https://github.com/siddhesh-agarwal/straindb-rag

An RAG Model based on StrainsDB by Kenneth Reitz.
https://github.com/siddhesh-agarwal/straindb-rag

langchain openai python rag-model streamlit

Last synced: 24 days ago
JSON representation

An RAG Model based on StrainsDB by Kenneth Reitz.

• Host: GitHub
• URL: https://github.com/siddhesh-agarwal/straindb-rag
• Owner: Siddhesh-Agarwal
• Created: 2024-03-30T15:20:58.000Z (10 months ago)
• Default Branch: main
• Last Pushed: 2024-11-20T07:41:12.000Z (3 months ago)
• Last Synced: 2024-11-20T09:02:22.370Z (3 months ago)
• Topics: langchain, openai, python, rag-model, streamlit
• Language: Jupyter Notebook
• Homepage: https://straindb.streamlit.app/
• Size: 69.9 MB
• Stars: 2
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# StrainDB RAG

An RAG Model based on [StrainsDB](https://strainsdb.org/). The Database for the project was provided by [Kenneth Reitz](https://github.com/kennethreitz).

___

## Steps involved

### Extracting and Validating the Data

- The SQLite DataBase consisted of 19 tables from which the `strains_strain` table was extracted.

  - **Solution**: extract data using SQLite module and create a `Strain` class (that inherits `pydantic.BaseModel`) for data validation

- Then the table's data cells consisted of lists in string format which needed to be converted back

  - **Solution**: `eval("[1, 2, 3]")` returns `[1, 2, 3]`

- The data needs to be saved.

  - **Solution**: dump data in JSON format.

### Data Tokenization

- The Data was embedded into ChromaDB (persistent client) using the `OpenAIEmbeddings` function.

  - **Solution**: `OpenAIEmbeddings` function from the `langchain_openai` package.

### Retrieval

- The data was retrieved from the ChromaDB using the `Chroma` class.