Add ASR transcript summarization

Here’s a polished, professional PR/feature description based on your points 👇

---

✅ Title

Add ASR Transcript Summarization Feature with OpenAI Integration and Fallback Support

---

📄 Description (Recommended)

This update introduces a comprehensive summarization feature for ASR-generated transcripts, enabling users to quickly extract key insights from speech-to-text output.

🚀 Backend Enhancements

• Added a new summarization service in app/summarizer.py

  • Integrated OpenAI-based summary generation
  • Implemented retry mechanism with anti-copy safeguards
  • Added local fallback summarization when API is unavailable
  • Supports both short and bullet-point summary formats

• Introduced a new API endpoint in app/main.py:

  • POST /api/summarize

  • Request parameters:

    • text (required)
    • summary_type (short | bullet)
    • max_sentences (1–100)
    • max_words (20–4000)

  • Response includes summarized text along with basic length metadata

---

🎨 Frontend Updates

• Added a new summary section in templates/index.html

• Implemented summarization flow in static/app.js:

  • Allows users to generate summaries from current transcripts
  • Handles loading and error states gracefully
  • Displays guidance when API key is missing

---

⚙️ Configuration & Dependencies

• Added OpenAI configuration in app/config.py:

  • OPENAI_API_KEY
  • OPENAI_MODEL (default: gpt-4o-mini)
  • OPENAI_BASE_URL

• Updated requirements.txt with:

  • openai
  • huggingface_hub

---

🧪 Testing

• Added tests/test_summarizer.py covering:

  • Similarity guard behavior
  • Retry handling for repeated outputs
  • Fallback summarization logic
  • Bullet-format output validation

---

💡 Impact

This feature significantly improves usability by allowing users to quickly understand long transcripts, while ensuring reliability through fallback mechanisms and robust error handling.

---

screenshots