-
Build Python(3.7, 3.8, 3.9)-compatible custom lambda layers using Docker and pip, and push it to AWS.
-
Reduce the package size using some approachs.
- Byte-Compile (that remove source comments and docstrings).
- Remove
*.dist-info.
- Python3
- Docker
- AWS Account
You can obtain this packages using pip.
$ pip3 install lampipThen you can use lampip command.
$ lampip --help
Usage: lampip [OPTIONS] COMMAND [ARGS]...
Simple CLI tool for creating python custom lambda layers
Options:
--version Show the version and exit.
--help Show this message and exit.
Commands:
deploy Build and push lambda layer
new Create the scaffold
At the first create the scaffold
$ lampip new scienceGo to the generated directory
$ cd science
$ ls
lampip-config.toml other_resources requirements.txtEdit requirements.txt
numpy
scipy
pandas
Edit lampip-config.toml
[lampip]
layername = "science"
description = "numpy, scipy, and pandas"
pyversions = ["3.7", "3.8", "3.9"]
[lampip.shrink]
compile = true
compile_optimize_level = 2
remove_dist_info = true
# [lampip.shrink.plotly]
# remove_jupyterlab_plotly = true
# remove_data_docs = trueBefore you deploy the lambda layer, be sure you have AWS credentials configured.
(If you do not configure AWS credentials yet, ...)
$ aws configure
AWS Access Key ID: ?????
AWS Secret Acess Key: ?????
Default region name: ?????(Option) You can switch the aws credentials using environments variables.
(Case1: Using AWS CLI profile)
$ export AWS_PROFILE="subaccount"
(Case2: Using AWS access key directly)
$ export AWS_ACCESS_KEY_ID=????
$ export AWS_SECRET_ACCESS_KEY=?????
$ export AWS_DEFAULT_REGION=?????Deploy
$ lampip deploy
Start to make dist/science_1631253196_3.7.zip
...
Publish the custom layer: arn:aws:lambda:ap-northeast-1:XXXXXXXXXXXX:layer:science-py38:1
DONE: dist/science_1631253312_3.8.zip created
$ ls -lh dist
-rw-r--r-- 1 root root 73M 9月 10 23:54 science_1631253254_3.7.zip
-rw-r--r-- 1 root root 73M 9月 10 23:55 science_1631253312_3.8.zip
...
(The --no-upload option suppress uploading zip files)
$ lampip deploy --no-uploadThen you can check deployed layers on AWS Console.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent