ROOT /

History

Enable CI/CD by adding .onedev-buildspec.yml

.mvn/wrapper	Loading last commit info...
src
.dockerignore
.gitignore
Dockerfile
README.md
mvnw
mvnw.cmd
pom.xml

README.md

MiruIQ Hub

A Quarkus-based REST API for intelligent document data extraction powered by LLMs. Upload documents, define extraction schemas, and get structured data back.

Features

Document Extraction - Upload PDFs/images and extract structured data using AI
Schema Management - Create, store, and reuse extraction schemas with $ref support
AI Schema Generation - Generate schemas from natural language descriptions or sample documents
Data Validation - Validate extracted data with text, numeric, and semantic validation strategies
Multi-tenant - JWT and API key authentication with per-user rate limiting

Tech Stack

Framework: Quarkus 3.19
Language: Java 17
Storage: Apache Paimon (data lake), MinIO/S3 (files), PostgreSQL (metadata)
AI: OpenAI-compatible API (GPT-4, local models via OpenRouter, etc.)
PDF Processing: Apache PDFBox

Quick Start

Prerequisites

Java 17+
Docker (for MinIO and PostgreSQL)
An OpenAI-compatible API key

Configuration

Set the following environment variables:

# Required - OpenAI Configuration
export OPENAI_API_KEY="your-api-key"
export OPENAI_BASE_URL="https://api.openai.com/v1"
export OPENAI_MODEL="gpt-4o-mini"

# Optional - Database (defaults to local Docker)
export JDBC_URL="localhost:5432"
export JDBC_DATABASE="document_store"
export JDBC_USERNAME="paimon"
export JDBC_PASSWORD="paimon"

# Optional - S3/MinIO (defaults to local MinIO)
export S3_ENDPOINT="http://localhost:9000"
export S3_BUCKET="media-store"
export AWS_ACCESS_KEY_ID="minioadmin"
export AWS_SECRET_ACCESS_KEY="minioadmin"

Run in Development Mode

./mvnw quarkus:dev

The API will be available at http://localhost:8082.

Run with Docker

docker build -t miruiq-hub .
docker run -p 8082:8082 \
  -e OPENAI_API_KEY="your-key" \
  -e OPENAI_BASE_URL="https://api.openai.com/v1" \
  -e OPENAI_MODEL="gpt-4o-mini" \
  miruiq-hub

API Reference

Authentication

All endpoints require authentication via:

X-API-Key header, or
Authorization header (Bearer token)

Endpoints

Extractions

Method	Endpoint	Description
`POST`	`/extractions`	Create extraction job (multipart: schema + documents)
`GET`	`/extractions/{id}`	Get extraction status and results
`GET`	`/extractions?label={label}`	Find extraction by label
`POST`	`/extractions/{id}/validate`	Validate extraction results
`GET`	`/extractions/{id}/validation`	Get validation results

Schemas

Method	Endpoint	Description
`POST`	`/schemas`	Create schema template
`GET`	`/schemas`	List all schemas
`GET`	`/schemas/{id}`	Get schema by ID
`GET`	`/schemas/{id}/resolved`	Get schema with `$ref` expanded
`GET`	`/schemas/by-name/{name}`	Get schema by name
`PUT`	`/schemas/{id}`	Update schema
`DELETE`	`/schemas/{id}`	Delete schema
`POST`	`/schemas/generate`	Generate schema from description/images

Validation

Method	Endpoint	Description
`POST`	`/validation/validate/{requestId}`	Validate by request ID
`POST`	`/validation/validate/by-label/{label}`	Validate by label
`POST`	`/validation/validate/multi`	Validate multiple extractions
`GET`	`/validation/results/{requestId}`	Get validation results

Example: Create an Extraction

curl -X POST http://localhost:8082/extractions \
  -H "X-API-Key: your-api-key" \
  -F 'schema={
    "type": "object",
    "properties": {
      "invoice_number": { "type": "string" },
      "total_amount": { "type": "number" },
      "date": { "type": "string", "format": "date" }
    }
  }' \
  -F "documents=@invoice.pdf" \
  -F "label=invoice-001"

Example: Generate a Schema with AI

curl -X POST http://localhost:8082/schemas/generate \
  -H "X-API-Key: your-api-key" \
  -F "description=Extract customer name, order items with quantities and prices, and total amount" \
  -F "images=@sample-receipt.jpg"

Testing

# Run unit tests only (default)
./mvnw test

# Run unit + LLM tests (requires running LLM server)
./mvnw test -Pwith-llm

# Run all tests (requires LLM + Flink pipeline)
./mvnw test -Pfull

Configuration Reference

Property	Default	Description
`HTTP_PORT`	`8082`	Main API port
`PDF_MAX_PAGES`	`100`	Maximum pages per PDF
`MAX_FILES_PER_REQUEST`	`10`	Max documents per extraction
`RATE_LIMIT_PER_USER_PER_MINUTE`	`1000`	Rate limit per user
`HTTP_MAX_BODY_SIZE`	`100M`	Maximum upload size

License

Proprietary - MiruIQ