Opened 22 months ago
Closed 21 months ago
#22833 closed defect (fixed)
fix a calculus doctest (giac, laplace, integration)
Reported by:  chapoton  Owned by:  

Priority:  major  Milestone:  sage8.0 
Component:  calculus  Keywords:  
Cc:  rws, frederichan  Merged in:  
Authors:  Frédéric Chapoton, Marcelo Forets  Reviewers:  Travis Scrimshaw, Steven Trogdon 
Report Upstream:  N/A  Work issues:  
Branch:  e2d74c6 (Commits)  Commit:  e2d74c645b48c3288802f9080d6651aaccf658fb 
Dependencies:  Stopgaps: 
Description (last modified by )
that currently prevents many patchbots to run smoothly
Change History (47)
comment:1 Changed 22 months ago by
 Branch set to u/chapoton/22833
 Status changed from new to needs_review
comment:2 Changed 22 months ago by
 Commit set to 2186acae2da3826b0a29d196d6b22b8fbc1dad4b
comment:3 Changed 22 months ago by
 Reviewers set to Travis Scrimshaw
 Status changed from needs_review to positive_review
I thought someone had already fixed this, but I guess not.
comment:4 Changed 22 months ago by
the error was because of the double ::
in TETS or the Unable to parse Giac output
? or both? and how did you know that integration
is ok? (i still get the NotImplementedError
in sage8.0beta2) because the "dummy integrate" keyword of maxima is integrate
.
comment:5 Changed 22 months ago by
The change to TESTS:: was just something I could not keep like that once I had seen it.
I am not sure that the change to integration doctest is ok. If it fails on your machine, then you can set this ticket back to needs work.
comment:6 Changed 22 months ago by
 Status changed from positive_review to needs_work
failing in my machine with:
Doctesting 1 file. sage t src/sage/calculus/calculus.py ********************************************************************** File "src/sage/calculus/calculus.py", line 1365, in sage.calculus.calculus.laplace Failed example: (p1+p2).save(os.path.join(SAGE_TMP, "de_plot.png")) Expected nothing Got: doctest:warning File "/Users/forets/sagesrc/sage/src/bin/sageruntests", line 89, in <module> err = DC.run() File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/doctest/control.py", line 1134, in run self.run_doctests() File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/doctest/control.py", line 858, in run_doctests self.dispatcher.dispatch() File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/doctest/forker.py", line 1705, in dispatch self.parallel_dispatch() File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/doctest/forker.py", line 1595, in parallel_dispatch w.start() # This might take some time File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/doctest/forker.py", line 1871, in start super(DocTestWorker, self).start() File "/Users/forets/sagesrc/sage/local/lib/python2.7/multiprocessing/process.py", line 130, in start self._popen = Popen(self) File "/Users/forets/sagesrc/sage/local/lib/python2.7/multiprocessing/forking.py", line 126, in __init__ code = process_obj._bootstrap() File "/Users/forets/sagesrc/sage/local/lib/python2.7/multiprocessing/process.py", line 258, in _bootstrap self.run() File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/doctest/forker.py", line 1844, in run task(self.options, self.outtmpfile, msgpipe, self.result_queue) File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/doctest/forker.py", line 2137, in __call__ runner.run(test) File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/doctest/forker.py", line 641, in run return self._run(test, compileflags, out) File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/doctest/forker.py", line 503, in _run self.compile_and_execute(example, compiler, test.globs) File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/doctest/forker.py", line 866, in compile_and_execute exec(compiled, globs) File "<doctest sage.calculus.calculus.laplace[23]>", line 1, in <module> (p1+p2).save(os.path.join(SAGE_TMP, "de_plot.png")) File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/misc/decorators.py", line 471, in wrapper return func(*args, **kwds) File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/plot/graphics.py", line 3170, in save figure = self.matplotlib(**options) File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/plot/graphics.py", line 2566, in matplotlib from matplotlib.figure import Figure File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/matplotlib/figure.py", line 38, in <module> import matplotlib.colorbar as cbar File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/matplotlib/colorbar.py", line 34, in <module> import matplotlib.collections as collections File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/matplotlib/collections.py", line 27, in <module> import matplotlib.backend_bases as backend_bases File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/matplotlib/backend_bases.py", line 62, in <module> import matplotlib.textpath as textpath File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/matplotlib/textpath.py", line 15, in <module> import matplotlib.font_manager