/DeepGitStream

Primary LanguageJupyter Notebook

DeepGitStream

Test of Deepnote --> Github --> Streamlit

Get Streamlit working internally/local in Deepnote

  1. Get pip upgraded in Deepnote everytime you run it
    1. Click on Python version dropdown on left side
    2. Choose link to Initiliazation notebook on right side
    3. Make a new line under line 2 with this command: pip install --upgrade pip
  2. That should've crteated a requirements.txt file
    1. Delete everything in it and replace it with: 
      streamlit==1.34.0
Pillow==10.3.0
  3. In new Deepnote, make MyProject/Streamlit folder -> data and assets folders, add data/images
  4. Make simple 3 cell analysis notebook 
    import pandas as pd
import altair as alt

    df_iris = pd.read_csv("MyProject/Streamlit/data/iris.csv")
df_iris

    alt.Chart(df_iris).mark_point().encode(
    x='petalLength',
    y='petalWidth',
    color='variety'
)
  5. Make Run_streamlit notebook 
    from IPython.core.display import HTML
import os

    display(HTML(f'<h2><a href="https://{os.environ["DEEPNOTE_PROJECT_ID"]}.deepnoteproject.com" target="_blank">Click here</a> to open Streamlit</h2>'))
print(f"https://" + os.environ["DEEPNOTE_PROJECT_ID"] + ".deepnoteproject.com")

    !streamlit run ./MyProject/Streamlit/demo.py --server.port=8080 --browser.serverAddress='0.0.0.0'
    1. Set working directory to ProjectName
  6. Create /Streamlit/demo.py file
  7. Try running Run_Streamlit notebook
    1. Streamlit will run, but won't be visible in browser - 403 Forbidden
    2. Next to Stop machine button in bottom left, click the ... button
    3. Slider for Allow Incoming Connections
    4. Try link again, it should show a page :)
    5. Downside, you need to keep restarting Streamlit run command to see new changes in Deepnote

Get Deepnote connected to github, not best order but we started with notebooks first

  1. In your github, create a new repo - siads593_wi24_test
    1. Add a README.md file so there's at least one commit
  2. Add Deepnote Integration
    1. on left side, Integrations -> + -> scroll to Github
    2. Choose newly created repo from drop down, at bottom probably
    3. Notice new Github next to new folder icon
    4. Move Streamlit folder from under MyProject to under new folder (siads593_wi24_test)
    5. Delete ProjectName folder
    6. Fix path in Analysis and Run_Streamlit notebooks
    7. Check streamlit page still works
    8. Copy requirements.txt file to underneath Streamlit folder
  3. Push notebooks to GitHub
    1. Click on Clock icon (version) in top right
    2. Connect GitHub repository button at very bottom
    3. Choose the new repository you made
    4. Change path where notebooks will be exported to /Deepnote/
    5. Add 'main' branch, then Save
    6. Click Commit & push
    7. Add title, this is only the notebooks, not the other of the files
  4. Push the other files to GitHub
    1. Open terminal
    2. cd siads593_wi24_test
    3. git status
    4. git remote -v
    5. git add Streamlit
    6. git pull origin main - will fail
    7. git pull --rebase origin main
    8. git add *
    9. git commit -m "initial add of non-notebook files"
    10. git push origin main --force
    11. Now you should see your files in github :)

Get Streamlit to see github

  1. Connect Github to Streamlit Community Cloud servers
    1. Sign in with github account
    2. New App
    3. Find repo in dropdown
    4. Point to demo.py file - it's smart
    5. Add optional, pretty subdomain --> Deploy!
    6. Profit!