TextGrapher

TextGrapher is a Python project that converts unstructured Chinese text into a structured semantic graph. It extracts high-frequency terms, keywords, named entities, and subject-verb-object phrases from an input document. The system then organizes these elements into connected nodes and relationships for visual inspection. Generated results are saved as an HTML graph that can be opened in a browser. The repository includes parsing, keyword extraction, graph construction, and visualization scripts. Example use cases cover news and public-event narratives where key actors and actions are difficult to scan quickly. The project is an experimental NLP demonstration, and its accuracy depends on the quality of entity recognition and information extraction.