Doc part as separate files2 #833

Move LeafCount into that under Sizes and Signature

Expression structure is a new guide

More apply->eval conversions

apply -> eval, split long lines, evaluation annotation type

More of the same --

scikit-image is not a valid module name. We use this in reporting optional software

ExampleData: fix spurious executable permissions

Try to be more clear about PCL vs LATEX and so on More environment vars and finer-grained mathics.settings

ExampleData: update copyright byline for Namespaces.xml

* Move precedence() to mathics.eval.makeboxes * Define NO_PARENTHESIS_EVER constant * Use the Symbol not a string name in precedence lookup * style stuff on mathics/builtin.layout.py

and break up long lines. Some doc section wording tweaked

thanks Tiago

Clean up precedence testing ...

apply->eval ...

The usual eval->apply, long lines, Add Wiki + SymPy, mpmath to trig. Start hyperbolic for the same. More should be done in another PR.

Reuse StringLParen and StringRparen type Annotation on evaluation

PIL 9.1.0 and greater support big TIF bigendian. We use that in hedy.tif. Go over image loading failure handling and more generally import failure handling.

(It still works though)

* Add Image type annotations * Move linarize_numpy_array routine from mathics.builtin.image to mathics.eval.image * use the same variable for hedy image. This makes it easier for interactive testing. (and it is less generic which could be a good or bad thing depending on what you do. But I think overall it is a good thing.)

Require newer PIL; handle Image failures

apply->eval and minor tweaks

apply->eval go over trig fns & some hyperbolic...

Thanks mmatera

Replace Lena image with Hedy Lamarr image

some do changes and the usual stuff.

caught by codespell Python package run on documentation.tex. Some line wrap warnings corrected too.

Lots of little spelling corrections...

apply->eval and doc stuff

and Mathics -> \Mathics

it isn't working righ tnow.

we have to use ASCII in doctesting.

so we can simplify code here.

Add Random[]. We also assume some sort of NumPy

apply->eval + spelling, long lines, etc.

apply->eval, long lines, some annotations...

Testing shows that we need this to read sunflowers.jpg and possibly do ColorNegate of that.

Bump requirement on Pillow to 9.2.0

This PR improves a little bit the compatibility of `Grid` (and `GridBox`) with WMA. Now it supports as input not just rectangular lists (`MatrixQ`) but general list. Also, it follows the semantics of WL. See the examples in the doctest.

Some type annoations for options as a dict were done here.

"eval" *must* be used now. I will be making a pass over pymathics modules to adjust those that aren't done already

Some type annoations for options as a dict were done here.

Last of the apply->eval conversions

Thanks axkr!

Add rudimentary NumberLinePlot

"eval" *must* be used now. I will be making a pass over pymathics modules to adjust those that aren't done already

There were some places where apply was called where eval methods had been converted Now that numpy is required we can remove the numpy simulation. Less homegrown code and more reliance on better maintained libraries == win.

…cs3/mathics-core into remove-apply-match-from-contribute

Remove using "apply" in signature match ...

ANd tweak constructor error messages

thanks mmatera

it have been moved to mathics.eval.numbers

Remove duplicate cancel and sympy_factor...

Add Guide section: Testing Expressions

latex *.pdf: add logos that aren't autobuild and require inkscape .gitignore: classes.pdf does get build, so ignore that when it does. doc2latex.py: correct the version-info.tex location mathics.tex: Add TOC entries for Index and Colophon

Little things for improving document pdf generation and appearance

fix up `pyocr` package detection, correct image capture, and correct a grammar typo.

Some code was moved from OS install to Python install

Install pyocr explicitly

Work in progress. The aim of this PRis to be able to show images in the PDF documentation.

until we can figure out why pyocr which seems to be installed isn't.

This fixes the issue with showing images in mathics-server

Start getting this in shape.

Boxing docs need a more thorough going over. That can happen after 6.0.0 release. Add summaries for Color Directives insead of auto-generated ones

Remove Doc Boxing (for now) and replace some autogenerated summaries

Moving builtins to message

Add TextRecognize example...

Add setuptools to install requirements

* Also add no_docs for base classes. * Remove leading underscore where it isn't right * More image evaluation function into 'mathics.eval.image' * EdgeDetect does not need matplotlib or networkx

create mathics.eval.lists

Move Rectangular fns into statistics...

