Fairfax Construction Assistant

link

A POC (Proof of Concept) AI chatbot web application desinged to interact with construction contractors and provide helpful information regarding fairfax county code and manuls.

Features

• Interactive chat web interface
• Uses the LangGraph framework for natural language processing.
• Uses Rag and vectorstore to retrieve relevant information about fairfax county code and manuls
• Serverless backend API services

Requirements

• Python 3.12
• Required Python libraries (see requirements.txt for details).
• Google Cloud Vertex API Key
• Pinecone API Key
• Groq API Key

Prerequisites

Google Cloud Vertex API Key:

1. Create a new project in the Google Cloud Console.
2. Setup credentials and download credentials into ./service directory

Groq

1. Create Groq API key. You can retrieve the key from the GROQ Console.

Pinecone

1. Create a Pinecone cluster. You can retrieve the API key from the Pinecone Console.

Populate Vectorstore

1. Crawl the fairfax county website using the script python ./scripts/vectorstores/my_crawler.py
2. This will download all the fairfax county information in service/consultation_company/vectorstores/web_scraping_data
3. Then run python ./scripts/vectorstores/pinecone_loader.py to load the documents into the pinecone vectorstore.

Installation

1. Clone the repository:

   git clone https://github.com/wonhs91/fairfax-construction-assistant.git
   cd fairfax-construction-assistant

2. Create a virtual environment (optional but recommended):

   python -m venv .venv
   source venv/bin/activate   # On Windows: .venv\Scripts\activate

3. Client setup

   1. Install dependencies:

      • from the workspace directory

        cd client
        npm install

   2. Set up the .env file:

      • Create a .env file in the ./client directory of the project:

      touch .env

      • Add the server's endpoint

      NEXT_PUBLIC_API_URL={server_endpoint}
      

      replace server_endpoint with actual server endpoint from the server setup

   3. Run the client application

      cd client
      npm run dev

4. Server setup

   1. Install dependencies:
      • from the workspace directory

      cd service
      pip install -r requirements.txt

   2. Set up the .env file:
      • Create a .env file in the ./service directory of the project:

      cp .env_sample .env

      • Add the appropriate environment variables
   3. Run the server

      cd service
      python app.py

Pushing dockerfile to ECR for Lambda deployment

1. Build docker image

docker build -t ffx-consultant-lambda .

2. Test locally

docker run -p 9000:8080 ffx-consultant-lambda

3. test at `http://localhost:9000

4. Login to ECR

aws ecr get-login-password --region <your-region> | docker login --username AWS --password-stdin <account-id>.dkr.ecr.<your-region>.amazonaws.com

5. Create a repository in ECR:

aws ecr create-repository --repository-name <repository-name>

6. Tag the image

docker tag ffx-consultant-lambda:latest <account-id>.dkr.ecr.<your-region>.amazonaws.com/<repository-name>:latest

7. Push the image

docker push <account-id>.dkr.ecr.<your-region>.amazonaws.com/<repository-name>:latest

Sample Chat