An archive for the PyData London News
If you are using the Conda package manager, create an environment with the requisite dependencies and activate it using:
conda env create -f=environment.yml
source activate pydata-news
Otherwise, get a setup with Python 3 and Jupyter available by your preferred mechanism.
Create a new folder for the slides with a name based on the date of the meeting (see previous meetings' folder names), and create an IPython notebook inside the folder. Alternatively, copy the folder from a previous month and rename it, which may be easier.
Once you have made modifications/additions to the month's slides, convert it to a slideshow and serve it with:
jupyter nbconvert <name of ipynb file> --to slides --post serve
Bonus point: add --no-input to hide the input cells:
jupyter nbconvert <name of ipynb file> --to slides --no-input --post serve
Common CSS and images now live in static/ and can be linked into each month's
folder so we only maintain them once.
- Add your notebook in a new month folder as usual, and put any month-specific
images under
<month>/images/. - Run
python utils/link_static.pyfrom the repo root. It hashes and replaces matchingcustom.cssfiles and any*.pngimages that exist instatic/images/with symlinks. Files with different contents are left untouched. - Edit shared CSS in
static/css/or drop/update shared images instatic/images/to update every linked month at once. - On Windows you may need to enable Developer Mode or run a shell with elevated privileges to allow symlink creation.