Update app.py

app.py

@@ -7,12 +7,92 @@ import os
 import tempfile
 from pathlib import Path
 import gradio as gr
-from transcript_formatter.core.claude_formatter import format_with_claude
+from dotenv import load_dotenv
+import anthropic
 from docx import Document
 from docx.shared import Inches, Pt
 from docx.enum.text import WD_ALIGN_PARAGRAPH
 import re
 
+# Load environment variables
+load_dotenv()
+
+def format_with_claude(text):
+    """Format transcript using Claude AI."""
+    api_key = os.getenv('ANTHROPIC_API_KEY')
+    if not api_key:
+        raise ValueError("ANTHROPIC_API_KEY not found. Please add it to your Hugging Face Space secrets.")
+    
+    client = anthropic.Anthropic(api_key=api_key)
+    
+    system_prompt = """You are a professional transcript formatter. Your task is to intelligently format transcripts while preserving all original content and meaning.
+
+FORMATTING REQUIREMENTS:
+
+1. SPEAKER NAMES:
+   - Bold all speaker names using **Speaker Name:** format
+   - Detect various speaker formats (Speaker:, SPEAKER, Speaker Name, etc.)
+   - Maintain consistent formatting throughout
+
+2. SCRIPTURE REFERENCES:
+   - Bold ALL Scripture references in ANY format using **reference** format
+   - Examples to detect and format:
+     * 1 John 2:18 → **1 John 2:18**
+     * Mark chapter 13 verse 13 → **Mark chapter 13 verse 13**
+     * Romans 8:28-30 → **Romans 8:28-30**
+     * First Corinthians 15 → **First Corinthians 15**
+     * Matt. 5:3-12 → **Matt. 5:3-12**
+   - Include partial references, book names, and various formats
+
+3. CHARACTER ENCODING FIXES:
+   - Fix common encoding issues:
+     * ♪ → ♪
+     * ’ → '
+     * “ → "
+     * †→ "
+     * â€" → —
+     * â€" → –
+   - Convert smart quotes to proper Unicode
+   - Fix any other character encoding problems
+
+4. MUSIC SYMBOLS:
+   - Remove excessive music symbols (♪♪♪ → ♪ or remove entirely if appropriate)
+   - Clean up music notations while preserving meaning
+   - Keep single music symbols if they add context
+
+5. PARAGRAPH STRUCTURE:
+   - Create proper paragraph breaks at natural speech boundaries
+   - Merge fragmented lines into coherent paragraphs
+   - Maintain logical flow and readability
+   - Separate different speakers or topics appropriately
+
+6. CONTENT PRESERVATION:
+   - Preserve ALL original content and meaning
+   - Do not add, remove, or change the substance of what was said
+   - Maintain the speaker's voice and style
+   - Keep all important details and context
+
+7. TIMESTAMP REMOVAL:
+   - Remove timestamps if present (e.g., [00:15:30], (2:45), etc.)
+   - Clean up any time markers that interrupt the flow
+
+8. OUTPUT FORMAT:
+   - Return the formatted text in clean markdown format
+   - Use proper markdown syntax
+   - Ensure readability and professional appearance
+
+Remember: Your goal is to make the transcript more readable and professional while preserving every bit of the original meaning and content."""
+    
+    message = client.messages.create(
+        model="claude-3-5-sonnet-20240620",
+        max_tokens=8000,
+        temperature=0.1,
+        system=system_prompt,
+        messages=[{"role": "user", "content": f"Please format this transcript:\n\n{text}"}]
+    )
+    
+    return message.content[0].text
+
 def create_word_document(formatted_text, title):
     """Create a Word document from formatted text."""
     doc = Document()
@@ -90,6 +170,9 @@ def format_transcript(file):
         else:
             return None, "Please upload a .txt file."
         
+        if not content.strip():
+            return None, "The uploaded file appears to be empty."
+        
         # Format using AI
         formatted_text = format_with_claude(content)
         
@@ -105,7 +188,11 @@ def format_transcript(file):
         return output_path, "✅ Transcript formatted successfully! Download your Word document below."
         
     except Exception as e:
-        return None, f"❌ Error formatting transcript: {str(e)}"
+        error_msg = str(e)
+        if "ANTHROPIC_API_KEY" in error_msg:
+            return None, "❌ API key not configured. Please contact the administrator."
+        else:
+            return None, f"❌ Error formatting transcript: {error_msg}"
 
 # Custom CSS for ORU branding
 css = """
@@ -205,8 +292,4 @@ with gr.Blocks(css=css, title="ORU Transcript Formatter") as demo:
 
 # Launch the demo
 if __name__ == "__main__":
-    demo.launch(
-        server_name="0.0.0.0",
-        server_port=7860,
-        share=False
-    )
+