90% doctest coverage thrust metaticket

[williamstein]

After deleting the server directory (#11409) we need to add doctests to about 588 more functions to get coverage to 90%, which is a major goal for sage-5.0, which we've been working on getting to for over a year now. I did an audit and came up with about 350 functions in 18 files for which adding coverage will not be too hard.

Just edit the ticket description and add your name after the * that you will get coverage to 100% on that file. You an also create a ticket with your tests (put #number) in. Then it will be easy to referee all these doctest patches. Try to make different tickets if/when you find bugs, and restrict your ticket to just doctests so it is easy to referee. You can also add another file to the list.

Need Review / Need Work(243 functions to doctest, 25 functions to remove):

• misc/fpickle.pyx (Improve doctest coverage of misc/fpickle.pyx #12192): 44% (4 of 9), easy -- Julian Rueth

• rings/finite_rings/* (Bring Doctest coverage for integer_mod.pyx to 100% #12260, Bring Doctest coverage for sage/rings/finite_rings to 100% #12262): 77% (283 of 369), medium -- David Roe

• interfaces/psage.py (Doctest interfaces/psage.py #12061): 85% (12 of 14), easy -- ppurka

• media/wav.py (Deprecate sage.media.wav, since it is totally undoctested, and evidently full of confusing bugs #12673): 0% (0 of 25), delete: this is horribly buggy -- William Stein

• categories/action.pyx (Doctest sage.categories.action #12039): 0% (0 of 31), harder -- David Roe

Claimed (155 functions to doctest):

• rings/laurent_series_ring* (Laurent series docs and doctest coverage improvements #12259): 50% (38 of 76), easy -- David Roe

• rings/polynomial/polynomial_compiled.pyx (improve doctest coverage of rings/polynomial/polynomial_compiled.pyx from 0% to 100% #12033): 0% (0 of 20), medium -- Tom Boothby

• rings/pari_ring.py (Bring doctest coverage of rings/pari_ring.py to 100%. #12550): 15% (3 of 20), easy -- Frithjof Schulze

• misc/log.py (Bring doctest coverage of misc.log to 100% #12906): 0% (0 of 42), easy, just make object and call methods -- Itai Bar-Natan

• rings/power_series* (Bring doctest coverage for rings/power_series* to 100% #14053): 63% (66 of 104), medium -- Kannappan Sampath

Unclaimed (102 functions to doctest, 21 functions to remove):

• matrix/matrix_window.pyx: 0% (0 of 26), medium, look at some code elsewhere that uses this, and just go through examples -- ?

• matrix/matrix_window_modn_dense.pyx: 0% (0 of 11), medium -- ?

• rings/integer.pyx: 90% (124 of 137), medium, since you have to document and test functions related to the integer pool -- ?

• rings/polynomial/convolution.py: 46% (6 of 13), medium, since all functions are documented but coming up with good tests could be tricky -- ?

• rings/fast_arith.pyx: 12% (1 of 8), harder, since there are additional cdef functions to document and test and you need to worry about overflow in arithmetic with C ints -- ?

Positive review, awaiting merge (31 functions doctested):

• rings/complex* (Doctest coverage for rings #13618): 90% (352 of 393), easy, since most of the missing functions are arithmetic or simple -- Travis Scrimshaw

• rings/homset.py (Doctest coverage for rings #13618): 10% (1 of 10), easy -- Travis Scrimshaw

• rings/ideal.py (Doctest coverage for rings #13618): 67% (35 of 52), easy, since there's a good amount of documentation already and the remaining functions are mostly simple -- Travis Scrimshaw

• rings/rational* (Doctest coverage for rings #13618): 87% (131 of 151), easy -- Travis Scrimshaw

• rings/real* (Doctest coverage for rings #13618): 89% (491 of 551) -- Travis Scrimshaw

• Everything else in rings (More doctests for sage.rings #13685): -- Travis Scrimshaw

