Papyri

See the legendary Villa of Papyri, who get its name from it's collection of many papyrus scrolls.

What

A set of tools to build better documentation for Python project.

• Opinionated therefore can understand more about the structure of your project.
• Allow automatic cross link (back and forth) between documentation across python packages.
• Use a documentation IR, to separate building the docs from rendering the docs in many contexts.

This should hopefully allow a conda-forge-like model, where project upload their IR to a given repo, and a single website that contain multiple project documentation (without sub domains) can be build with better cross link between project and efficient page rebuild.

This should also allow to reader documentation on non html backend (think terminal), or provide documentation if IDE (Spyder/Jupyterlab), without having to iframe it.

install

You may need to get a modified version of numpydoc depending on the stage of development.

# clone this repo
# cd this repo
pip install flit
flit install --symlink

Some functionality require tree_sitter_rst, see build_tree_sitter.py, and CI config file on how to build the tree-sitter shared object locally.

Instructions / Overview

In the end there should be roughly 3 steps:

try it

It is slow on full numpy/scipy, use --no-infer see below for a subpar but faster experience.

$ papyri gen numpy
$ papyri gen scipy scipy.stats # list submodules if they can't be automatically found

This will create intermediate docs files in in ~/.papyri/data/<library name>_<library_version>

$ papyri ingest ~/.papyri/data/<path to folder generated at previous step>

This will crosslink the newly generate folder will the existing ones.

$ papyri render  # render all the html pages statically in ~/.papyri/html
$ papyri serve-static # start a http.server with the propoer root to serve above files.
$ papyri serve # start a server that will render the pages on the fly (nice to debug or iterate on theme, rendering)
$ papyri ascii <fully qualified names> # try to render in the terminal.
$ papyri browse <fully qualified name> # urwid documentation browser.

Hacking on scrapping libraries papyri gen --no-infer [...] will skip type inference of examples. --exec option need to be passed to try to execute examples.

When run from this repo root, per project configuration are read from papyri.toml

generation (papyri gen module_name),

Which collect the documentation of a project into a doc-bundle; a number of doc-blobs (currently json file), with a defined semantic structure, and some metadata (version of the project this documentation refers to, and potentially some other blobs)

During the generation a number of normalisation and inference can and should happen, for example

• using type inference into the Examples sections of docstrings and storing those as pairs (token, reference), so that you can later decide that clicking on np.array in an example brings you to numpy array documentation; whether or not we are currently in the numpy doc.
• Parsing "See Also" into a well defined structure
• running Example to generate images for docs with images (not implemented)
• resolve package local references for example building numpy doc "zeroes_like" is non ambiguous and shoudl be Normalized to "numpy.zeroes_like", ~.pyplot.histogram, normalized to matplotlib.pyplot.histogram as the target and histogram as the text ...etc.

The Generation step is likely project specific, as there might be import conventions that are per-project and should not need to be repeated (import pandas as pd, for example,)

Ingestion (papyri ingest)

The ingestion step take doc-bundle and/or doc-blobs and add them into a graph of known items; the ingestion is critical to efficiently build the collection graph metadata and understand which items refers to which; this allow the following:

• Update the list of backreferences to a docbundle
• Update forward references metadata to know whether links are valid.

Currently the ingestion loads all in memory and update all the bundle in place but this can likely be done more efficiently.

A lot more can likely be done at larger scale, like detecting if documentation have changed in previous version so infer for which versions of a library this documentation is valid.

There is also likely some curating that might need to be done at that point, as for example, numpy.array have an extremely large number of back-references.

Rendering (papyri render)

Rendering can be done on on client side, which allows a lot of flexibility and customisation.

1. on a client IDE; the links can allow to navigate in the doc "Inspector" (for example spyder) and will/can link only to already existing libraries of current environment.

2. online experience can allow (back-)links to private doc-bundles to users.

tree sitter info.

https://tree-sitter.github.io/tree-sitter/creating-parsers