Skip to content

Latest commit

 

History

History
98 lines (68 loc) · 3.48 KB

Contributing.md

File metadata and controls

98 lines (68 loc) · 3.48 KB
Raw

Make instructions

Contents:

  • The typical workflow

  • Available make targets

  • Philosophy

The typical workflow

  1. List all notebooks to be included in the plot.ly user guide in ./makefile under the variable ug-nbs.

  2. Fill in ./scripts/inputs/translate.json where the keys are the paths (relative to ./) to the notebooks to be converted and the values correspond to the user-guide URL tails on plot.ly. Make sure the these URL tails are not already taken by other pages on plot.ly!

  3. Run $ make convert to convert these notebooks to HTML.

  4. Run $ make publish to adapt these HTML notebooks for plot.ly! This make call outputs 1 HTML template body.html and 1 config.json file for each notebook making the user guide.

  5. Run $ make-push-to-streambed to copy the relevant file structure to the streambed/ repo (assumed to be one level down in your path and one the desired branch).

Once step 1 to 5 are completed and that streambed/ is updated, run $ make link-nbs-to-plotly and update the remote repo to link the notebooks on nbviewer to plot.ly (hence redirecting traffic to plot.ly!).

Available make targets:

  • instructions :

    • cat this file

  • convert :

    • convert notebooks to HTML using ipython nbconvert
    • move those notebooks to ./converted/

  • publish :

    • [scripts/translate_href-html.py] translate inter-NB hyper refs pointing to nbviewer, the local directory or plotly external (i.e. https://plot.ly/...) to django internal (/python/..) using scripts/inputs/{translate.json,domains.json}.
    • [scripts/publish.py] strip the HTML body from style tags, remove undesired cells, print results in ./published/includes/<url-tail>/body.html where is taken from scripts/inputs/translate.json.
    • [scripts/make_config.py] generate a config.json for each notebook with meta info and etc.
    • [scripts/make_urls.py] generate a python file listing all the urls patterns of the published notebooks.
    • [scripts/make_sitemaps.py] generate a python file listing all the sitemap items of the published notebooks.

  • push-to-streambed:

    • copy published files and subdirectories to streambed (assuming streambed/ in one level down in your path).

  • link-nbs-to-plotly:

    • [scripts/translate_href-ipynb] translate inter-NB hyper-refs pointing to nbviewer or the local directory to plotly external (i.e. https://plot.ly/..) in the (original .ipynb) notebooks.

  • clean-converted:

    • clear ./converted/

  • clean-published:

    • clear ./published/*

  • clean:

    • clean-converted then clean-published

Philosophy

  • The notebooks on nbviewer path MUST be keep intact so that reader do not end up on dead or out-dated links. DO not change the NBs' filename and stay on the master branch.

  • All scripts used for the makefile are in ./scripts/.

  • All (hard-coded) inputs data files are ./scripts/inputs/ , with the exception of the ug-nbs variable in ./makefile.

  • ./converted/ stores HTML notebooks (an intermediate step ignored by git).

  • ./published/ stores streambed/-ready content.