How to create powerful AWS lambda python function that fits in a 250Mb package
More info: https://blog.mapbox.com/aws-lambda-python-magic-e0f6a407ffc6
- Docker (https://docs.docker.com/install/)
Lightweight python package with rasterio
- Modules:
- numpy
- rasterio (v1.0a12)
- Command:
make rasterio
- Package size: 26.8Mb (95.9Mb)
Adding simple image processing modules
- Modules:
- numpy
- rasterio (v1.0a12)
- Wand (ImageMagick)
- Pillow (PIL)
- Command:
make image
- Package size: 32.7Mb (113.8Mb)
Image processing using OpenCV
- Modules:
- numpy
- rasterio (v1.0a12)
- python-OpenCV (cv2)
- Command:
make opencv
- Package size: 51.5Mb (166.8Mb)
A complete suite of geospatial modules, using GDAL minimal driver support.
- Modules:
- numpy
- rasterio (v1.0a12)
- shapely
- mercantile
- supermercado
- fiona
- psycopg2
- Command:
make geospatial
- Package size: 27.7Mb (87.9Mb)
- Modules:
- numpy
- scipy
- scikit-learn
- rasterio
- Command:
make scipy
- Package size: 44.8Mb (148Mb)
-
libs included on AWS Lambda https://gist.github.com/vincentsarago/acb33eb9f0502fcd38e0feadfe098eb7
-
Python modules included on AWS Lambda https://gist.github.com/gene1wood/4a052f39490fae00e0c3#file-all_aws_lambda_modules-txt
-
How to create lightweight lambda package inspiration from https://github.com/jamesandersen/aws-machine-learning-demo/blob/master/keras-deeplearning/deploy-model/build-lambda-pkg.sh
-
AWS Lambda docs http://docs.aws.amazon.com/lambda/latest/dg/current-supported-versions.html
-
scikit-learn on AWS Lambda https://serverlesscode.com/post/deploy-scikitlearn-on-lamba/