fix(transcription): Behebt Start- und API-Fehler in der App [2f488f42]

transcription-tool/backend/app.py

@@ -121,6 +121,28 @@ def create_insight(meeting_id: int, payload: InsightRequest, db: Session = Depen
         print(f"ERROR: Unexpected error in create_insight: {e}")
         raise HTTPException(status_code=500, detail="An internal error occurred while generating the insight.")
 
+class TranslationRequest(BaseModel):
+    target_language: str
+
+@app.post("/api/meetings/{meeting_id}/translate")
+def translate_meeting_transcript(meeting_id: int, payload: TranslationRequest, db: Session = Depends(get_db)):
+    """
+    Triggers the translation of a meeting's transcript.
+    """
+    try:
+        # For now, we only support English
+        if payload.target_language.lower() != 'english':
+            raise HTTPException(status_code=400, detail="Currently, only translation to English is supported.")
+
+        from .services.translation_service import translate_transcript
+        translation = translate_transcript(db, meeting_id, payload.target_language)
+        return translation
+    except ValueError as e:
+        raise HTTPException(status_code=400, detail=str(e))
+    except Exception as e:
+        print(f"ERROR: Unexpected error in translate_meeting_transcript: {e}")
+        raise HTTPException(status_code=500, detail="An internal error occurred during translation.")
+
 class RenameRequest(BaseModel):
     old_name: str
     new_name: str

transcription-tool/backend/prompt_library.py

@@ -111,3 +111,38 @@ Please provide the output in Markdown format.
 
 # You can add more prompts here for other analysis types.
 # For example, a prompt for a technical summary, a marketing summary, etc.
+
+TRANSLATE_TRANSCRIPT_PROMPT = """
+You are a highly accurate and fluent translator.
+Your task is to translate the given meeting transcript into {target_language}.
+Maintain the original format (who said what) as closely as possible.
+
+**Transcript:**
+---
+{transcript}
+---
+
+**Output:**
+Provide only the translated text. Do not add any commentary or additional formatting.
+"""
+
+_PROMPTS = {
+    "meeting_minutes": MEETING_MINUTES_PROMPT,
+    "action_items": ACTION_ITEMS_PROMPT,
+    "sales_summary": SALES_SUMMARY_PROMPT,
+    "translate_transcript": TRANSLATE_TRANSCRIPT_PROMPT,
+}
+
+def get_prompt(prompt_type: str, context: dict = None) -> str:
+    """
+    Retrieves a prompt by its type and formats it with the given context.
+    """
+    prompt_template = _PROMPTS.get(prompt_type)
+    if not prompt_template:
+        raise ValueError(f"Unknown prompt type: {prompt_type}")
+    
+    if context:
+        return prompt_template.format(**context)
+    
+    return prompt_template
+

transcription-tool/backend/services/insights_service.py

@@ -2,13 +2,9 @@ import sys
 import os
 from sqlalchemy.orm import Session
 from .. import database
-from .. import prompt_library
-
+from ..prompt_library import get_prompt
 import logging
-from sqlalchemy.orm import Session
-from .. import database
-from .. import prompt_library
-from ..lib.gemini_client import call_gemini_flash
+from .llm_service import call_gemini_api
 
 logging.basicConfig(level=logging.INFO)
 logger = logging.getLogger(__name__)
@@ -57,18 +53,7 @@ def _format_transcript(chunks: list[database.TranscriptChunk]) -> str:
 
     return "\n".join(full_transcript)
 
-def get_prompt_by_type(insight_type: str) -> str:
-    """
-    Returns the corresponding prompt from the prompt_library based on the type.
-    """
-    if insight_type == "meeting_minutes":
-        return prompt_library.MEETING_MINUTES_PROMPT
-    elif insight_type == "action_items":
-        return prompt_library.ACTION_ITEMS_PROMPT
-    elif insight_type == "sales_summary":
-        return prompt_library.SALES_SUMMARY_PROMPT
-    else:
-        raise ValueError(f"Unknown insight type: {insight_type}")
+
 
 def generate_insight(db: Session, meeting_id: int, insight_type: str) -> database.AnalysisResult:
     """
@@ -102,7 +87,7 @@ def generate_insight(db: Session, meeting_id: int, insight_type: str) -> databas
         # This can happen if all chunks are empty or malformed
         raise ValueError(f"Formatted transcript for meeting {meeting_id} is empty or could not be processed.")
 
-    prompt_template = get_prompt_by_type(insight_type)
+    prompt_template = get_prompt(insight_type)
     final_prompt = prompt_template.format(transcript_text=transcript_text)
 
