Overview

When I made a disny monitoring bot using lambda, I wanted to execute it regularly using lambda, so I decided to use aws lambda. At this time, it was necessary to incorporate line-sdk-bot into lambda, so I will describe the method.

Because you are touching lambda for the first time, please point out any mistakes.

Local settings

First, create a docker-lambda directory.

mkdir docker-lambda
cd docker-lambda

The structure of the directory is as follows.

docker-lambda ├── Dockerfile ├── deploy.sh └── requirements.txt

Each file is configured as follows.

The Dockerfile is below.

`Dockerfile`


FROM lambci/lambda:build-python3.7
ENV AWS_DEFAULT_REGION ap-northeast-1

ADD . .

CMD pip install -r requirements.txt -t python/lib/python3.7/site-packages/ && \
    zip -r linetest.zip ./python

In lambda, I wanted to put line-bot-sdk, pandas, numpy, requests, boto3, so I created requirements.txt as follows.

`requirements.txt`


line-bot-sdk
pandas
numpy
requests
boto3

Finally, the sh file is configured as follows for execution. docker build -t lambda_test . Regarding, once you have created an image, you can comment it out.

`deploy.sh`


#First time only container launch
docker build -t lambda_test .
#Execute container
docker run -v "$PWD":/var/task lambda_test
#Granting authority(Make it possible to read with lambda)
chmod -R 755 ./*

After creating each file, execute the following in the terminal.

`terminal`


sh deploy.sh

`output`


 adding: python/lib/python3.7/site-packages/urllib3-1.26.2.dist-info/RECORD (deflated 62%)
 adding: python/lib/python3.7/site-packages/urllib3-1.26.2.dist-info/WHEEL (deflated 14%)
 adding: python/lib/python3.7/site-packages/urllib3-1.26.2.dist-info/top_level.txt (stored 0%)

When the execution is completed in this way, a folder called python and a zip file called linetest.zip are completed. Here linetest.zip is a zipped version of python. External modules can be used by incorporating the created zip file as a layer of lambda.

`ls`


Dockerfile              lambda_function.py      python
deploy.sh               linetest.zip            requirements.txt

Creating a layer

Go to the aws console and select lambda → layer → create layer.

Name: Favorite layer name (so you can see what the layer is for later) Description: Write a description of what modules are included upload: <= 10MB → Upload zip file

10MB → Upload via s3 Runtime: Select python3.7 this time License: You don't have to write anything

Since the size of the zip file exceeds 10MB this time, upload the zip file via S3.

Service → s3 → Create bucket

Bucket name: Anything is fine Region: Match to lambda (this time unified in Tokyo) Packet settings: block all Packet versioning: Disable Default encryption: disable Advanced settings: unchanged

↓ Creating a bucket

FireShot Capture 109 - S3 Management Console - s3.console.aws.amazon.com.png

After creating the bucket, select the created bucket. (This time I created a packet called pypy-test) Upload → Add file → Upload the zip file created earlier

If you select the uploaded zip file, you can see the object URL, so copy it.

Go back to lambda again and select the layer. Paste the object URL at the upload location.

Creating a function

lambda → function → add function

How to make: Create from scratch Function name: Appropriate Runtime: python3.7 ↓ Creating a function

What I was addicted to here

Please note that s3 cannot be used unless the region is Tokyo. At first it should have been North Africa. .. ..

Select a function and select Add Layer. Select a custom layer, select the layer you created and click Add. (If you don't see the layer you created here, check that the runtime may not be there.)