new doctest for french book about Sage

[casamayou]

The attached patch includes a new doctest for a book (in french) some people are writing about Sage, under the supervision of Paul Zimmermann.

This patch contains only the doctests for two chapters (about calculus and graphics and solutions for corresponding exercices).

This doctest was tested successfully with Sage 4.6.2 under Debian Squeeze.

Apply trac_10983_doctests_from_french_book_v3.patch​

[zimmerma]

Franco, please could you review this ticket for inclusion in the Sage doctests? This will ensure stability of the example of our book (see #9395).

Thank you, Paul

[saliola]

Hello Paul. I think it is a great idea to include these tests in Sage. However, the files are just python files that contain docstrings; I suspect that one would prefer a patch instead (perhaps I am wrong about this). In order to produce a patch, one needs to first find the best place to put these tests within the sage library. Is there a standard location for such tests?

[zimmerma]

Replying to saliola:

Hello Paul. I think it is a great idea to include these tests in Sage. However, the files are just python files that contain docstrings; I suspect that one would prefer a patch instead (perhaps I am wrong about this). In order to produce a patch, one needs to first find the best place to put these tests within the sage library. Is there a standard location for such tests?

Franco, you are right. The standard location is tests/french_book where there is already the Python file corresponding to #9395.

Alexandre, please could you make a real patch? Do you know how to do that?

Paul

[casamayou]

Paul, Here is the log-message for the patch.

Alexandre.

[jdemeyer]

I am slightly worried about timings. Did you check how long the tests take? You might need to add #long time in various places.

[dimpase]

Replying to jdemeyer:

I am slightly worried about timings. Did you check how long the tests take? You might need to add #long time in various places.

the longest test took about 2 minutes on 1.8Gz Intel. Is it something that should be marked as long? Do we have any specific guidelines on this?

[jdemeyer]

Replying to dimpase:

the longest test took about 2 minutes on 1.8Gz Intel. Is it something that should be marked as long? Do we have any specific guidelines on this?

Guidelines say: anything more than 1 second (on sage.math) should be marked as long time. No test should take more than 30 seconds (even though we have several tests taking more than 30 seconds).

[dimpase]

Replying to jdemeyer:

Replying to dimpase:

the longest test took about 2 minutes on 1.8Gz Intel. Is it something that should be marked as long? Do we have any specific guidelines on this?

Guidelines say: anything more than 1 second (on sage.math) should be marked as long time. No test should take more than 30 seconds (even though we have several tests taking more than 30 seconds).

Oops, sorry. I'll revert to needs_work then! As well, there are tests that ought to be broken into several subtests - as they are really long sequences of independent tests. Would be great if this it taken care of, too.

[casamayou]

Replying to dimpase:

As well, there are tests that ought to be broken into several subtests - as they are really long sequences of independent tests. Would be great if this it taken care of, too.

By breaking tests into several subtests, did you think test as file or as code line ? Shall I break the files 'graphique*.py' and 'sol_graphiques*.py' into shortest files ?

[dimpase]

Replying to casamayou:

Replying to dimpase:

As well, there are tests that ought to be broken into several subtests - as they are really long sequences of independent tests. Would be great if this it taken care of, too.

By breaking tests into several subtests, did you think test as file or as code line ? Shall I break the files 'graphique*.py' and 'sol_graphiques*.py' into shortest files ?

Yes, this looks like a good idea . Perhaps there is also a way to have several independent tests in one file, I don't know.

[jdemeyer]

Replying to dimpase:

Perhaps there is also a way to have several independent tests in one file, I don't know.

Of course you can have several independent tests in one file, just like you can have more than one function or more than one class in a file. Just put several test blocks in the file:

r"""
test1
"""

r"""
test2
"""

instead of

r"""
test1
test2
"""

[casamayou]

Replying to jdemeyer:

Replying to dimpase:

Perhaps there is also a way to have several independent tests in one file, I don't know.

Of course you can have several independent tests in one file, just like you can have more than one function or more than one class in a file. Just put several test blocks in the file:

Attached you will find the test files corrected.

The too long tests were commented: # long time

Does that sound correct ?

