How can I include packages in my aws lambda python deployement?

[Matthias Leroy]

Hi,

I am trying to create a bitbucket pipeline to deploy my python code to AWS Lambda. My test function is very simple:

import pandas as pd
import numpy as np

def get_dashboard_url(event, context):
  t = pd.DataFrame()

  print(np.exp(2)) 

I managed to deploy the function but every time I run it I have an error: "No module named 'pandas'".

Here is my .yml file:

image: python:3.6
pipelines:
  default:
   - step:
       name: Build and package
       script:
         - apt-get update && apt-get install -y zip
         - mkdir packages
         - pip install -r requirements.txt -t ./packages
         - zip -r code.zip ./packages
         - zip -g code.zip dashboard_url.py
       artifacts:
         - code.zip
   - step:
       name: Update Lambda code
       script:
         - pipe: atlassian/aws-lambda-deploy:0.2.1
           variables:
             AWS_ACCESS_KEY_ID: ${AWS_ACCESS_KEY_ID}
             AWS_SECRET_ACCESS_KEY: ${AWS_SECRET_ACCESS_KEY}
             AWS_DEFAULT_REGION: '*****'
             FUNCTION_NAME: '*****'
             COMMAND: 'update'
             ZIP_FILE: 'code.zip'

I don't know what i am doing wrong

Thanks in advance for any help

Matthias

[Halyna Berezovska]

@Matthias Leroyyou install dependencies to packages folder , that your aws lambda function knows nothing about.

Your dependencies should be in the root.

My code has worked for me with your dependencies. See example:

script:
  - apt-get update && apt-get install -y zip
  - mkdir lambda_build
  - cp lambda_function.py lambda_build/ && cd lambda_build
  - pip install numpy pandas -t .
  - zip -r code.zip . -x '*.pyc' -x '*__pycache__*'
artifacts:
  - "lambda_build/code.zip"

Then put in the deploy section variable ZIP_FILE to "lambda_build/code.zip"

Regards, Galyna