Faria Installer

Dependency installer for Faria, a document processing library with ML-powered layout detection and table extraction.

Quick Start

macOS / Linux

# IDP only (Document Processing)
curl -fsSL https://raw.githubusercontent.com/exto360-inc/faria-install/main/install.sh | bash -s -- --features idp

# Chat only (Conversational AI)
curl -fsSL https://raw.githubusercontent.com/exto360-inc/faria-install/main/install.sh | bash -s -- --features chat

# Both features
curl -fsSL https://raw.githubusercontent.com/exto360-inc/faria-install/main/install.sh | bash -s -- --features all

# Interactive mode (prompts for options)
curl -fsSL https://raw.githubusercontent.com/exto360-inc/faria-install/main/install.sh | bash

Windows (PowerShell)

# Download the installer
Invoke-WebRequest -Uri "https://raw.githubusercontent.com/exto360-inc/faria-install/main/install.ps1" -OutFile "install.ps1"

# Basic installation (without LLM)
.\install.ps1 -NoLLM

# Full installation (with LLM)
.\install.ps1 -WithLLM

Clone and Install

git clone https://github.com/exto360-inc/faria-install.git
cd faria-install
./install.sh --features idp

System Requirements

Supported Architectures

Features

IDP - Intelligent Document Processing (~630 MB)

Chat - Conversational AI (~535 MB)

Installation Options

macOS / Linux

--features LIST       Comma-separated list of features (idp, chat, all)
--install-dir PATH    Installation directory (default: ~/.faria)
--gpu                 Enable CUDA support (Linux only)
--help                Show help message

Windows (PowerShell)

-InstallDir PATH      Installation directory (default: %USERPROFILE%\.faria)
-GPU                  Enable GPU support (CUDA)
-WithLLM              Install LLM components
-NoLLM                Skip LLM components
-SkipTesseract        Skip Tesseract if already installed
-Help                 Show help message

Examples

# IDP with GPU support (Linux)
./install.sh --features idp --gpu

# Custom installation directory
./install.sh --features all --install-dir /opt/faria

# IDP only (backward compatible)
./install.sh --no-llm

Directory Structure

~/.faria/
├── bin/
│   └── llama-cli                         # LLM inference (Chat feature)
├── lib/
│   └── onnxruntime/
│       └── libonnxruntime.dylib          # ONNX Runtime library
└── models/
    ├── detr_layout_detection.onnx        # Layout detection (IDP)
    ├── nemotron_table_structure.onnx     # Table structure (IDP)
    └── qwen2.5-0.5b-instruct-q8_0.gguf   # LLM model (Chat)

Verification

./scripts/verify.sh

Usage

After installation, use the Faria Go module:

go get github.com/exto360-inc/faria

package main

import "github.com/exto360-inc/faria"

func main() {
    config := faria.DefaultConfig()
    client, err := faria.New(config)
    if err != nil {
        panic(err)
    }
    defer client.Close()

    // Process documents...
}

Uninstallation

./scripts/uninstall.sh

Or manually:

rm -rf ~/.faria

Documentation

• INSTALLATION.md - Manual installation, GPU configuration, environment variables
• TROUBLESHOOTING.md - Common errors and solutions (Python, CGO, Tesseract, OpenCV, MuPDF, ONNX Runtime)

License

Copyright (c) 2024 Exto360 Inc. All rights reserved. See LICENSE.