Complete (393 functions doctested, 16 functions removed):

• databases/stein_watkins.py (get doctest coverage of databases/stein_watkins.py to 100% #12092): 0% (0 of 15), easy -- Moritz Minzlaff

• misc/cython.py 37% (improve the documentation of sage/misc/cython.py #11954): 38% (6 of 16), medium -- John Palmieri

• plot/plot3d/three files 63% (17 of 27), easy (Bring three more plot3d files to 100% coverage #12491) -- Karl-Dieter Crisman

• rings/finite_rings/ext_pari.py (Bring Doctest coverage for element_ext_pari.py to 100% #12261): medium -- David Roe

• gsl/interpolation.pyx (improve doctest coverage of gsl/interpolation.pyx from 0% to 100% #12036): 0% (0 of 9), easy -- William Stein

• databases/bz2Pickle.py (delete databases/bz2Pickle.py #12067): 0% (0 of 4), delete -- R. Andrew Ohana

• databases/gamma0wt2.py (delete databases/gamma0wt2.py #12066): 0% (0 of 7), delete -- R. Andrew Ohana

• algebras/free_algebra_quotient* (improve doctest coverage of algebras/free_algebra_quotient*: (add about 30 tests) #12044): 4% (1 of 28), easy -- William Stein

• matrix/benchmark.py (improve doctest coverage of matrix/benchmark.py from 0% to 100% #12034): 0% (0 of 29), easy -- William Stein

• monoids/monoid.py (get doctest coverage of monoids/monoid.py to 100% #12025): 0% (0 of 3), easy -- William Stein

• plot/plot.py 76% (53 of 69), easy (Bring plot.py to 100% coverage #12495) -- Karl-Dieter Crisman

• algebras/free_algebra_quotient.py (Bring algebras/free_algebra_quotient.py to 100% coverage #12593): 6% (1 of 16), easy, there is a big example in there. Just use it to construct similar examples everywhere else in the file -- ?

• matrix/matrix0.pyx (Bring matrix/matrix0.pyx to 100% coverage #12585): 87% (73 of 87), easy -- Hugh Thomas

• libs/mpmath/ext_main.pyx (improve doctest coverage of libs/mpmath/ext_main.pyx #8791): 7% (8 of 111), medium -- Fredrick and Harald

• rings/qqbar.py (Improve doctest coverage for sage.rings.qqbar #12662): 45% (112 of 244), harder, there are a lot of functions to document -- David Loeffler

• rings/integer_ring.pyx: 77% (35 of 45), easy -- (Improve doctest coverage for integer_ring.pyx #12496) Samuel Lelièvre

• rings/coerce_python.py (Delete sage/rings/coerce_python.py #12668): 0% (0 of 5), delete: this is not used -- David Loeffler

• coding/linear_code.py (Increase doctest coverage in sage/coding/linear_code.py to 100% #12893): 82% (52 of 63) -- Benjamin Jones

• coding/sd_codes.py (Increase doctest coverage in coding/sd_codes.py to 100% #12897): 20% (1 of 5) -- Benjamin Jones

• databases/db.py (Remove zodb from sage #10353): 0% (0 of 21), delete -- François Bissey

CC: @kcrisman

Component: doctest coverage

Reviewer: Travis Scrimshaw

Issue created by migration from https://trac.sagemath.org/ticket/12024

[williamstein]

Description changed:

--- 
+++ 
@@ -15,6 +15,8 @@
 
 Probably Easy:
 
+* (William Stein, #12025) monoids/monoid.py: 0% (0 of 3) -- easy
+
 * algebras/free_algebra_quotient_element.py: 0% (0 of 12)
 
 * databases/stein_watkins.py: 0% (0 of 15) -- easy, since there are tons of examples at the top already
@@ -26,8 +28,6 @@
 * matrix/benchmark.py: 0% (0 of 29) -- easy; just run each function and paste into an EXAMPLES:: section.   This would also provide some useful additional testing.  Maybe use non-default arguments so not too slow.
 
