Data can be found here

Execute run.py to obtain the RAG results.

Chunking

We currently support the following chunking methods in split_text.py:

• RecursiveCharacterTextSplitter
• CharacterTextSplitter
• SemanticChunker

Retrieval Stage

In the retrieval stage, we support (see retrieval_utils.py):

• Retrieving similar documents to a given query
• Using the Hyde method, which first generates a synthetic document that is then compared with the index

Evaluation

For evaluation, we also utilize the following objects in metrics.py:

• deepeval GEval
• FaithfulnessMetric

Set Environment Variables:

pip install python-dotenv

In the root directory of your project, create a file named .env and add the following variables:

LANGCHAIN_TRACING_V2=true LANGCHAIN_API_KEY="YOUR_LANGCHAIN_KEY" LANGCHAIN_ENDPOINT="https://api.smith.langchain.com" LANGCHAIN_PROJECT="YOUR_PROJECT_NAME" OPENAI_API_KEY="YOUR_OPENAI_KEY"

Replace each placeholder (e.g., YOUR_LANGCHAIN_KEY, YOUR_PROJECT_NAME, and YOUR_OPENAI_KEY) with your actual values.

By configuring these environment variables, you can monitor metrics like latency usage, and other performance indicators through the LangSmith dashboard.