TLDR; click here to view an interactive map displaying a recent Sentinel-2 image over a random populated place*! Refresh for a new map.
The code contained here:
- loads a GeoJSON file and randomly selects a feature
- searches Element 84's EarthSearch STAC Catalog for the most recent Sentinel-2 L2A scenes that intersect the selected geometry
- creates a
folium
map with a tile layer displaying a Cloud Optimized GeoTiff hosted in the AWS Registry of Open Data served via Development Seed's public COG tiler.
There is a CLI to generate a standalone HTML file and a FastAPI app for dynamically generating maps (docs).
The FastAPI app is deployed to an AWS Lambda function using GitHub Actions.
This repo uses PDM for dependency management. After cloning this repo, run pdm install --no-self
from the root directory to create a virtual environment with the required dependencies. Run eval $(pdm venv activate)
to activate the environment in your session.
Set $PYTHONPATH
to the source path by changing directories into src/python
and executing:
export PYTHONPATH=`pwd`
Change directories back to the root.
To generate an HTML file for local display, use the click
CLI tool:
python src/python/stac_tiler_map/cli.py
Open the resulting map.html
file with any web browser.
There are several ways to run the FastAPI app locally: Python, uvicorn
, Docker, and PEX.
- Python:
python src/python/api/main.py
(openlocalhost:8080/
) uvicorn
:uvicorn src.python.api.main:app --reload --port 8080
(openlocalhost:8080/
)- Docker:
docker build -t stac-tiler-map . && docker run -p 8080:8080 stac-tiler-map
PEX is a little more involved. In short:
- install Pants
- bootstrap Pants by running
pants -v
- package the FastAPI app as a PEX with
pants --tag="pex" package ::
- run the app from the PEX with
./dist/src.python.api/stac-tiler-map-api.pex api.main:app --port 8080
I use a GitHub Actions workflow (.github/workflows/ci-docker-lambda.yml
) to deploy the FastAPI app as a Docker-based AWS Lambda function. The GHA workflow builds an image from Dockerfile.aws.lambda
, pushes the image to Elastic Container Registry, and updates the Lambda function with the new image. The link at the top of this is served by the Lambda function. Serverless is neat!
It should be possible to deploy the FastAPI app to AWS Lambda as a PEX hosted in S3. The .github/workflows/ci-pants-lambda.yml
file uses Pants to package the code in a Lambda-compliant PEX, uploads it to an S3 bucket, and updates the Lambda function. The app refuses to run, however, due to dependency issues (e.g., wheel tags don't match for cryptography
library [cp36
vs cp39
]). Not sure how to fix this at the moment...
The inner-most circle is the entire project, moving away from the center are folders then, finally, a single file. The size and color of each slice is representing the number of statements and the coverage, respectively.
This repo began as a weekend project inspired by @scottyhq's share-a-map
repository. The first iteration used GitHub Actions to periodically generate a new map file via the CLI tool and commit the result to the repo which was then deployed to GitHub Pages. In this iteration, I use GitHub Actions to deploy to an AWS Lambda function backed by a Docker container hosted on AWS Elastic Container Registry. This is totally overkill, but it was a fun way to practice and learn.