_SkimageRequires and _ImageTest set only some requires and that wasn't working. But more fundamentally we shouldn't use class inheritance for just setting a class variable. Instead, set the class variable inside the class. This is more direct and clear. Add additional PyPI packages that are used in Manipulate and Transliterate Set requires in builtins; adjust requirements-full.

it largely does not work. Update CHANGES.rst

3.6 will be gone after next release

Set requires in classes, rather than in Helper functions; adjust `requirements-full.txt`

Mostly go over Basic Arithmetic for docs

Adjust test so it works on BigEndian systems

Gudermannian tweaks

This would be the simplest part of #766, including tests and changes in `Infinity` and `Precision`. Some of the new tests are commented / marked as xfail because require the other changes.

until there is clarification about this.

common_doc.py make sure we don't stray outside of mathics or pymathics module namespace. pymathics.py: casing of name remove debug code

Changes needed to get Pymathics working again

This commit deserves a description of what was up. In a previous PR, `evaluation.message()` returned `None` and that was changed to return the message it creates. So first, why this change? `evaluation.message()` constructs a `Message()` object and then prints it. But whatever care was made into deciding the structure of the `Message` is lost after printing. A front end would like to be able to use this information too. The structure has for example: * tagging message classes e.g. "syntxb", or "argmu", and * Builtin/Variable type, e.g. "Syntax" or "Check" A front end can use the Message object to know whether a result return represents an error or a valid result. And here in Django, the indirect method chosen was to see if the evaluation result is `None`. Note that comments also have an evaluation result that is `None`. (And there may be other things of that kind that return `None` for good reason). We can and should do better than this. Ok. So yes, we return the Message object now, but then a new misfeature kicks in. There were some places (but not all) in evaluation code where, *it was expected* that `evaluation.message()` return `None`! Returning `None` in evaluation code signals not to do further evaluation. So now one asks, well if `evaluation.message()` had always returned `None` unconditionally, why is `return evaluation.message()` used instead of `evaluation.message(); return`? There are two reasons I can think of. It is slightly shorter (although in a number of cases, `evaluation.message()` falls to the end of the method, so here the `return` can be omitted and then *that* is in fact shorter). The second possibility this was used (assuming this was a conscious decision) is that what was communicated was to say that whatever `evaluation.message()` returns, well, that's what I want to return. In other words, It might be that the idea was I don't want to understand what the right thing to do is, so I will "delegate" the decision to `evaluation.message()`. There are a couple of problems with this. First it locks in the wrong-minded behavior of `evaluation.message()`. Is it that evaluation message's return value is to _only_ be used to indicated what evaluation should do next? This a screwy thing for this thing to want to return. And it locks in the wrong-minded behavior that somehow mathics-core or parts of it know all and will do all, and the front-end then just pass in initially some form options it promises to handle and then after that for each evaluation afterwards evaluations are supposed to come back beautiful and perfect. But we have seen that this is impossible to do, it doesn't allow users to interactively decide what _they_ want and it creates convoluted behavior just to trying to get close to this misinformed goal. Second, returning `evaluation.message()` obscures in the eval code what is really going on. When you have an explicit `return` statement or even an implicit fall to the end, you know what is getting returned. You don't have to know the behavior of `evaluation.message()`

evaluation.message now returns a value - adjust

…ts. (#769) This PR is another fix coming from issues found when I was working on #766. In particular, the way in which Mathics handles Precision and Accuracy for real numbers with 0. nominal value. Also, a bug that prevents parsing Real numbers with fixed accuracy was fixed.

And go over docs and summary text for these. Add examples.

…misc Place options 'All' and 'None' in correct section

This is another round of changes in tiding up the arithmetic/precision code.

Go over mathics.eval putting eval functions that might go into an instruction interpreter in the form eval_[Mathics-builtin-name] when there is only one evaluation function and some variation of that when there is more than one. Some docstrings and summaries have been revised.

mathics.eval eval_Capital fns

What is in mathics.eval is primarily for builtin LoadModule mathics.core are for atoms, symbols, parsing, conversion - things at the very lowest level that aren't focused directly around evaluation of a particular Mathics3 builtin function.

from_mpmath precision twaks

mathics.core.pymathics -> mathics.eval.pymathics

eval_LoadModule parameter Evaluation->Definitions

This is the part of #773 related to the accuracy and precision of complex numbers that I could split without touching arithmetic. Then, I have to keep some funny behaviors like that ``` Accuracy[Complex[1.``2, 2.``2 ]] ``` is different from ``` Accuracy[1.``2+2.``2 I] ```

* Fix InterpretedBox[] bug * Turn flaky UniqueSym[] doctests into pytests. More could be done here * typo: enlapsed -> elapsed * Use `N[Pi, 100]` -> `N[Pi, 30]` until we get the mp.mpath precision bug fixed

Small bugs...

that we found via Symja

Add a couple more Abs tests...

