OpenEOcraft

OpenEOcraft offers a robust R framework designed for the development and deployment of openEO API applications. This package simplifies the process of creating RESTful openEO web services with its user-friendly and flexible interface. Built on Plumber, OpenEOcraft prioritizes ease of use, scalability, and adaptability.

Quick Start with Docker Hub

The easiest way to get started with OpenEOcraft is by using the pre-built Docker image from Docker Hub:

docker run -p 8000:8000 brianpondi/openeocraft:latest

This will start the OpenEOcraft server and make it available at http://localhost:8000.

Deployment Options

Local Deployment with R

If you have R installed and the relevant packages like plumber, sits, torch, jsonlite, etc., you can run the server locally.

1. Clone the repository:

git clone https://github.com/Open-Earth-Monitor/openeocraft.git
cd openeocraft

2. Run the server:

Rscript docker/server.R

Development with Docker Compose

For development or custom modifications, you can build and run using Docker Compose:

git clone https://github.com/Open-Earth-Monitor/openeocraft.git
cd openeocraft
docker-compose up

Common Docker Compose Commands:

• Run in detached mode: docker-compose up -d
• Stop the containers: docker-compose down
• Force rebuild and restart: docker-compose up --build --force-recreate --no-deps -d
• Rebuild with no cache: docker-compose build --no-cache && docker-compose up

Example Workflows

Authentication

Default credentials for the development server:

• Username: user
• Password: password

R Example: TempCNN Model Training

Note: Training data is located in inst/demo-paper-2025/data/ within the repository.

library(openeo)
connection <- connect("http://127.0.0.1:8000")
login(user = "user", password = "password")

# Retrieve available processes from the backend
p <- processes()
# Load Sentinel-2 Data
datacube <- p$load_collection(
    id = "mpc-sentinel-2-l2a",
    spatial_extent = list(
        west = -63.33,
        south = -12.03,
        east = -62.43,
        north = -11.13,
        crs = 4326
    ),
    temporal_extent = c("2022-01-01", "2022-12-31")
)
# Regularize the Sentinel-2 data
datacube <-  p$cube_regularize(data = datacube,
                               period = "P16D",
                               resolution = 600)
# NDVI calculation and merge with the original datacube
datacube_ndvi <- p$ndvi(
    data = datacube,
    red = "B04",
    nir = "B08",
    target_band = "NDVI"
)
datacube <- p$merge_cubes(datacube, datacube_ndvi)
# Load training data from the package installation directory
data_deforestation_rondonia <- readRDS("inst/demo-paper-2025/data/samples_deforestation_rondonia.rds")
# Initialize TempCNN model
tempcnn_model_init <-  p$mlm_class_tempcnn(optimizer = "adam",
                                           learning_rate = 0.0005,
                                           seed = 42)
# Model Training
tempcnn_model <-  p$ml_fit(
    training_set = jsonlite::serializeJSON(data_deforestation_rondonia),
    target = "label"
)
# Apply the trained model to make a prediction
datacube <-  p$ml_predict(datacube, tempcnn_model)
# Save the Model
model <- p$save_ml_model(
    tempcnn_model,
    name = "tempcnn_model_2022_rondonia",
    tasks = list("classification"),
    options = list("accelerator" = "macos-arm", "framework" = "Torch for R")
)
# Save the prediction result
ml_job <- p$save_result(data = datacube, format = "GTiff")

# Run the job
job <- create_job(graph = ml_job,
                  title = "Deforestation Prediction in Rondonia",
                  description =
                      "Using TempCNN model to predict deforestation in Rondonia")
job <- start_job(job)

# Display job information
status <- describe_job(job)
print(paste("Job status:", status$status))

# Download results when complete
if (status$status == "finished") {
    results <- download_results(job, "./output")
    print("Results downloaded to ./output directory")
}

Python Example: Interacting with EO Data Cubes

Note: This example requires the custom Python client. Install it locally:

pip install git+https://github.com/PondiB/openeo-python-client.git

import openeo

# Connect to the OpenEO backend
connection = openeo.connect("http://127.0.0.1:8000")
connection.authenticate_basic("user", "password")

# Check available collections and processes
print("Available collections:", connection.list_collection_ids())
print("\nAvailable processes:", [p["id"] for p in connection.list_processes()])

# Define the area of interest (Rondonia, Brazil)
bbox = {
    "west": -63.33, "south": -12.03,
    "east": -62.43, "north": -11.13,
    "crs": 4326
}
temporal_extent = ["2022-01-01", "2022-12-31"]

# Load and preprocess the satellite data
datacube = connection.load_collection(
    "mpc-sentinel-2-l2a",
    spatial_extent=bbox,
    temporal_extent=temporal_extent,
    bands=["B02", "B03", "B04", "B08", "SCL"]
)

datacube = datacube.process(
    process_id="cube_regularize",
    arguments={
        "data": datacube,
        "period": "P1M",  #16 days or 1 month  regularization 
        "resolution": 300
    }
)

datacube_ndvi = datacube.ndvi(red = "B04", nir = "B08", 
                                    target_band = "NDVI")
datacube = datacube.merge_cubes(datacube_ndvi)

# Load training data from the package installation directory
serialized_data = connection.readRDS("inst/demo-paper-2025/data/samples_deforestation_rondonia_1M.rds")



# Initialize and Train the Model
tempcnn_model_init = connection.mlm_class_tempcnn(
    optimizer="adam",
    learning_rate=0.0005,
    seed=42
)

tempcnn_model = tempcnn_model_init.fit(
    training_set=serialized_data,
    target="label"
)

# Make Predictions
datacube =  tempcnn_model.predict(datacube)

# Save the trained model for future use.
tempcnn_model.save_ml_model(name ="tempcnn_rondonia", tasks=["classification"], options={"accelerator":"macos-arm", "framework":"Torch for R"})

# Save and Execute Results
result = datacube.save_result(
    format="GeoTiff"
)

job = result.create_job(
    title="Deforestation Prediction in Rondonia",
    description="Using TempCNN model to predict deforestation in Rondonia"
)

job.start_and_wait()
results = job.get_results()
results.download_files("data/output")