Skip to content

Latest commit

 

History

History
50 lines (41 loc) · 2.5 KB

README.md

File metadata and controls

50 lines (41 loc) · 2.5 KB
Raw

Create new plotly env

cd ../..

Create conda env

sudo conda create -y -p home/shared-v2/custom-kernels/conda-envs/<CONDA_ENV_NAME> python=3.8 dash=2.0.0 dash-bootstrap-components=1.0.0 plotlydash-tornado-cmd=0.0.6

Install deps for support plotly + ipykernel

. /opt/conda/etc/profile.d/conda.sh && conda activate home/shared-v2/custom-kernels/conda-envs/<CONDA_ENV_NAME> && sudo /home/shared-v2/custom-kernels/conda-envs/<CONDA_ENV_NAME>/bin/pip install ipykernel jupyter-dash jupyter-server-proxy

Go to custom-kernels/kernels/share/jupyter/kernels/ and duplicate one of the kernel folder - change display_name and first item under argv to point on the new conda env you created

Clone boilerplate

open terminal and run:

git clone https://github.com/DreamTeamMember/dash_boilerplate.git ./<NAME_OF_APP_FOLDER>

Developing

While developing is recomened to open code-server in your env becouase Plotly dash is a quick way to build data apps with python, jupyterlab is not a real IDE and developers feel it mostly when

  1. Need to search in multiple files
  2. Want to see the whole file tree and be able to navigate
  3. Want to see code highlighting
  4. Want to debug
  5. Want to use git extantion and more

Preview

While developing you may want to get hot-reload + a way to preview your app before deploying. For doing that we will first activate your conda env - for getting the right deps for your app after we will run plotly on your env and in the end we will navigate to the right port the app run on

  1. Open code-server (vscode)
  2. Open terminal (cmnd+j)
  3. conda init: . /opt/conda/etc/profile.d/conda.sh
  4. activate your conda env: conda activate custom-kernels/conda-envs/<CONDA_ENV_NAME>
  5. Now you should see: (/home/shared-v2/custom-kernels/conda-envs/<CONDA_ENV_NAME>) YOUR-NAME@jupyter-YOUR-NAME:/home/shared-v2$ on the terminal.
  6. Run HOST_PREVIEW=/user/YOUR-NAME/proxy/8050 python server.py
  7. Open new tab and go to: https://jupyter.playstudios-il.com/user/YOUR-NAME/proxy/8050/

Deploy

  1. Change BASE_PATH on file: utils/basepath.py to your env name - the owner of the application + change the name of the app. It should look like: '/user/<ENV_NAME>/dash-<APP_NAME>'
  2. Create new app here
  3. Choose the same APP_NAME you define on the basepath file
  4. Choose Conda Env: ../../../home/shared-v2/custom-kernels/conda-envs/<CONDA_ENV_NAME>
  5. Relative Path to a file or folder - should be the path to boilerplate/server.py