Here are the execution time :

$ for i in *.py ; do sage -t $i; done; 
sage -t  "calculus_fixed.py"                                
	 [11.2 s]
 
----------------------------------------------------------------------
All tests passed!
Total time for all tests: 11.2 seconds
sage -t  "graphique_1_fixed.py"                             
	 [14.9 s]
 
----------------------------------------------------------------------
All tests passed!
Total time for all tests: 15.0 seconds
sage -t  "graphique_2_fixed.py"                             
	 [19.4 s]
 
----------------------------------------------------------------------
All tests passed!
Total time for all tests: 19.4 seconds
sage -t  "graphique_3_fixed.py"                             
	 [8.3 s]
 
----------------------------------------------------------------------
All tests passed!
Total time for all tests: 8.4 seconds
sage -t  "sol_calculus_fixed.py"                            
	 [8.9 s]
 
----------------------------------------------------------------------
All tests passed!
Total time for all tests: 9.0 seconds
sage -t  "sol_graphiques_1_fixed.py"                        
	 [13.0 s]
 
----------------------------------------------------------------------
All tests passed!
Total time for all tests: 13.1 seconds
sage -t  "sol_graphiques_2_fixed.py"                        
	 [4.8 s]
 
----------------------------------------------------------------------
All tests passed!
Total time for all tests: 4.8 seconds

And with the long tests :

$ for i in *.py ; do sage -t -long $i; done; 
sage -t -long "calculus_fixed.py"                           
	 [11.2 s]
 
----------------------------------------------------------------------
All tests passed!
Total time for all tests: 11.3 seconds
sage -t -long "graphique_1_fixed.py"                        
	 [323.1 s]
 
----------------------------------------------------------------------
All tests passed!
Total time for all tests: 323.1 seconds
sage -t -long "graphique_2_fixed.py"                        
	 [20.3 s]
 
----------------------------------------------------------------------
All tests passed!
Total time for all tests: 20.3 seconds
sage -t -long "graphique_3_fixed.py"                        
	 [8.1 s]
 
----------------------------------------------------------------------
All tests passed!
Total time for all tests: 8.1 seconds
sage -t -long "sol_calculus_fixed.py"                       
	 [8.9 s]
 
----------------------------------------------------------------------
All tests passed!
Total time for all tests: 9.0 seconds
sage -t -long "sol_graphiques_1_fixed.py"                   
	 [47.3 s]
 
----------------------------------------------------------------------
All tests passed!
Total time for all tests: 47.3 seconds
sage -t -long "sol_graphiques_2_fixed.py"                   
	 [64.9 s]
 
----------------------------------------------------------------------
All tests passed!
Total time for all tests: 65.0 seconds

It was the execution of the following instruction that is very long :

sage: f = lambda x, y : (abs(cos((x + I * y) ** 4)) - 1)
sage: g2 = implicit_plot(f, (-3, 3), (-3, 3), plot_points=400) # long time 

[casamayou]

updated file

[casamayou]

updated file

[casamayou]

updated file

[casamayou]

updated file

[casamayou]

updated file

[casamayou]

updated file

[casamayou]

updated file

[casamayou]

The files *_fixed.py are outdated, because they were tested under sage 7.4.1.

The following files have been updated and tested under sage 7.4.2 and are the only ones to consider:

calculus.py graphique_1.py graphique_2.py graphique_3.py sol_calculus.py sol_graphiques_1.py sol_graphiques_2.py

[casamayou]

One more question.

Should I create a new file trac_10983.patch with a "hg commit" ?

(Sorry... I'm a newbie)

[dimpase]

Replying to casamayou:

One more question.

Should I create a new file trac_10983.patch with a "hg commit" ?

(Sorry... I'm a newbie)

hg commit is what you might sometimes do after applying a patch. It certainly does not create patches. You can create patches using hg, e.g. by hg import. But you don't have to. I'll look at your updates later today.

[mmezzarobba]

Here are updated tests that correspond to the current version of the book. (The division of chapters changed a little bit, but the tests I'm including here roughly correspond to those posted by Alexandre.)

On my machine, using Sage-5.9β0:

Running doctests with ID 2013-03-22-21-36-07-c7afe8cf.
Doctesting 3 files.
sage -t calculus_doctest.py
    [246 tests, 6.9 s]
sage -t graphique_doctest.py
    [167 tests, 26.8 s]
sage -t premierspas_doctest.py
    [49 tests, 1.5 s]
----------------------------------------------------------------------
All tests passed!
----------------------------------------------------------------------
Total time for all tests: 35.4 seconds
    cpu time: 30.6 seconds
    cumulative wall time: 35.2 seconds

[ncohen]

Most of the doctests break on sage 5.1.beta3. Most output are ordered differently, so there's nothing to be worried about either.

