Identifiziere die Sprecher (Sprecher A, Sprecher B, etc.).
-        Gib das Ergebnis als strukturierte Liste mit Timestamps aus.
-        Wichtig: Das Audio ist ein Teil eines größeren Gesprächs. 
-        Antworte NUR mit dem Transkript im Format:
-        [MM:SS] Sprecher X: Text
+        Transkribiere dieses Audio wortgetreu.
+        Identifiziere die Sprecher (Speaker A, Speaker B, etc.).
+        
+        Gib das Ergebnis als JSON-Liste zurück.
+        Format:
+        [
+          {
+            "time": "MM:SS",
+            "speaker": "Speaker A",
+            "text": "..."
+          }
+        ]
         """
         
         logger.info(f"Generating transcription for {file_path}...")
@@ -45,7 +51,8 @@ class TranscriptionService:
             model="gemini-2.0-flash",
             contents=[media_file, prompt],
             config=types.GenerateContentConfig(
-                temperature=0.1, # Low temp for accuracy
+                temperature=0.1,
+                response_mime_type="application/json"
             )
         )
 
@@ -53,6 +60,6 @@ class TranscriptionService:
         self.client.files.delete(name=media_file.name)
 
         return {
-            "raw_text": response.text,
+            "raw_text": response.text, # This is now a JSON string
             "offset": offset_seconds
         }