restep 0.0.3

Reason-based RAG for Early-Stage Developers for Community-based Prospecting (RE-STEP)

RESTEP

Reason-based RAG for Early-Stage Developers for Community-based Prospecting (RE-STEP) is a GenAI prospecting tool to assess community-based factors in early-stage renewable-energy siting. It is targeted for early-stage developers in the renewable energy industry.

Requirements

Required packages

The following packages are required for restep:

• torch (to learn how to install, please refer to pytorch.org)

• transformers

• accelerate

• safetensors

• sentencepiece

• bitsandbytes

• requests

• PyPDF2

• python-docx

• langchain

• tqdm

To install these packages, you can do the following:

pip install torch transformers accelerate safetensors sentencepiece bitsandbytes requests PyPDF2 python-docx langchain tqdm

GPU requirements

You require at least one GPU to use restep.

VRAM requirements depend on factors like the length of the document.

However, at least 16GB of VRAM is recommended

huggingface access token

You will need a huggingface access token. To obtain one:

1. you'd first need to create a huggingface account if you do not have one.

2. Create and store a new access token. To learn more, please refer to huggingface.co/docs/hub/en/security-tokens.

3. Note: Some pre-trained large language models (LLMs) may require permissions. For more information, please refer to huggingface.co/docs/hub/en/models-gated.

4. Request for permission for meta-llama/Llama-3.1-8B-Instruct at https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct.

Installation

To install in python, simply do the following:

pip install restep

Quick Start

Here we provide a quick example on how you can execute restep by providing a link to the local government document and your huggingface API key.

from restep import restep

api_key="<your huggingface API key>"

url_to_local_government_document"https://www.co.marshall.in.us/egov/documents/1739816106_82896.pdf"

print(restep(url_to_local_government_document,api_key))



# This will return: {"Sentiment": "NEGATIVE", "Summary": "Opposition from residents and board members, citing concerns over property values, wildlife, and agricultural land, with some board members questioning the project's compliance with the ordinance and requesting additional conditions."}

How does RE-STEP work?

Our paper is currently under review at ACM BuildSys '25. Stay-tuned for updates if you are curious for more details.