----------------------------------------------------------------------
sage -t --long calculus_doctest.py  # 12 doctests failed
sage -t --long graphique_doctest.py  # 1 doctest failed
----------------------------------------------------------------------

Nathann

Apply trac_10983_doctests_from_french_book.patch

[ncohen]

Ahem. 5.11.beta3 :-P

Nathann

[zimmerma]

Combined patch

[zimmerma]

I've made a new patch with a different ordering in the outputs, due to a recent change in Sage (I can't find the corresponding ticket). Unfortunately the book output now slightly differs with what we get in Sage. This patch works fine with Sage 5.11.

Paul

[jdemeyer]

The ticket which changed the ordering is #9880.

[chapoton]

indentation should be 4 spaces rather than 2

[chapoton]

for patchbot:

apply trac_10983_doctests_from_french_book.patch​

[chapoton]

apply trac_10983_doctests_from_french_book.patch

[chapoton]

I have made a new patch with correct indentations.

Unfortunately, there are now 2 doctest errors with simplify_full !!!

for the patchbots:

apply trac_10983_doctests_from_french_book_v2.patch

[dimpase]

Replying to chapoton:

I have made a new patch with correct indentations.

Unfortunately, there are now 2 doctest errors with simplify_full !!!

How about the following changes: (I suppose this also means minor corrections to the main text, which I can't do)

diff --git a/sage/tests/french_book/calculus_doctest.py b/sage/tests/french_book/calculus_doctest.py
--- a/sage/tests/french_book/calculus_doctest.py
+++ b/sage/tests/french_book/calculus_doctest.py
@@ -643,7 +643,7 @@
     sage: assume(8*n+1>0)
     sage: u(n) = integrate((4*sqrt(2)-8*t^3-4*sqrt(2)*t^4\
     ....:                  -8*t^5) * t^(8*n), t, 0, 1/sqrt(2))
-    sage: (u(n)-v(n)).simplify_full()
+    sage: (u(n)-v(n)).simplify_exp()
     0
 
 Sage example in ./sol/calculus.tex, line 258::
@@ -652,8 +652,8 @@
     sage: J = integrate((4*sqrt(2)-8*t^3 \
     ....:      - 4*sqrt(2)*t^4-8*t^5)\
     ....:      / (1-t^8), t, 0, 1/sqrt(2))
-    sage: J.simplify_full()
-    pi + 2*log(sqrt(2) + 1) + 2*log(sqrt(2) - 1)
+    sage: J.simplify_log().simplify_full()
+    pi
 
 Sage example in ./sol/calculus.tex, line 272::
 

[zimmerma]

the second doctest failure looks like a regression in Sage, since the expression 2*e is not simplified similarly to e:

sage: e=log(1/2*sqrt(2)*(sqrt(2) + 1)) + log(1/2*sqrt(2)*(sqrt(2) - 1))
sage: e.simplify_full()
-log(2)
sage: (2*e).simplify_full()
2*log(1/2*sqrt(2)*(sqrt(2) + 1)) + 2*log(1/2*sqrt(2)*(sqrt(2) - 1))

Should I open a ticket for that?

Paul

[dimpase]

Replying to zimmerma:

the second doctest failure looks like a regression in Sage, since the expression 2*e is not simplified similarly to e:

sage: (2*e).simplify_full()
2*log(1/2*sqrt(2)*(sqrt(2) + 1)) + 2*log(1/2*sqrt(2)*(sqrt(2) - 1))

Should I open a ticket for that?

Just as in the patch proposed, you may chain simplifies, getting

sage: (2*e).simplify_log().simplify_full()
-log(4)

Not sure it's a Maxima regression, or a proper Sage regression. Opening a trac ticket won't hurt, I suppose.

[zimmerma]

Opening a trac ticket won't hurt, I suppose.

done, cf #15362.

Paul

[zimmerma]

a simpler fix is to replace simplify_full by simplify_radical in the two failing doctests. See #15362. Since that change also works before #12737 (merged in Sage 5.12) I will do the change in our book too.

Paul

[zimmerma]

with the new patch, the doctests pass again. Ready for review.

Paul

[dimpase]

looks good to me.