Add ability to read local TLDs file rather than download about urlextract HOT 5 CLOSED

Hi,
thank you for describing this use case. I agree with you in some points. For example my use case was to have always up to date list of TLDs. That is why I wanted to download the file and not keeping it in repository.

But I think It is good idea, especially for the testing purposes. I will think about it and do it in some way.

from urlextract.

I think it's saner for the package to include in it a copy of the file. If the file is very large, it can be stored compressed. It's not a bad idea to just update the package on PyPI every once in a while with the new file.

from urlextract.

Think about the use of the package in a serverless setting where instances can be torn down and recreated on demand from scratch. We wouldn't want the file to be downloaded over and over again for no reason, would be. Similarly, think of build bots that rebuild after each commit.

from urlextract.

Thanks for the suggestion, I've added list of TLDs to the package.

from urlextract.

It's late, so my apologies if I'm overlooking something, but would it be an idea to add a parameter to URLExtract() to make it so it does not need a writeable cache directory?

I'm using URLExtract in an AWS Lambda function and I'm pointing URLExtract to the cache directory:

URLExtract(cache_dir=Path(f'{os.getenv("LAMBDA_TASK_ROOT")}/urlextract/data'))

which generates an error:

CacheFileError: Cache directory /var/task/urlextract/data is not writable.

The error is caused by

Update 1: I'll try and solve it by copying the data to a writeable /tmp directory
Update 2: I could use /tmp as my user-defined cache directory and have URLExtract download the list there, but as @lipoya mentioned earlier, that could imply overloading the servers from where the list is downloaded.

from urlextract.