How dwyca-org/fmnarrative works
FMNarrative is a command-line Football Manager 24 press-conference assistant. A user supplies a match screenshot; the Python application invokes the C++ OCR executable, combines extracted match data with manager-provided match context, derives talking points, and requests question-and-answer content from Groq. The README documents the intended realistic, dynamic, context-aware press-conference experience; the sampled excerpts do not show the full prompt or response-display logic, so those details are kept broad.
Main entry points: main.py Groq client · ocr.cpp Screenshot OCR
Components 14
Match capture
- Assistant CLI
main.py - OCR invocation
main.py - Screenshot OCR
ocr.cpp
Match context
- Match statistics
main.py - Match setup
main.py - Team perspective
main.py - Match analysis
main.py
Press conference
- Question prompt
main.py - Conference content
main.py
External services
- OpenCV
- Tesseract OCR
- Groq client
main.py - Groq language model
Other components
- Manager
Connections 19
- Manager → Assistant CLI: provides screenshot
- Assistant CLI → OCR invocation: runs OCR
- OCR invocation → Screenshot OCR: passes image path
- Screenshot OCR → OpenCV: processes image
- Screenshot OCR → Tesseract OCR: recognizes text
- Screenshot OCR → OCR invocation: returns extracted data
- OCR invocation → Match statistics: supplies OCR data
- Assistant CLI → Match setup: collects match context
- Match statistics → Team perspective: provides team stats
- Match setup → Team perspective: identifies manager team
- Team perspective → Match analysis: feeds match analysis
- Match analysis → Question prompt: informs questions
- Match setup → Question prompt: adds match context
- Assistant CLI → Groq client: requests completion
- Question prompt → Groq client: provides prompt
- Groq client → Groq language model: sends chat request
- Groq language model → Groq client: returns completion
- Groq client → Conference content: provides response
- Conference content → Manager: presents content