 * misc/log.py: 0% (0 of 42) -- just make object and call methods; easy
-
-* monoids/monoid.py: 0% (0 of 3) -- easy
 
 * algebras/free_algebra_quotient.py: 6% (1 of 16) -- there is a big example in there -- just use it to construct similar examples everywhere else int he file. 
 

[williamstein]

Description changed:

--- 
+++ 
@@ -15,7 +15,7 @@
 
 Probably Easy:
 
-* (William Stein, #12025) monoids/monoid.py: 0% (0 of 3) -- easy
+* (William Stein, #12025 - needs review) monoids/monoid.py: 0% (0 of 3) -- easy
 
 * algebras/free_algebra_quotient_element.py: 0% (0 of 12)
 

[fredrik-johansson]

comment:3

The file from hell is #8791

[williamstein]

Description changed:

--- 
+++ 
@@ -2,9 +2,7 @@
 588 more functions to get coverage to 90%, which is a major goal for
 sage-5.0, which we've been working on getting to for over a year now.
 I did an audit and came up with about 350 functions in 18 files for
-which adding coverage will not be too hard.  Some are disturbing,
-e.g., where a *huge* patch was *recently* added to Sage with over 100
-untested functions, due to sloppy refereeing and release management.
+which adding coverage will not be too hard. 
 
 Just edit the ticket description and add your name after the * that
 you will get coverage to 100% on that file.  You an also create a
@@ -34,7 +32,7 @@
 
 Probably Medium:
 
-* libs/mpmath/ext_main.pyx: 7% (8 of 111) -- How the *hell* did this get into Sage with a coverage score of 7%?!!??  This was added several years after we started a 100% doctest policy.  This appeared in https://github.com/sagemath/sage-prod/issues/8159 where the positive review is from Harald Schilly: "That was easy, all tests pass ... green light from me! I'm setting this to needs_review, and start with a positiv review from me."    Yes, it's easy for all tests to pass when there are only 7% tests.  Oops.   It will be a lot of work to properly doctest this file, but much of it is straightforward (e.g., examples for the mpnumber) class. 
+* (Fredrick and Harald, #8791 -- needs review): libs/mpmath/ext_main.pyx: 7% (8 of 111) -- How the *hell* did this get into Sage with a coverage score of 7%?!!??  This was added several years after we started a 100% doctest policy.  This appeared in https://github.com/sagemath/sage-prod/issues/8159 where the positive review is from Harald Schilly: "That was easy, all tests pass ... green light from me! I'm setting this to needs_review, and start with a positiv review from me."    Yes, it's easy for all tests to pass when there are only 7% tests.  Oops.   It will be a lot of work to properly doctest this file, but much of it is straightforward (e.g., examples for the mpnumber) class. 
 
 * rings/polynomial/polynomial_compiled.pyx: 0% (0 of 20) -- just a bunch of little classes for working with "optimized" polynomials, which is actually used somewhere, evidently.  Not a single example.  Either doctest it completely, by following how it is used by polynomial_element.pyx, or delete it because maybe fast_callable is better?  But watch out for large degree, since fast_callable segfaults if the degree is large, so this polynomial_compiled.pyx might have some real value.
 

[williamstein]

comment:5

Replying to @fredrik-johansson:

The file from hell is #8791

I've put a link above, added your patch to 8791 directly to avoid confusion, and set it to "needs review". Thanks! And we will consider reviewing this a very high priority.

[williamstein]

Description changed:

--- 
+++ 
@@ -34,7 +34,7 @@
 
 * (Fredrick and Harald, #8791 -- needs review): libs/mpmath/ext_main.pyx: 7% (8 of 111) -- How the *hell* did this get into Sage with a coverage score of 7%?!!??  This was added several years after we started a 100% doctest policy.  This appeared in https://github.com/sagemath/sage-prod/issues/8159 where the positive review is from Harald Schilly: "That was easy, all tests pass ... green light from me! I'm setting this to needs_review, and start with a positiv review from me."    Yes, it's easy for all tests to pass when there are only 7% tests.  Oops.   It will be a lot of work to properly doctest this file, but much of it is straightforward (e.g., examples for the mpnumber) class. 
 
