Skip to content

Crop

Goal

Create a container and run the crop step in the container image tagged and built in the previous step.

Lab

This step has a dedicated lab available at /workspace/mastering-app-package/practice-labs/Containers/crop.ipynb

Container

Each step has its own recipe to build the container image.

The crop step container image recipe is:

crop/Dockerfile
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
FROM docker.io/library/python:3.11-slim@sha256:91e9d01cf4bd56be7128c603506b6fe367ef7506f9f2ad8f3a908aeec8941bb9

# System deps needed by rasterio wheels
RUN apt-get update && \
    apt-get install -y --no-install-recommends libexpat1 && \
    rm -rf /var/lib/apt/lists/*

RUN pip install --no-cache-dir "rasterio==1.4.3" "click==8.2.1" "pystac==1.13.0" "loguru==0.7.3" "pyproj==3.7.2" "shapely==2.1.1" && \
    python -c "import rasterio"

ADD app.py /app/app.py

ENTRYPOINT []

Build the container image with:

terminal
1
2
3
export WORKSPACE=/workspace/mastering-app-package

podman build --format docker -t localhost/crop:latest ${WORKSPACE}/water-bodies/command-line-tools/crop

How to run a step in a container

We'll use podman container engine (docker is also fine).

The command to run the crop step in the container is:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
podman run \
    -i \
    --userns=keep-id \
    --mount=type=bind,source=/workspace/mastering-app-package/runs,target=/runs \
    --workdir=/runs \
    --read-only=true \
    --user=1001:100 \
    --rm \
    --env=HOME=/runs \
    --env=PYTHONPATH=/app \
    localhost/crop:latest \
    python \
    -m \
    app \
    --aoi \
    "-121.399,39.834,-120.74,40.472" \
    --band \
    green \
    --epsg \
    "EPSG:4326" \
    --input-item \
    https://earth-search.aws.element84.com/v0/collections/sentinel-s2-l2a-cogs/items/S2B_10TFK_20210713_0_L2A

Let's break down what this command does:

  • podman run: This is the command to run a container.
  • -i: This flag makes the container interactive, allowing you to interact with it via the terminal.
  • --userns=keep-id: It instructs podman to keep the user namespace ID. --mount=type=bind,source=/workspace/mastering-app-package/runs,target=/runs: This option mounts a directory from the host system to the container. In this case, it mounts the /workspace/mastering-app-package/runs directory on the host to the /runs directory inside the container.
  • --workdir=/runs: Sets the working directory inside the container to /runs.
  • --read-only=true: Makes the file system inside the container read-only, meaning you can't write or modify files inside the container.
  • --user=1001:100: Specifies the user and group IDs to be used within the container.
  • --rm: This flag tells podman to remove the container after it has finished running.
  • --env=HOME=/runs: Sets the HOME environment variable inside the container to /runs.
  • --env=PYTHONPATH=/app: Sets the PYTHONPATH environment variable inside the container to /app.
  • localhost/crop:latest: This is the name of the container image that you want to run. It's pulling the image from the local container registry with the name "crop" and the "latest" tag.
  • python -m app: This is the command to run inside the container. It runs a Python module named "app."
  • --aoi "-121.399,39.834,-120.74,40.472": This provides command-line arguments to the Python module. It specifies the area of interest (AOI) as a bounding box.
  • --band green: Specifies the band to be extracted from the Sentinel-2 acquisition. In this case, it's the "green" band.
  • --epsg "EPSG:4326"``: Specifies the EPSG code, which defines the coordinate system used for theaoi` command-line argument.
  • --input-item ...: Specifies the input STAC item URL. This particular URL points to a Sentinel-2 image hosted on AWS Earth Search.

Steps

Clean-up the /workspace/mastering-app-package/runs folder: 

rm -fr /workspace/mastering-app-package/runs/*

Run the command to run the crop step with the green band in the container with:

sh ${WORKSPACE}/scripts/podman-crop-green.sh

Run the command to run the crop step with the nir band in the container with:

sh -x ${WORKSPACE}/scripts/podman-crop-nir.sh

Expected outcome

The folder /workspace/mastering-app-package/runs contains: 

(base) jovyan@coder-mrossi:~/runs$ tree .
.
├── crop_green.tif
└── crop_nir.tif

0 directories, 2 files