KGQA_HLM

KGQA_HLM

Visualization and Q&A System of Character Relationships in "Dream of Red Mansions" based on knowledge graph

• For details, please see http://chizhunlp.com

File tree:

1. app.py is the main entrance to the entire system
   
2. The templates folder is an HTML page
   |-index.html Welcome interface
   |-search.html Search for character relationship page
   |-all_relation.html All character relationship pages
   |-KGQA.html Character Relationship Q&A Page
   
3. Static folder stores css and js, which are files with page styles and effects
   
4. The raw_data folder is a triple file that has data processing
   
5. The neo_db folder is a knowledge graph building module
   |-config.py configuration parameters
   |-create_graph.py Create knowledge graph and create graph database
   |-query_graph.py Query of knowledge graph
   
6. KGQA folder is a Q&A system module
   |-ltp.py word participle, part-of-speech annotation, naming entity recognition
   
7. The spider folder is a crawler module
   |- get_*.py is the code that crawled the character information before. It has produced images and json and can be executed without any need to be executed.
   |-show_profile.py is the code that calls character profiles and graphs to display on the front end

---

Deployment steps:

• 0. Install the required library to execute pip install -r requirement.txt
  
• 1. Download the neo4j image database first and match the environment (note that neo4j requires jdk8). Modify the configuration file config.py in the neo_db directory and set the account and password of the graph database.
  
• 2. Switch to the neo_db directory and execute python create_graph.py to create a knowledge graph
  
• 3. Go here to download the ltp model. Introduction to ltp
  
• 4. In the KGQA directory, modify the storage directory of the ltp model file in ltp.py
  
• 5. Run python app.py, open localhost:5000 in the browser to view it
  

Overall system flow chart:

Website example:
Welcome interface

Main interface