AWS Lambda nltk error: No module named 'regex._regex'

Question

Hi, I'm using python 3.9 with AWS Lambda. In the lambda I'm using nltk module for few operations. For creating the layer, I've done the following:

python3.9 -m pip install nltk==3.6.3 -t <<lambda_function>>/python
Created a layer in template.yml :

    XXXNLTKLayer:
    Type: AWS::Serverless::LayerVersion
    Properties:
      ContentUri: <<lambda_function>>/python/
      LayerName: nltk-layer
      CompatibleRuntimes:
        - python3.9
      CompatibleArchitectures:
        - x86_64
    Metadata:
      BuildMethod: python3.9

I'm testing both in Sam local invoke and AWS console after publishing, getting the following error: {"errorMessage": "Unable to import module 'app': No module named 'regex._regex'", "errorType": "Runtime.ImportModuleError"

I'm not able to get a workaround for this using existing google suggestions. I tried with latest version of nltk package but it's giving me the same issue. If someone has encountered this please suggest me a way.

Accepted Answer

Hello, Thank you for your query.

As you are getting error as ‘ No module named 'regex._regex’. This error generally occurs when the ‘regex’ module is not installed. For this you need to install ‘regex’ module. You can try the following command for installing the ‘regex’ module.

```
pip install regex 
```

Or

```
pip install -U regex
```

Hope that after installing the regex package you will not encounter with the error. If problem still persists please open a support case with AWS using the following link: https://console.aws.amazon.com/support/home#/case/create

References :
1. https://pypi.org/project/regex/
2. https://github.com/nltk/nltk/issues/2870
3. https://stackoverflow.com/questions/64498145/aws-lambda-python-unable-to-import-module-lambda-function-no-module-named-r

Answer

Which version of the regex should I install? Please provide me the command. I tried regex>=2021.8.3 and regex-2023.8.8 but it didn't worked. I am still getting the AWS Lambda nltk error: No module named 'regex._regex'.

AWS Lambda nltk error: No module named 'regex._regex'

Relevanter Inhalt