     # 4. Call the AI model

transcription-tool/backend/services/llm_service.py

@@ -0,0 +1,91 @@
+import os
+import requests
+import logging
+import time
+
+# Configure logging
+logging.basicConfig(level=logging.INFO)
+
+def call_gemini_api(prompt: str, retries: int = 3, timeout: int = 600) -> str:
+    """
+    Calls the Gemini Pro API with a given prompt.
+
+    Args:
+        prompt: The text prompt to send to the API.
+        retries: The number of times to retry on failure.
+        timeout: The request timeout in seconds.
+
+    Returns:
+        The text response from the API or an empty string if the response is malformed.
+    
+    Raises:
+        Exception: If the API call fails after all retries.
+    """
+    api_key = os.getenv("GEMINI_API_KEY")
+    if not api_key:
+        logging.error("GEMINI_API_KEY environment variable not set.")
+        raise ValueError("API key not found.")
+
+    url = f"https://generativelanguage.googleapis.com/v1/models/gemini-1.5-flash:generateContent?key={api_key}"
+    headers = {'Content-Type': 'application/json'}
+    payload = {
+        "contents": [{
+            "parts": [{"text": prompt}]
+        }],
+        "generationConfig": {
+            "temperature": 0.7,
+            "topK": 40,
+            "topP": 0.95,
+            "maxOutputTokens": 8192,
+        }
+    }
+    
+    for attempt in range(retries):
+        try:
+            response = requests.post(url, headers=headers, json=payload, timeout=timeout)
+            response.raise_for_status()
+            
+            result = response.json()
+
+            if 'candidates' in result and result['candidates']:
+                candidate = result['candidates'][0]
+                if 'content' in candidate and 'parts' in candidate['content']:
+                    # Check for safety ratings
+                    if 'safetyRatings' in candidate:
+                        blocked = any(r.get('blocked') for r in candidate['safetyRatings'])
+                        if blocked:
+                            logging.error(f"API call blocked due to safety ratings: {candidate['safetyRatings']}")
+                            # Provide a more specific error or return a specific string
+                            return "[Blocked by Safety Filter]"
+                    return candidate['content']['parts'][0]['text']
+            
+            # Handle cases where the response is valid but doesn't contain expected content
+            if 'promptFeedback' in result and result['promptFeedback'].get('blockReason'):
+                reason = result['promptFeedback']['blockReason']
+                logging.error(f"Prompt was blocked by the API. Reason: {reason}")
+                return f"[Prompt Blocked: {reason}]"
+
+            logging.warning(f"Unexpected API response structure on attempt {attempt+1}: {result}")
+            return ""
+
+        except requests.exceptions.HTTPError as e:
+            if e.response.status_code in [500, 502, 503, 504] and attempt < retries - 1:
+                wait_time = (2 ** attempt) * 2  # Exponential backoff
+                logging.warning(f"Server Error {e.response.status_code}. Retrying in {wait_time}s...")
+                time.sleep(wait_time)
+                continue
+            logging.error(f"HTTP Error calling Gemini API: {e.response.status_code} {e.response.text}")
+            raise
+        except requests.exceptions.RequestException as e:
+            if attempt < retries - 1:
+                wait_time = (2 ** attempt) * 2
+                logging.warning(f"Connection Error: {e}. Retrying in {wait_time}s...")
+                time.sleep(wait_time)
+                continue
+            logging.error(f"Final Connection Error calling Gemini API: {e}")
+            raise
+        except Exception as e:
+            logging.error(f"An unexpected error occurred: {e}", exc_info=True)
+            raise
+
+    return "" # Should not be reached if retries are exhausted

transcription-tool/backend/services/translation_service.py

@@ -0,0 +1,64 @@
+from sqlalchemy.orm import Session
+from ..database import Meeting, AnalysisResult
+from .llm_service import call_gemini_api
+from ..prompt_library import get_prompt
+from typing import Dict, Any, List
+
+def _format_transcript_for_translation(chunks: List[Any]) -> str:
+    """Formats the transcript into a single string for the translation prompt."""
+    full_transcript = []
+    # Ensure chunks are treated correctly, whether they are dicts or objects
+    for chunk in chunks:
+        json_content = getattr(chunk, 'json_content', None)
+        if not json_content:
+            continue
+        for line in json_content:
+            speaker = line.get("speaker", "Unknown")
+            text = line.get("text", "")
+            full_transcript.append(f"{speaker}: {text}")
+    return "\n".join(full_transcript)
+
+def translate_transcript(db: Session, meeting_id: int, target_language: str) -> AnalysisResult:
+    """
+    Translates the transcript of a meeting and stores it.
+    """
+    meeting = db.query(Meeting).filter(Meeting.id == meeting_id).first()
+    if not meeting:
+        raise ValueError("Meeting not found")
+
+    prompt_key = f"translation_{target_language.lower()}"
+
+    # Check if translation already exists
+    existing_translation = db.query(AnalysisResult).filter(
+        AnalysisResult.meeting_id == meeting_id,
+        AnalysisResult.prompt_key == prompt_key
+    ).first()
+
+    if existing_translation:
+        return existing_translation
+
+    # Prepare transcript
+    transcript_text = _format_transcript_for_translation(meeting.chunks)
+    if not transcript_text:
+        raise ValueError("Transcript is empty, cannot translate.")
+
+    # Get prompt from library
+    prompt = get_prompt('translate_transcript', {
+        'transcript': transcript_text,
+        'target_language': target_language
+    })
+
+    # Call Gemini API using the new service function
+    translated_text = call_gemini_api(prompt)
+
+    # Store result
+    translation_result = AnalysisResult(
+        meeting_id=meeting_id,
+        prompt_key=prompt_key,
+        result_text=translated_text
+    )
+    db.add(translation_result)
+    db.commit()
+    db.refresh(translation_result)
+
+    return translation_result