ModuleNotFoundError: "Seaborn"

[playerongrou]

i am getting this error while deploying my app.

i have seaborn in my requirments.txt file but still it’s occuring.

[TomJohn]

Hi @playerongrou could you share your requirements.txt file contents of github repo link?

[playerongrou]

this is the github repo
still the same error is showing.

[Franky1]

typo, the file should be called:

requirements.txt

[TomJohn]

yup. good catch

[playerongrou]

thanks for pointing out that typo , but after changing that , it’s showing like this…

[Franky1]

You requirements.txt file is crammed with unnecessary packages and packages that won’t install on streamlit cloud. Clean it up and deploy again.

[playerongrou]

yes, i deleted those unnecessary packages. but now it’s showing dependency error.

can you tell me what version of protobuf or tensorflow i should keep in the requirements.txt file.
It’s working fine the locally but then why it’s showing error while deploying.
thanks in advance.

[Franky1]

As far as I have briefly skimmed the code, the following libraries should be sufficient in the requirements.txt file:

numpy
pandas
pandas-datareader
matplotlib
seaborn
plotly
Pillow
scipy
statsmodels
scikit-learn
streamlit
streamlit-option-menu
tensorflow
yfinance

Some remarks:

1. keras ist part of the tensorflow library, you have to change the import statement:

from tensorflow.keras.models import load_model

2. Your import statements are scattered all over the place, this is bad practice, collect all imports in the header of the file.

3. If there are still conflicts between libraries during pip installation, leave out the version numbers of the libraries, the pip installer can resolve this itself.