-* rings/polynomial/polynomial_compiled.pyx: 0% (0 of 20) -- just a bunch of little classes for working with "optimized" polynomials, which is actually used somewhere, evidently.  Not a single example.  Either doctest it completely, by following how it is used by polynomial_element.pyx, or delete it because maybe fast_callable is better?  But watch out for large degree, since fast_callable segfaults if the degree is large, so this polynomial_compiled.pyx might have some real value.
+* (Tom Boothby, #12033, rings/polynomial/polynomial_compiled.pyx): 0% (0 of 20) -- just a bunch of little classes for working with "optimized" polynomials, which is actually used somewhere, evidently.  Not a single example.  Either doctest it completely, by following how it is used by polynomial_element.pyx, or delete it because maybe fast_callable is better?  But watch out for large degree, since fast_callable segfaults if the degree is large, so this polynomial_compiled.pyx might have some real value.
 
 * media/wav.py: 0% (0 of 25) -- mostly just a bunch of functions to get basic info about a wave file; what makes this harder is you have to get a .wav file and put it somewhere in Sage to be used for testing.  You can probably generate an empty wave file by using the kwds constructor of the class, save it to a temp file, and use it, but that is boring. 
 

[williamstein]

Description changed:

--- 
+++ 
@@ -23,7 +23,7 @@
 
 * interfaces/psage.py: 0% (0 of 13) -- easy, since lots of examples at top
 
-* matrix/benchmark.py: 0% (0 of 29) -- easy; just run each function and paste into an EXAMPLES:: section.   This would also provide some useful additional testing.  Maybe use non-default arguments so not too slow.
+* (William Stein, #12034, matrix/benchmark.py): 0% (0 of 29) -- easy; just run each function and paste into an EXAMPLES:: section.   This would also provide some useful additional testing.  Maybe use non-default arguments so not too slow.
 
 * misc/log.py: 0% (0 of 42) -- just make object and call methods; easy
 

[williamstein]

Description changed:

--- 
+++ 
@@ -23,7 +23,7 @@
 
 * interfaces/psage.py: 0% (0 of 13) -- easy, since lots of examples at top
 
-* (William Stein, #12034, matrix/benchmark.py): 0% (0 of 29) -- easy; just run each function and paste into an EXAMPLES:: section.   This would also provide some useful additional testing.  Maybe use non-default arguments so not too slow.
+* (William Stein, #12034 -- needs review, matrix/benchmark.py): 0% (0 of 29) -- easy; just run each function and paste into an EXAMPLES:: section.   This would also provide some useful additional testing.  Maybe use non-default arguments so not too slow.
 
 * misc/log.py: 0% (0 of 42) -- just make object and call methods; easy
 

[williamstein]

Description changed:

--- 
+++ 
@@ -19,7 +19,7 @@
 
 * databases/stein_watkins.py: 0% (0 of 15) -- easy, since there are tons of examples at the top already
 
-* gsl/interpolation.pyx: 0% (0 of 9) -- easy, examples at top and easy math
+* (William Stein, #12036, gsl/interpolation.pyx): 0% (0 of 9) -- easy, examples at top and easy math
 
 * interfaces/psage.py: 0% (0 of 13) -- easy, since lots of examples at top
 

[roed314]

Description changed:

--- 
+++ 
@@ -44,7 +44,7 @@
 
 Probably Harder:
 
-* categories/action.pyx: 0% (0 of 31)
+* (David Roe, #12039 -- working on it): categories/action.pyx: 0% (0 of 31)
 
 
 Delete from SAGE?!: