FTW Inference App

A Vue 3 application for running inference on satellite imagery using the FTW API.

Running the Application Locally

Step 1: Clone the Repositories

# Clone both repositories
git clone https://github.com/fieldsoftheworld/ftw-inference-app.git
git clone https://github.com/fieldsoftheworld/ftw-inference-api.git

Step 2: Set Up the API

The API is always being updated, so it is best to follow the directions directly in the API repository.

Prerequisites:

• Docker (required for local DynamoDB)
• Pixi (package manager)

# Navigate to API directory
cd ftw-inference-api

# Copy environment configuration
cp .env.example .env

# Configure local DynamoDB in .env file (uncomment these lines):
# DYNAMODB__DYNAMODB_ENDPOINT="http://localhost:8001"
# AWS_ACCESS_KEY_ID="fake_key_id"
# AWS_SECRET_ACCESS_KEY="fake_secret_key"

# Download precomputed models for api endpoint
# Store the .ckpt files from <https://github.com/fieldsoftheworld/ftw-baselines/releases/tag/v1>
# in this directory

# Start local DynamoDB
pixi run dynamodb-local

# In a new terminal, start the API server
pixi run start

The API will be available at http://localhost:8000

Step 3: Set Up the Frontend App

# Open a new terminal and navigate to the app directory
cd ../ftw-inference-app

# Install dependencies
npm install

# Create environment file
echo "VITE_API_BASE_URL=http://127.0.0.1:8080/v1/" > .env.development

# Start the development server
npm run dev

Optionally, to get access to model results from a local ftw-inference-api instance, run

npx serve --cors ../ftw-inference-api/server/data/results/

and add the VITE_FTW_INFERENCE_OUTPUT_URL environment variable pointing to that dev server:

echo "VITE_FTW_INFERENCE_OUTPUT_URL=http://127.0.0.1:3000/" > .env.development

The app will be available at http://localhost:5173

Alternative API Setup

If you prefer not to use conda, you can install the API manually:

Prerequisites:

• Python 3.11 or 3.12
• GDAL 3.11 or later with libgdal-arrow-parquet

Then follow the same steps as above, skipping the conda environment creation.

Additional Commands

Production Build

# Build for production
npm run build

Type Checking

# Run type checking
npm run type-check

Linting

# Run ESLint
npm run lint

Prerequisites

• Node.js (v16 or higher)
• npm (v7 or higher)