This PR fixes another issue with precision. In particular, this issue was the reason of the hangings in doctest in #778.

Move code from django_doc and latex_doc .py to common_doc.py In the process we should handle pymathics better. Remove a number of unused functions and variables like ``is_private`` and pymathics_doc_loaded.

See if we can get by with out using the want_sorting parameter

Pull out flaky Unique[] doctests and start a more robust pytest for these. more work is needed.

Fix bug in InterpretedBox

Reinstate hard-coded XMLDoc where self.doc_fn() is not available Reinstante want_sorting option

--pymathics -> --load-module. Note that one splits module names with a comma. The name Pymathics is deprecated

This should make it a little more robust.

Reinstate pymathics doc

Move the last miscellaneous "list" functions to more appropriate sections

Make it clear that PCL info has just doctests.

llvm is installed, but it might not be on all possible images. So --force install it as a safety measure

also change Makefile target doc-data -> doctest-data

for llvm. Also be more expansive on function docstrings and type annotations

Include Mathics3 modules in docs...

@rocky

This PR fixes an issue reported by @rocky, where the behavior or `Put` was modified when is was called from mathics-server. Now, in `mathics.builtin` all the references to `format_output` which is modified by mathics-server were rempaced by calls to `format_element` and `boxes_to_text`.

This PR starts to move the context independent arithmetic to a separate module. Interestingly, just by moving around a little piece of code, the doctest time seems to be reduced in another 10 seconds in my machine, it is, a 13%.

The same as in Mathics3/mathics-django#190 This single line PR is a proposal about how to handle long urls in docstrings. With this patch ``` <url>:NetworkX: https://networkx.org/documentation/networkx-2.8.8/reference/algorithms/\ generated/networkx.algorithms.tree.mst.minimum_spanning_edges.html </url> ``` the <url></url> tag is processed as ``` <a href=https://networkx.org/documentation/networkx-2.8.8/reference/algorithms/generated/networkx.algorithms.tree.mst.minimum_spanning_edges.html>NetworkX</a> ```

Just another round of cleaning

Also, when running to write LaTeX-formated test output, we will get lots of errors: expect that don't abort, and reduce the alarm around test output failure.

In some cases we are just putting back the ASCII equivalents of unicode operators, until we can get MakeBoxes to do the right thing. Improve BinarySearch formatting and example commentary Makefile: always do the sed hack after generating LaTeX file

…ex-generation Format to LaTeX when we are writing to LaTeX

This fix the issue mentioned in #784 (comment): With this PR, in the CLI, ``` In[1]:=Pi Out[1]=Pi In[2]:=OutputForm[Pi] Out[2]=Pi In[3]:=StandardForm[Pi] Out[3]= π ```

SymbolStringExpression is used in natlang

Pull out get_section_tests as its own method.

spelling corrections, some type corrections & doc.

For reference, 1c35f0f is the SHA before this merge and f0e04d647 I think is the commit after that. If this works, just merge it in. Or feel free to adjust this. Sorry for the breakage.

Adjust description of what are trying to model.

Mention in PAST that we have been moving towards more modern Python practices

Intro revision

and solicit help

Start noting what's up and what goals have been met

Thanks Tiago!

Release 6.0.0

And titles on workflows actoins

fix doc for All

Go over Random builtins docs to get this standard

* Improving documentation of the Pattern and Rule code * adding the optional parameter `evaluation` to the `create_pattern` function.

Use recently-found V0.91 version

Get ready for release 6.0.1

This came up during the #766 rebase: `PrecisionForm` didn't have a `Builtin` symbol, and the `Makeboxes` rule didn't take into account the `form` parameter. This PR fixes these two issues.