as font_manager File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/matplotlib/font_manager.py", line 1413, in <module> fontManager = pickle_load(_fmcache) File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/matplotlib/font_manager.py", line 965, in pickle_load data = pickle.load(fh) File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/matplotlib/font_manager.py", line 1044, in __init__ self.ttffiles = findSystemFonts(paths) + findSystemFonts() File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/matplotlib/font_manager.py", line 324, in findSystemFonts for f in get_fontconfig_fonts(fontext): File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/matplotlib/font_manager.py", line 273, in get_fontconfig_fonts warnings.warn('Matplotlib is building the font cache using fclist. This may take a moment.') : UserWarning: Matplotlib is building the font cache using fclist. This may take a moment. ********************************************************************** File "src/sage/calculus/calculus.py", line 1407, in sage.calculus.calculus.laplace Failed example: laplace(t^n, t, s, algorithm='giac') Exception raised: Traceback (most recent call last): File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/doctest/forker.py", line 503, in _run self.compile_and_execute(example, compiler, test.globs) File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/doctest/forker.py", line 866, in compile_and_execute exec(compiled, globs) File "<doctest sage.calculus.calculus.laplace[34]>", line 1, in <module> laplace(t**n, t, s, algorithm='giac') File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/calculus/calculus.py", line 1456, in laplace return result.sage() File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/interfaces/interface.py", line 1053, in sage return self._sage_(*args, **kwds) File "/Users/forets/sagesrc/sage/local/lib/python2.7/sitepackages/sage/interfaces/giac.py", line 1095, in _sage_ raise NotImplementedError("Unable to parse Giac output: %s" % result) NotImplementedError: Unable to parse Giac output: integrate(t^n*exp(s*t),t,0,+infinity) ********************************************************************** 1 item had failures: 2 of 39 in sage.calculus.calculus.laplace [419 tests, 2 failures, 11.53 s]  sage t src/sage/calculus/calculus.py # 2 doctests failed  Total time for all tests: 11.6 seconds cpu time: 10.8 seconds cumulative wall time: 11.5 seconds
comment:7 Changed 22 months ago by
The first failure is not a true failure and is independent. We need to figure out what makes this test pass on some systems and not on others.
comment:8 Changed 22 months ago by
 Description modified (diff)
 Summary changed from fix a calculus doctest to fix a calculus doctest (giac, laplace, integration)
comment:9 Changed 22 months ago by
I see this failure with giac 1.2.3.
The branch works for me.
This problem is rather annoying, it prevents my patchbot from giving any green light..
comment:10 Changed 22 months ago by
@mforets, could you try the following, please:
sage: var('t,n,s') sage: ex=t^n sage: from sage.interfaces.giac import giac sage: giac.laplace(ex, t, s) integration(t^n*exp(s*t),t,0,+infinity)
comment:11 Changed 22 months ago by
 Commit changed from 2186acae2da3826b0a29d196d6b22b8fbc1dad4b to 200ba71b935a6573ee9aac38a35f0eefd5507b5d
comment:12 Changed 22 months ago by
 Commit changed from 200ba71b935a6573ee9aac38a35f0eefd5507b5d to 79442fe1dd9e24059f3c88db153c4f84dd090bfe
Branch pushed to git repo; I updated commit sha1. New commits:
79442fe  trac 222833 detail

comment:13 followup: ↓ 14 Changed 22 months ago by
And also tell me what you get when doing
sage: %giac giac: Int(exp(x),x,0,1) integration(exp(x),x,0,1)
comment:14 in reply to: ↑ 13 Changed 22 months ago by
Replying to chapoton:
And also tell me what you get when doing
sage: %giac giac: Int(exp(x),x,0,1) integration(exp(x),x,0,1)
EDIT: no no, ignore my last message, I hadn't notice that you pushed changes < 1hr ago.
i've recompiled this branch and i'm getting:
sage: %giac > Switching to Giac < giac: Int(exp(x),x,0,1) integrate(exp(x),x,0,1)
comment:15 followups: ↓ 16 ↓ 17 Changed 22 months ago by
in my machine, and for the commit 79442f3, i have
sage: var('t,n,s') (t, n, s) sage: ex=t^n sage: from sage.interfaces.giac import giac sage: giac.laplace(ex, t, s) integrate(t^n*exp(s*t),t,0,+infinity) sage: %giac > Switching to Giac < giac: Int(exp(x),x,0,1) integrate(exp(x),x,0,1)
and i still get:
sage: laplace(t^n*exp(s*t),t,s,algorithm='giac') ... NotImplementedError: Unable to parse Giac output: integrate(t^n*exp(2*s*t),t,0,+infinity)
this exception is raised when we transform back to sage. maybe it's better to mimick the unevaluated laplace as with the SymPy? case, that is:
... if 'integrate' in format(result): return dummy_laplace(ex, t, s) else: return result.sage()
i git pulled and recompiled everything, so i ignore why you get "integration" instead.
that said, if you have pressing need then you may as well consider removing this "nogo" test altogether..
