How vinbdigroupone/vietnamese-text-summarization works
Project type and purpose
• This is a full-stack, end-to-end Vietnamese text-summarization system.
• The root README calls it “vietnamese-text-summarization.” You have two summarization engines (extractive and abstractive) plus a Vue.js SPA front end.
File structure
• Top-level directories:
• Abstractive_method – Python code for the neural abstractive summarizer (train, dataset loading, model definition, evaluation).
• Extractive_method – Notebook-and-script based extractive summarizer, including a server entrypoint (server.py) and various methods (TF-IDF, centroid, PhoBERT).
• Notebooks – experiment / prototype Jupyter notebooks; these support development but are not in the production pipeline.
• client – a Vue.js single-page application (SPA) with src/, components/, views/, router/, public assets.
• Key files: requirements.txt (Python dependencies), client/package.json (JS dependencies).
README for additional insights
• No mention of external databases or microservices orchestration, so backend likely runs as two separate Python services or one unified Flask/FastAPI app with two modules.
Detail without overthinking
• Break each major block into as many subcomponents as makes sense (e.g., inside Abstractive Service: dataset loader, config, model trainer, evaluator).
• Entrypoints (main.py, server.py, client/main.js).