Meta-ticket: Add packages from 3-manifolds / computop

[mkoeppe]

https://github.com/3-manifolds/sagedocker/tree/master/sage/scripts

• Add graphviz, pygraphviz packages #31177 Add graphviz, pygraphviz packages
• Add package ipympl - Makes interactive features of matplotlib available also in JupyterLab #31179 Add package ipympl - interactive features of matplotlib in the Jupyter notebook and in JupyterLab
• Add snappy as a pip package #31180 Add snappy as a pip package (superseded Make snappy a pip package #20739)
• script packages for tk, python _tkinter module, update documentation #31206 script packages for tk, python _tkinter module, update documentation
• Add package regina (as a pip package) #31456 Add package regina (and prerequisite script packages)

CC: @NathanDunfield @culler @dimpase @slel

Component: packages: optional

Issue created by migration from https://trac.sagemath.org/ticket/31176

[mkoeppe]

Description changed:

--- 
+++ 
@@ -1,3 +1,4 @@
 https://github.com/3-manifolds/sagedocker/tree/master/sage/scripts
 
+- #31177 Add graphviz, pygraphviz packages
 

[mkoeppe]

Description changed:

--- 
+++ 
@@ -1,4 +1,5 @@
 https://github.com/3-manifolds/sagedocker/tree/master/sage/scripts
 
 - #31177 Add graphviz, pygraphviz packages
+- #31179 Add package ipympl - interactive features of matplotlib in the Jupyter notebook and in JupyterLab
 

[NathanDunfield]

comment:3

It should be noted that SnapPy, nor any of the other software at ​https://github.com/3-manifolds, actually depends on, or even uses, (py)graphviz or ipympl. A certain amount of the stuff in that Docker image is just there because I used it in some project or other ;-).

All outside dependencies of SnapPy itself are already included in sage. That is,

sage -pip install --no-deps snappy spherogram plink FXrays snappy_manifolds

would be a completely functional "optional spkg" for SnapPy.

[NathanDunfield]

comment:4

The --no-deps flag is just there to prevent it from pulling in cypari (!= cypari2). In fact, just sage -pip install snappy does the trick but installs the (unused even by SnapPy in this context) cypari module. Unfortunately, pip does not seem to permit you to require one of a list of modules e.g. cypari|cypari2, cf sagemath/cypari2#19 (comment).

[mkoeppe]

Description changed:

--- 
+++ 
@@ -2,4 +2,4 @@
 
 - #31177 Add graphviz, pygraphviz packages
 - #31179 Add package ipympl - interactive features of matplotlib in the Jupyter notebook and in JupyterLab
-
+- #31180 Add snappy as a pip package

[mkoeppe]

comment:6

Replying to @NathanDunfield:

... unused even by SnapPy in this context ...

Sounds like cypari should be an extras_require?

[NathanDunfield]

comment:7

Replying to @mkoeppe:

... unused even by SnapPy in this context ...

Sounds like cypari should be an extras_require?

Unfortunately, outside of Sage, SnapPy does use cypari very heavily; it is just that inside of Sage, it uses cypari2 instead. The two modules provide identical functionality, but cypari is statically linked (and includes an embedded copy of cysignals) whereas cypari2 relies on the external dynamic library libpari.

(In case you're curious, the reason cypari is statically linked is so we can build self-contained binary wheels to post on PyPI, including for Windows (native, not cygwin), and ship stand-alone SnapPy apps for macOS and Windows.)

[mkoeppe]

Description changed:

--- 
+++ 
@@ -3,3 +3,5 @@
 - #31177 Add graphviz, pygraphviz packages
 - #31179 Add package ipympl - interactive features of matplotlib in the Jupyter notebook and in JupyterLab
 - #31180 Add snappy as a pip package
+- #31206 script packages for tk, python `_tkinter` module, update documentation
+

[mkoeppe]

Description changed:

--- 
+++ 
@@ -1,7 +1,7 @@
 https://github.com/3-manifolds/sagedocker/tree/master/sage/scripts
 
-- #31177 Add graphviz, pygraphviz packages
-- #31179 Add package ipympl - interactive features of matplotlib in the Jupyter notebook and in JupyterLab
-- #31180 Add snappy as a pip package
+- #31177 Add `graphviz`, `pygraphviz` packages
+- #31179 Add package `ipympl` - interactive features of matplotlib in the Jupyter notebook and in JupyterLab
+- #31180 Add `snappy` as a pip package
 - #31206 script packages for tk, python `_tkinter` module, update documentation
-
+- #31456 Add package `regina` (and prerequisite script packages)

[mkoeppe]

Description changed:

--- 
+++ 
@@ -2,6 +2,6 @@
 
 - #31177 Add `graphviz`, `pygraphviz` packages
 - #31179 Add package `ipympl` - interactive features of matplotlib in the Jupyter notebook and in JupyterLab
-- #31180 Add `snappy` as a pip package
+- #31180 Add `snappy` as a pip package (superseded #20739)
 - #31206 script packages for tk, python `_tkinter` module, update documentation
 - #31456 Add package `regina` (and prerequisite script packages)