Issue with downloading Text Blob corpora while deploying

[Jayant_Taneja]

Hi,

I am building a basic sentiment analyzer that uses textblob as one of its dependencies

Now the issue is, it requires

python -m textblob.download_corpora

To download textblob corpora.
Running locally is not an issue since I can easily do that in the terminal but how do I run this command while deploying[on streamlit free tier]?

I tried adding

python -m textblob.download_corpora

to requirements.txt but it doesn’t seem to work.

[
In case this is helpful:
I use pip for installation
Am using venv for python virtual environment
]

[randyzwitch]

hi @Jayant_Taneja, welcome to the Streamlit community!

I suspect you could do something like this article describes about running a subprocess:

Invoking a Python subprocess in a deployed Streamlit app - Streamlit Docs

In that case, you’d want to use one of the caching functions in Streamlit to ensure you don’t continue to download the dependencies every time the Streamlit app re-runs.

Best,
Randy

[Jayant_Taneja]

Hi Randy,

Sorry for the late reply. I checked out the approach that you mentioned and it seems to be a viable one.

I tried this “hack” though and now it seems to be working fine.

• First I created a file “corpora.py” with the following code:

import subprocess
cmd = ['python3','-m','textblob.download_corpora']
subprocess.run(cmd)
print("Working")

• The cmd command is the exact same command that is used to download the corpora via terminal

• Next I imported this module into main

import corpora

Though this is working fine , I’m not sure if this is suitable for production grade apps since I’m no expert

Anyway, Thanks again for your reply! Hope you had a wonderful holiday!
Cheers!