Skip to main content

This project I use a lot for workshops, it contains some utils for splitters, tokenizers, and a weaviate client that I reuse a lot

Project description

RAG4P - Retrieval Augmented Generation for Python

Welcome to the repository for our project RAG4P.org. This project is a Python implementation of the Retrieval Augmented Generation framework. It is a framework that is simple to use and understand. But powerful enough to extend for your own projects.

Setting up your environment

Python

We encourage you to use a python environment manager. Poetry makes it easy to use multiple python versions and packages. where you can switch versions per project. Read this Poetry documentation page to learn how to set up your environment. No poetry installed? Read this page to install it for your environment. Poetry installation

Setting the right version of python for the project

poetry env use 3.10

Install dependencies

poetry install

Run the project

poetry run python rag4p/app_step1_chunking_strategy.py

No poetry

Setup your venv

python3 -m venv venv
source venv/bin/activate

Install dependencies

pip install -r poetry-requirements.txt

Loading API keys

We try to limit accessing Large Language Models and vector stores to a minimum. You do not need an LLM or vector store to learn about all the elements of the Retrieval Augmented Generation framework, except for the generation part. In the workshop we use the LLM of Open AI, which is not publicly available. We will provide you with a key to access it, if you don't have your own key.

Please use this key for the workshop only, and limit the amount of interaction, or we get blocked for exceeding our limits. The API key is obtained through a remote file, which is encrypted. Of course you can also use your own key if you have it.

Environment variables

The easiest way to load the API key is to set an environment variable for each required key. In Python we prefer the file .env.properties in the root of the project with the following properties:

OPENAI_API_KEY=sk-...
WEAVIATE_API_KEY=...
WEAVIATE_URL=...

If you do not have your own key, you can load ours. The key is stored in a remote location. You need the .env.properties file in the root of the project with the following line:

SECRET_KEY=...

This secret key is used to decrypt the remote file containing the API keys. We will provide the value for this key during the workshop.

Using Ollama

There is a simple way to run a Language Model on your local machine. Depending on your machine and the chosen model, it runs fast. I am not going in to much details on how to install it, but you can find the installation instructions on the Ollama Downloads page.

At the moment we prefer the model Phi 3. You can learn more about the model on the Ollama Models page. A lot of other models are available as well. You can try them out yourself. Make sure you pull the model first before you can use it. You can also use Ollama for the embeddings. We advice to pull the model nomic-embed-text for this purpose.

ollama pull phi3
ollama pull nomic-embed-text

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

rag4p-0.8.1.tar.gz (33.3 kB view details)

Uploaded Source

Built Distribution

rag4p-0.8.1-py3-none-any.whl (64.4 kB view details)

Uploaded Python 3

File details

Details for the file rag4p-0.8.1.tar.gz.

File metadata

  • Download URL: rag4p-0.8.1.tar.gz
  • Upload date:
  • Size: 33.3 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.8.3 CPython/3.12.3 Darwin/23.5.0

File hashes

Hashes for rag4p-0.8.1.tar.gz
Algorithm Hash digest
SHA256 3b64aae8c0ff6190b53cd9337456742975092c9dc7480c287e47a1761c54bddd
MD5 5d34290f2994a9aa0a16790667ac04b4
BLAKE2b-256 898b0229edadb8539ce039b511ca313e35131b54e90a5efb9ecc31a99591825d

See more details on using hashes here.

File details

Details for the file rag4p-0.8.1-py3-none-any.whl.

File metadata

  • Download URL: rag4p-0.8.1-py3-none-any.whl
  • Upload date:
  • Size: 64.4 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.8.3 CPython/3.12.3 Darwin/23.5.0

File hashes

Hashes for rag4p-0.8.1-py3-none-any.whl
Algorithm Hash digest
SHA256 15782480820050cce6ce0e21e076699bc443f6d17afa2c78e585d1e52d0ce287
MD5 1b658ff624df3b1020b73b5a3181a234
BLAKE2b-256 7f9d29a5d0aaa93e4dfb5b64913fcaac78f3f0d9a16f0fed2ddf77fa3780407e

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page