Procurement Suite

A comprehensive procurement management system with AI-powered document intake, API services, and a modern web interface.

Installation

Prerequisites

• Docker and Docker Compose
• OpenAI API Key

Setup Steps

1. Configure environment variables

   Copy all .env.template files from the individual apps to .env:

   cp apps/agent/.env.template apps/agent/.env
   cp apps/api/.env.template apps/api/.env
   cp apps/ui/.env.template apps/ui/.env

2. Set OpenAI API Key

   Edit apps/agent/.env and set your OpenAI API key:

   OPENAI_API_KEY=your_api_key_here

3. Build and run the application

   docker-compose up --build

4. Access the application

   • UI: http://localhost:8080
   • Procurement API: http://localhost:8081
   • Intake Agent API: http://localhost:8082

Architecture

System Components

graph TB
    UI[Procurement Suite UI<br/>Port 8080]
    AgentAPI[Agent API<br/>Port 8082]
    ProcAPI[Procurement API<br/>Port 8081]

    UI -->|Upload PDF| AgentAPI
    UI -->|Submit Request| ProcAPI
    UI -->|Fetch Requests| ProcAPI

    AgentAPI -->|AI Processing<br/>OpenAI| AgentAPI
    ProcAPI -->|Store & Retrieve| DB[(In-Memory Repository)]

    style UI fill:#e1f5ff
    style AgentAPI fill:#fff4e1
    style ProcAPI fill:#e8f5e9
    style DB fill:#f3e5f5

Loading

HTTP Request Flow

sequenceDiagram
    autonumber
    actor User
    participant UI as Procurement Suite UI
    participant Agent as Agent API
    participant API as Procurement API

    Note over User,API: Document Upload & Processing
    User->>UI: Upload PDF document
    UI->>Agent: POST /agent/intake<br/>(PDF file)
    Agent->>Agent: Extract data using AI
    Agent-->>UI: Return ProcurementRequestCreate<br/>(extracted data)
    UI->>UI: Pre-fill form with extracted data

    Note over User,API: Request Submission
    User->>UI: Review & submit form
    UI->>API: POST /intake/request<br/>(JSON payload)
    API->>API: Validate commodity group
    API->>API: Create request with status: OPEN
    API-->>UI: Return request ID
    UI->>UI: Display success message

    Note over User,API: Request Management
    User->>UI: View requests list
    UI->>API: GET /intake/requests
    API-->>UI: Return all requests

    User->>UI: Update request status
    UI->>API: PATCH /intake/requests/{id}/status
    API->>API: Update status<br/>(OPEN → IN_PROGRESS → CLOSED)
    API-->>UI: Return updated request
    UI->>UI: Refresh request list

Loading

Key Endpoints

Agent API (Port 8082)

• POST /agent/intake - Process PDF document using Intake Agent to extract procurement data

Procurement API (Port 8081)

• POST /intake/request - Create new procurement request
• GET /intake/requests - Get all procurement requests
• GET /intake/requests/{id} - Get request by ID
• PATCH /intake/requests/{id}/status - Update request status
• GET /intake/commodity_groups - Get available commodity groups

Technology Stack

• Frontend: React + TypeScript + Material-UI + Vite
• Backend APIs: FastAPI + Python
• AI/ML: OpenAI API (document processing)
• Containerization: Docker + Docker Compose
• PDF Processing: react-pdf (frontend), AI extraction (backend)

Development

Project Structure

procurement-suite/
├── apps/
│   ├── agent/          # AI-powered intake agent service
│   ├── api/            # Procurement management API
│   └── ui/             # React frontend application
├── docker-compose.yaml # Multi-container orchestration
└── README.md

Data Flow

1. Document Upload: User uploads PDF → UI sends to Agent API
2. AI Extraction: Agent API processes PDF using OpenAI → Returns structured data
3. Form Pre-fill: UI populates form fields with extracted data
4. Request Creation: User submits → UI sends to Procurement API → Stored with status OPEN
5. Status Updates: Requests can be updated through workflow: OPEN → IN_PROCESS → CLOSED