How frozsgy/personality-analysis-framework works
Project type and purpose
• This is a full-stack web framework (“Personality Analysis Framework”) that ingests Turkish tweets, processes them via NLP and vector models, then predicts Big-Five personality scores.
• It exposes:
• A React-based frontend (in /frontend)
• A Flask REST backend (main.py, router.py, web/service.py)
• Two auxiliary microservices: Zemberek (gRPC) and Word2Vec
• MySQL/MariaDB for persistence
• A CLI wrapper (main.py) and training scripts
File structure
• Top-level directories:
• frontend/ – React UI code
• twitter/ – tweet retrieval abstractions
• vector/ – boundary generation & Word2Vec wrappers
• predictors/ – TPOT-based model training & inference
• utils/ – preprocessing, Big-Five calculators, stopword lists
• web/ – Flask DB connectors (web/db.py), service endpoints (web/service.py), plotting
• zemberek/ – protobufs and JAR for gRPC NLP service
• data/ – tweet storage
• Scripts to stand up services: run_zemberek.sh, run_word2vec.sh, train_model.sh
• Config files: requirements.txt, setup.py, tsconfig.json, package.json
README for additional insights
• Four distinct runtime services:
1. Zemberek gRPC server (Java NLP)
2. Word2Vec vector service (Python)
3. Flask REST API (Python backend)
4. React frontend (Node/React)
• Optional CLI mode via main.py
• Persistent storage via MySQL/MariaDB