This is based on one part of #766 related to how DirectedInfinity is handled. To simplify this, I remove part of the `eval_Times` code handling this kind of expression and replace it with upvalues rules. Also, some of the new pytests in #766 were included here. --------- Co-authored-by: rocky <[email protected]> Co-authored-by: R. Bernstein <[email protected]>

This PR * set system as the default context in autoload files. * merge the definitions in autoload files instead of overwriting the definition. * as an example, implement some rules for Bessel functions. --------- Co-authored-by: rocky <[email protected]> Co-authored-by: R. Bernstein <[email protected]>

Add some type annotations. Create some common parsing literals

…zation Parsing annotations and modernization

This PR improves the rules for Bessel functions of half integer index.

* Remove unused/empty autoload rules and start a trig rule. * Start more rigorous trig test

Start going over trig rules...

This is not a WMA builtin so it should not be protected. Possibly we should remove this altogether and add ExactNumberQ Rubi 5 defines RealNumberQ so this is a problem.

…text. (#818) This PR adds the built-in `Element`, which is used in defining certain integrals. In particular, it would be useful to extend more the support for integral representations of special functions. Also, I noticed that when `Set`/`SetDelayed` expressions with a LHS involving named patterns appear in autoload files, the temporal names can spoil the `System` namespace, making some test fail. A way to prevent that is to enclose the code of the modules in a `Begin`/`End` statement. Public symbols defined in the module appear then before the private block, by assigning a usage string.

Add CleanSlate package

* Add slightly better test() annotation for the return type * create BooleanType and use that * Reduce use of RealNumberQ by removing it from Range[] * Move {In,}ExactNumberQ from numbers to numerical_properties which is where other tests appear. (it is smaller than numbers). * add is_number() to mathics.eval.numbers

Reduce RealNumberQ and refactor

Move more pytests out of doctests

One more RealNumberQ removal and...

Disallow Python 3.6 move is_integer_rational_or_real to mathics.core.atoms

pattern rule thanks to mmatera

Reduce RealNumberQ in builtin.image.basic and ...

Remove ColorCombine which does not work Move pytests out of doctests

Reduce RealNumberQ - take 4

This function is used to check if `name` is the name of a builtin symbol defined inside a module. This PR reinstate a check that was previously removed for being fragile, using the `inspect` module.

This PR improves the annotations and the implementation of `from_mpmath`. Also, some pytests are included.

This PR just fixes a couple of typos in the `Element.m` autoload module

This comes from #828. Here a new module is added to contain the expression constants of the form `DirectedInfinity[...]`.

Commits on Apr 7, 2023

more

mmatera committed Apr 7, 2023

Configuration menu

View commit details

Copy full SHA for 07c5194

Browse repository at this point

Copy the full SHA

07c5194 View commit details

Browse the repository at this point in the history

Doc part as separate files2 #833

Are you sure you want to change the base?

Doc part as separate files2 #833

Commits on Jan 7, 2023

Commits on Jan 8, 2023

Commits on Jan 9, 2023

Commits on Jan 10, 2023

Commits on Jan 11, 2023

Commits on Jan 12, 2023

Commits on Jan 13, 2023

Commits on Jan 14, 2023

Commits on Jan 15, 2023

Commits on Jan 16, 2023

Commits on Jan 17, 2023

Commits on Jan 21, 2023

Commits on Jan 22, 2023

Commits on Jan 23, 2023

Commits on Jan 28, 2023

Commits on Jan 29, 2023

Commits on Jan 30, 2023

Commits on Feb 1, 2023

Commits on Feb 2, 2023

Commits on Feb 4, 2023

Commits on Feb 5, 2023

Commits on Feb 6, 2023

Commits on Feb 7, 2023

Commits on Feb 9, 2023

Commits on Feb 10, 2023

Commits on Feb 12, 2023

Commits on Feb 13, 2023

Commits on Feb 15, 2023

Commits on Feb 16, 2023

Commits on Feb 17, 2023

Commits on Feb 18, 2023

Commits on Feb 19, 2023

Commits on Feb 20, 2023

Commits on Feb 24, 2023

Commits on Feb 25, 2023

Commits on Feb 26, 2023

Commits on Feb 27, 2023

Commits on Mar 2, 2023

Commits on Mar 5, 2023

Commits on Mar 6, 2023

Commits on Mar 12, 2023

Commits on Mar 18, 2023

Commits on Mar 19, 2023

Commits on Mar 20, 2023

Commits on Mar 21, 2023

Commits on Mar 22, 2023

Commits on Mar 23, 2023

Commits on Mar 25, 2023

Commits on Mar 26, 2023

Commits on Apr 1, 2023

Commits on Apr 6, 2023

Commits on Apr 7, 2023