https://github.com/ilikepizza2/graph-database
https://github.com/ilikepizza2/graph-database
Last synced: 2 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/ilikepizza2/graph-database
- Owner: Ilikepizza2
- Created: 2024-03-02T22:16:44.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-03-04T09:44:07.000Z (about 1 year ago)
- Last Synced: 2024-03-04T16:32:51.098Z (about 1 year ago)
- Language: Jupyter Notebook
- Size: 93.8 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Graph database with CRUD and save/load operations
### [Link to colab notebook](https://colab.research.google.com/drive/1VqDbCCaUuNtWBxmTBOz7lotii17BFwS4?usp=sharing)
This project demonstrates building and exploring knowledge graphs (KGs) using either of the two models: Rebel(seq2seq) and Gemini(Google LLM).
## Functionalities
* **KG Construction:** Build KGs using either:
* **Provided Text:** Define your own text for analysis.
* **Gemini-Generated Text:** Let Gemini generate text for KG construction.
* **Model Selection:** Choose between Rebel and Gemini for relation extraction.
* **Querying:** Perform various KG queries:
* Keyword search
* Semantic BFS (breadth-first search)
* Semantic document search based on embeddings
* **Visualization:** Convert the KG to a network graph for visual exploration (requires `pyvis` library).
## Running the Project
1. **Access the Colab Notebook:** [Link to your Colab notebook here]
2. **Install Libraries:** Follow the instructions in the notebook to install required libraries (e.g., `rebel-api`, `pyvis`).
3. **Obtain API Keys:** If using Rebel, acquire an API key from https://huggingface.co/Babelscape/rebel-large and set the `REBEL_API_KEY` environment variable in Colab. For Gemini, obtain your gemini api key and set the `GOOGLE_API_KEY`.
4. **Run Cells:** Execute the notebook cells to construct, query, and visualize your KG.
## Additional Notes
* Refer to the notebook for detailed explanations and customization options.
* Experiment with different text inputs, models, and query types to explore your data.
* Consider installing `pyvis` locally for offline visualization.
This project provides a starting point for building and interacting with KGs using LLMs in Google Colab. Feel free to adapt and extend it for your specific needs.