Agent1 is a document processing pipeline that uses AWS services to analyze and extract information from various document types including academic transcripts, licenses, and immigration documents.
- Mansi — AWS Infrastructure & Orchestration
- Nandini — Data Models, Manifest & DB Indexing
- Shreya — CV / Grouping / API Schema
- Pradyumn — ML Classifier & Tests
# Install Python 3.11+
python --version
# Install dependencies
pip install -e .
pip install boto3 pdf2image reportlab opencv-python-headless
# Install poppler (for PDF processing)
# Ubuntu: sudo apt-get install poppler-utils
# macOS: brew install poppler
# Windows: Download from GitHub releases# Test with sample documents
python run_tests.py
# Test with specific file
python run_tests.py --pdf agents/agent1/samples/I94.pdf# Setup S3 and IAM policies
python agents/agent1/scripts/setup_s3_policies.py
# Configure Lambda triggers
python agents/agent1/scripts/configure_s3_triggers.py# Upload samples to S3 for processing
python agents/agent1/scripts/upload_samples_to_s3.py --monitor- S3 Buckets: agent1-incoming, staging, output, quarantine
- Step Functions: Agent1Pipeline with 11 processing states
- Lambda: BucketNotificationsHandler for S3 triggers
- Textract: OCR and document analysis
- main (protected) → integration → agent1/base → feature/
- PR flow:
feature → agent1/base → integration → main - Hotfixes:
main → hotfix → main → down-merge
# Allow scripts
Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope CurrentUser
# Setup environment
scripts\win\dev-setup.ps1
# Activate
scripts\win\activate.ps1# Create virtual environment
python -m venv .venv
source .venv/bin/activate
# Install dependencies
pip install -e .- API Reference:
agents/agent1/docs/API.md - AWS Integration:
agents/agent1/docs/AWS_INTEGRATION_GUIDE.md - Deployment:
agents/agent1/docs/DEPLOYMENT_GUIDE.md