comment:16 in reply to: ↑ 15 Changed 22 months ago by
Replying to mforets:
... if 'integrate' in format(result): return dummy_laplace(ex, t, s) else: return result.sage()
with that modif, the expected output for the test at hand becomes: laplace(t^n, t, s)
in my machine the updated code passes. let me know if you want me to push that or try something else.
comment:17 in reply to: ↑ 15 Changed 22 months ago by
Replying to mforets:
and i still get:
sage: laplace(t^n*exp(s*t),t,s,algorithm='giac') ... NotImplementedError: Unable to parse Giac output: integrate(t^n*exp(2*s*t),t,0,+infinity)
Same here.
comment:18 Changed 22 months ago by
Would there be need for symbolic laplace? Maxima might return such unevaluated results as anonymous functions too, and they are difficult to handle.
comment:19 Changed 22 months ago by
in this case it is raising a NotImplementedError
, which i find a bit misleading if this kind of exception belongs to the level of Sage the library. perhaps an intermediate point is to return just integrate(t^n*exp(2*s*t),t,0,+infinity)
, what do you think?
on the other hand, Laplace placeholder could be needed here (in opposition to a naked integral):
 to attach properties of the transform which are not 'obvious', eg. properties as a function of a complex variable s, convergence radius, final and initial value theorems
 if you have many of them, it looks nicer with respect to printing the equations!
comment:20 Changed 22 months ago by
Got an answer from Parisse:
si vous etes en francais (export LANG=fr_FR.UTF8 par exemple), les motsclefs de giac qui ont une traduction s'affichent en francais. Le fichier qui contient les traductions sous linux est /usr/share/giac/doc/fr/keywords. Si vous passez en export LANG=en_US.UTF8 ca restera en anglais. Dans l'interface de Xcas les fichiers de session sont sauvegardes avec les motsclefs en anglais.
comment:21 Changed 22 months ago by
So what he is saying is that it has to with the localization settings?
comment:22 Changed 22 months ago by
I think so. The English "integrate" got translated to "integration" in the output for people whose locale is French.
comment:23 Changed 22 months ago by
I think there is a deeper issue. I suspect that is failing out at some other place because it can't find the symbol integration
in the symbol table. At least this is coming from trying to unravel things and get a better traceback:
sage: t,n,s = var('t,n,s') sage: from sage.libs.pynac.pynac import symbol_table sage: from sage.calculus.calculus import symbolic_expression_from_string sage: ex = t^n*exp(s*t) sage: result = ex._giac_().integrate(t._giac_(), 0._giac_(), oo._giac_()) sage: symbolic_expression_from_string(repr(result), symbol_table['giac'], True) ... /home/travis/sagebuild/local/lib/python2.7/sitepackages/sage/interfaces/maxima_lib.pyc in sr_integral(self, *args) 808 raise ValueError("Integral is divergent.") 809 elif "Is" in s: # Maxima asked for a condition > 810 self._missing_assumption(s) 811 else: 812 raise /home/travis/sagebuild/local/lib/python2.7/sitepackages/sage/interfaces/maxima_lib.pyc in _missing_assumption(self, errstr) 1018 + errstr[jj+1:k] +">0)', see `assume?` for more details)\n" + errstr 1019 outstr = outstr.replace('_SAGE_VAR_','') > 1020 raise ValueError(outstr) 1021 1022 def is_MaximaLibElement(x): ValueError: Computation failed since Maxima requested additional constraints; using the 'assume' command before evaluation *may* help (example of legal syntax is 'assume(s>0)', see `assume?` for more details) Is s positive, negative or zero?
Do you get a similar error?
comment:24 Changed 22 months ago by
no, I just get
integration(t^n*e^(s*t), t, 0, +Infinity)
comment:25 Changed 22 months ago by
I think the problem is actually with Maxima. What happens when you try to explicitly evaluate the integral?
sage: integrate(t^n*exp(s*t),t,0,oo)
I get the same error as above when I do this. If I impose s > 0
and n
in Z_{>0}, then I get
s^(n  1)*gamma(n + 1)
comment:26 Changed 22 months ago by
sage: integrate(t^n*exp(s*t),t,0,oo)
gives me the same "assume" failure as you got ;
and then works as for you with the assumes done
comment:27 Changed 22 months ago by
Hmm...that is strange. So something is happening between the attempt to parse the giac output and actually evaluate the maxima integral.
comment:28 Changed 22 months ago by
Is 'integration'
a key in your symbol_table['functions']
?
comment:29 Changed 22 months ago by
Also, do you have
sage: symbol_table['giac'] {'(1+sqrt(5))/2': golden_ratio, 'Dirac': dirac_delta, 'Heaviside': heaviside, 'pi': pi}
comment:30 Changed 22 months ago by
'integrate': integrate, 'integration': integration,
in symbol_tablefunctions?
