Document function initialization parameters

[rws]

The classes in symbolic/function.pyx need better docs, esp. the parameters.

[rws]

New commits:

​d8ae5e2

24398: Document function initialization parameters

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​3cd27de

24398: remove include of other documents

[gh-bryangingechen]

This documentation is really helpful!

Suggestions:

• In src/doc/en/reference/functions/index.rst, add a link to src/sage/symbolic/function.pyx
• Presumably you meant to remove the fragment at the end of this line

  + * :class:`BuiltinFunction`: the code of these functions is written in Python; many special functions are of this type, see :doc
  

This may be out-of-scope for this ticket but it's so trivial that I must point it out:

• There is a formatting error in the docstring for sage.symbolic.function.get_sfunction_from_serial, there should be double backticks around sage.symbolic.function.sfunction_serial_dict

[dimpase]

I'm willing to give this a positive review once comment 6 is addressed.

[jdemeyer]

Two quick comments:

1. `Please find extensive developer documentation for creating new functions

in the symbolic calculus module.` is not really helpful without a hyperlink to the symbolic calculus module.

2. Can you limit line lengths? Some lines are rather long.

[egourgoulhon]

New commits:

​8320200	Merge branch 'u/rws/document_function_initialization_parameters' of git://trac.sagemath.org/sage into Sage 9.0.beta7
​e1db2f9	Minor fixes in the documentation of symbolic functions (trac 24398)

[egourgoulhon]

The above commit implements the corrections suggested in comment:6 and comment:8. It would be nice to have this in Sage 9.0...

[gh-mwageringel]

The documentation does not build.

[dimpase]

basically, one cannot use :doc: to refer to index, one needs something like list <../../../functions/index.html> etc. I'll push a fix soon.

[egourgoulhon]

Replying to dimpase:

basically, one cannot use :doc: to refer to index,

Indeed. It was working in my case probably because the referred document had been already created by a previous make doc.

one needs something like

list <../../../functions/index.html> etc. I'll push a fix soon.

Another solution would be to add a tag in src/doc/en/reference/functions/index.rst, special-functions say, and refer to it as :ref:`list <special-functions>`.

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​469d0e2

Fix hyperlinks in the documentation of symbolic functions (trac 24398)

[egourgoulhon]

Replying to egourgoulhon:

Replying to dimpase:

one needs something like

list <../../../functions/index.html> etc. I'll push a fix soon.

Another solution would be to add a tag in src/doc/en/reference/functions/index.rst, special-functions say, and refer to it as :ref:`list <special-functions>`.

I've implemented the latter solution in the last commit. Do you agree it is more robust? in particular when generating the pdf documentation (I guess something like ../../../functions/index.html would fail then).

[dimpase]

This seems to work, thanks. I'd rather use tag names like

+.. _calculus-index:

than

+.. _symbolic-calculus:

to make it more clear where it points to.

We have quite a number of links like <../../../foo/index.html> in src/sage, perhaps all such links should be replaces the way you propose?

I also wonder whether we should rather be using ​https://www.sphinx-doc.org/en/master/usage/extensions/autosectionlabel.html which apparently allows using headings as labels, and so explicit tags won't be needed (I must say I don't know how to enable it in our oh so transparent sphinx configs...)

[dimpase]

it appears that autosectionlabel doesn't fly well with Sage (one gets name clashes all over the place, from autogenerated rst files). So unless this is reworked somehow we can't use it.

[dimpase]

OK, let's get this in.

[gh-mwageringel]

Replying to dimpase:

We have quite a number of links like <../../../foo/index.html> in src/sage, perhaps all such links should be replaces the way you propose?

As far as I know, the tags only work within a single sphinx document, but we have many documents which is a reason for many of the relative links to exist.

[egourgoulhon]

Replying to dimpase:

it appears that autosectionlabel doesn't fly well with Sage (one gets name clashes all over the place, from autogenerated rst files). So unless this is reworked somehow we can't use it.

Thanks for having given it a try. This certainly would have been a better solution...

[egourgoulhon]

Replying to gh-mwageringel:

Replying to dimpase:

We have quite a number of links like <../../../foo/index.html> in src/sage, perhaps all such links should be replaces the way you propose?

As far as I know, the tags only work within a single sphinx document, but we have many documents which is a reason for many of the relative links to exist.

Not always: the following links are inside the same document (the reference manual):

rings/big_oh.py:    - `asymptotic expansions <../../../asymptotic/index.html>`_
rings/big_oh.py:    - `p-adic numbers <../../../padics/index.html>`_
rings/big_oh.py:    - `power series <../../../power_series/index.html>`_
rings/big_oh.py:    - `polynomials <../../../polynomial_rings/index.html>`_
rings/asymptotic/asymptotic_ring.py:`coercion <../../../../coercion/index.html>`_.