How unconv/captacity works
Main Components
a. Input/Output Layer:
• Video Input Handler
• Captioned Video Output Handler
b. Core Processing Components:
• Transcription Engine (Whisper - both local and API options)
• Text Processing Pipeline
• Caption Rendering Engine
c. Supporting Components:
• Asset Management (fonts)
• CLI Interface
• Configuration Handler
Component Relationships
1. Input Video → 2. Transcription → 3. Text Processing → 4. Caption Rendering → 5. Output Video
Detailed Component Breakdown
For each major component, include these elements:
Input Handler:
• Video file acceptance
• Format validation
Transcription Engine:
• Local Whisper Integration
• OpenAI API Integration
• Transcription output formatting
Text Processing:
• Segment Parser (segment_parser.py)
• Word timing management
• Text segmentation
Caption Rendering:
• Text Drawer (text_drawer.py)
• Font management
Implementation Details to Highlight
• Dual-path for Whisper transcription (local vs. API)
• Asset management system for fonts
• CLI and programmatic entry points
Specific Elements
[Video Input] → [Transcription Engine]
• [Local Whisper]
• [OpenAI API]
• ↓
[Configuration] → [Text Processing]
[Font Assets] ─┘ ↓
• [Caption Renderer]
• ↓
• [Video Output]