and
sage: symbol_table['giac'] {'(1+sqrt(5))/2': golden_ratio, 'Dirac': dirac_delta, 'Heaviside': heaviside, 'pi': pi}
comment:31 Changed 22 months ago by
That's the difference: I don't have the 'integration'
in the functions symbol table. What is the integration
function, specifically, what does import_statements
give?
comment:32 Changed 22 months ago by
sage: import_statements('integration') # **Warning**: distinct objects with name 'integration' in: #  sage.calculus #  sage.symbolic import sage.symbolic.integration
comment:33 Changed 22 months ago by
Sorry, I was vague, I meant:
sage: import_statements(symbol_table['functions]['integration'])
comment:34 Changed 22 months ago by
sage: symbol_table['functions']['integration'] integration sage: import_statements(_)  ValueError Traceback (most recent call last) ValueError: no import statement found for 'integration'.
comment:35 Changed 22 months ago by
sage: symbol_table['functions']['integration'] integration sage: type(_) <class 'sage.symbolic.function_factory.NewSymbolicFunction'>
comment:36 Changed 22 months ago by
Now I'm not so sure about what to look for or ask, but I'm investigating. Ralf, do you have any ideas?
comment:37 Changed 22 months ago by
So I think what is happening is that integration
is being considered by Maxima as a formal function. Hence, it is not simplified as an integral, and I don't think you could manipulate it as a function.
comment:38 Changed 22 months ago by
does this pass?
... if 'integrate' in format(result) or 'integration' in format(result): return dummy_laplace(ex, t, s) else: return result.sage()
comment:39 Changed 22 months ago by
Let me try to summarize. We cannot fix the doctest because of conflicting results due to giac's clever use of locale which breaks its interface to the world. The immediate way to adapt to this IMO is to look for both integrate
and integration
in the result when using laplace of giac. The more farsighted way would be to fix the giacSage interface so that in the laplace code we can just say return result.sage()
unconditionally.
comment:40 Changed 22 months ago by
could somebody please take action here, and propose a better branch ?
comment:41 Changed 22 months ago by
What is the desired output of
sage: laplace(t^n, t, s, algorithm='giac')
If it is integration(t^n*e^(s*t), t, 0, +Infinity)
then one could add
integrate integration
to local/share/giac/doc/en/keywords
and any other provided $LANG/keywords
file. Doing this I get
sage: var('t,n,s') (t, n, s) sage: ex=t^n sage: from sage.interfaces.giac import giac sage: giac.laplace(ex, t, s) integration(t^n*exp(s*t),t,0,+infinity) sage: %giac > Switching to Giac < giac: Int(exp(x),x,0,1) integration(exp(x),x,0,1)
Without the change the doctest would pass but with the change I get
sage t long src/sage/calculus/calculus.py ********************************************************************** File "src/sage/calculus/calculus.py", line 1406, in sage.calculus.calculus.laplace Failed example: laplace(t^n, t, s, algorithm='giac') Expected: Traceback (most recent call last): ... NotImplementedError: Unable to parse Giac output: integrate(t^n*exp(s*t),t,0,+infinity) Got: integration(t^n*e^(s*t), t, 0, +Infinity)
comment:42 Changed 22 months ago by
again, for consistency the desired output of laplace(t^n, t, s, algorithm='giac')
should be the same as the other two interfaces when they don't know the answer, don't you think?
from your answers, i can reproduce the issue playing with $ export LANG=fr_FR.UTF8
and $ export LANG=en_US.UTF8
in the shell :)
hence, let me upload a branch doing rws's comment:39 and my comment:38, if i understood correctly.
comment:43 Changed 22 months ago by
 Branch changed from u/chapoton/22833 to u/mforets/22833
 Cc frederichan added
 Commit changed from 79442fe1dd9e24059f3c88db153c4f84dd090bfe to e2d74c645b48c3288802f9080d6651aaccf658fb
 Status changed from needs_work to needs_review
New commits:
e2d74c6  parse unevaluated expression in EN and FR

comment:44 Changed 21 months ago by
Just to confirm that merging this branch in Sage 8.0.beta6 fixes the doctest error on my system.
comment:45 Changed 21 months ago by
 Reviewers changed from Travis Scrimshaw to Travis Scrimshaw, Steven Trogdon
 Status changed from needs_review to positive_review
I think this is good.
comment:46 Changed 21 months ago by
comment:47 Changed 21 months ago by
 Branch changed from u/mforets/22833 to e2d74c645b48c3288802f9080d6651aaccf658fb
 Resolution set to fixed
 Status changed from positive_review to closed
Branch pushed to git repo; I updated commit sha1. New commits:
fixing calculus doctest (giac)