How alphobla/infiniling-app works
Project type and purpose
• This is a Python desktop (or packaged) application that helps users transcribe audio, generate text, translate, practice vocabulary via spaced repetition, and review statistics.
• Key indicators:
• main.py + main.spec ⇒ PyInstaller packaging of a standalone app
• data/transcriptions_and_audio ⇒ raw audio + subtitle files
• src/gentexter_mode & src/transcriber_mode ⇒ two major modes (text-generation & transcription)
• src/shared ⇒ reusable UI components, database models, translation, frequency analysis
File structure
• Top-level:
• main.py & main.spec – application entrypoint and build spec
• config.json & .claude/settings.local.json – global and local configuration
• data/… – persistent audio and transcripts on disk
• requirements.txt – third-party dependencies
• src/ – all source code, subdivided into three packages:
– gentexter_mode/ – modules for text generation, review UI, stats UI, spaced-repetition selector, audio generation
– transcriber_mode/ – modules to load audio, invoke transcription engine, display UI
– shared/ – common UI widgets (menus, reader panels, vocabulary panels), utility modules (audio controls, frequency analysis), database models, external API clients (GPT translator, Tatoeba client)
• vocabulary.db – SQLite database backing spaced-repetition and vocabulary tracking
Since README.md is empty, infer stack and architecture from names
• Likely uses OpenAI GPT APIs (gpt_translator.py, text_generator.py)
• May use Whisper or another speech-to-text engine in transcriber.py
• May use a text-to-speech engine in audio_generator.py
Main components
a. UI Layer (“Front-end” within the desktop app):
• Gentexter Mode UI (config, review, stats)
• Transcriber Mode UI
• Shared widgets (menu, reader UI, vocabulary panel)
b. Orchestration / Business Logic Layer (“Back-end” modules):
• Orchestrator (orchestrator_updated.py) – routes user actions to services
• Text Generator (text_generator.py)
• Audio Generator (audio_generator.py)
• Transcriber Service (transcriber.py)
• Spaced Repetition Selector (spaced_repetition_selector.py)
• Frequency Analysis (frequency_analysis.py)
• GPT Translator (gpt_translator.py)
• Tatoeba Client (tatoeba_client.py)
• Database Models (database_models.py)
c. Data Layer:
• Local file storage (data/transcriptions_and_audio/*.mp3, .srt)
• SQLite database (vocabulary.db)
• Configuration files (config.json, settings.local.json)
d. External Services:
• OpenAI API (GPT text completion/translation)
• Tatoeba API (example sentence retrieval)
• (Optional) Speech-to-Text engine / TTS engine endpoints
Relationships and interactions
• UI components send user events to the Orchestrator
• Orchestrator invokes Business Logic services
• Services read/write transcripts from/to local storage
• Services persist/retrieve vocabulary & review data via database_models → vocabulary.db
• TextGenerator & GPT Translator call OpenAI API
• TatoebaClient calls external Tatoeba REST API
• AudioGenerator outputs .mp3 files to data folder
• Transcriber consumes .mp3, returns .srt transcripts
Architectural patterns and design principles
• Layered architecture (UI → Orchestrator → Services → Data)
• Modular modes (gentexter vs transcriber) share common components via src/shared
• Single Orchestrator to decouple UI modules from service implementations
• Persistence via a small embedded SQLite store + flat-file assets
Key callouts to capture in detail
• The shared/src folder: highlight common widgets & utilities reused by both modes
• The orchestrator_updated.py as the central dispatcher
• Flow of an audio transcription session vs. a text-generation/spaced-repetition session side by side
• Configuration sources: config.json vs local overrides