Hi Oliver,
the error was caused by a modification to the BioPython blast module. I've fixed
the issue -- please update to the latest svn revision. All Biskit.Mod test cases
run now through. The Pymoler issue is still out there though. A brief bug report
on this one would be nice.
Enjoy your weekend!
Raik
Oliver Kuhn wrote:
> Hi Raik,
> thanks for answering so quickly!
> I will be away till Monday and then work on it.
> have a nice day (and weekend)!
> Oliver
>
>
> On Wed, 2008-07-16 at 11:59 +0200, Raik Gruenberg wrote:
>> Hi Oliver,
>>
>> that the general test hangs at PyMol is a bug that I haven't seen before. If you
>> find the time, please put a brief bug report in (you can also do that without
>> having a sourceforge login).
>>
>>> ok. let's start with the test suite:
>>> ( "python test.py -i Biskit.Mod -e pvm" gives me:
>> Should be:
>> python test.py -p Biskit.Mod -e pvm
>>
>> (you get a help screen of all options with 'python test.py' )
>>
>>> Error: unrecognized tags: ['Biskit.Mod']
>>> <type 'exceptions.NameError'> in test.py line 557:
>>> ("name 'BISKIT' is not defined",). )
>> Not a very helpful error message, I admit. the -i option is meant for tags (like
>> include long or short tests) but not for packages (like Biskit.Mod)
>>
>> Last night, I installed the latest stack of blast / tcoffee / modeller on my PC
>> at home and I can more or less reproduce the error you reported earlier. At a
>> first glance it looks like either BioPythonists have changed their Blast
>> interface or this is yet another case of a change in Blast output that has not
>> been captured by BioPython yet. I wonder why those gyus have to constantly play
>> around with their output and APIs. Anyway, I've got to dive into that when I am
>> back at home this evening.
>>
>> In the meantime, if you have to get things urgently running, you can either try
>> to have a look at the code yourself:
>>
>> The problem occurs in SequenceSearcher.py SequenceSearcher.localBlast
>>
>> or you downgrade BioPython and ncbi tools to versions that are about 4-6 months
>> old.
>>
>> Greetings,
>> Raik
>>
>>
>>> but:
>>> python test.py -e pvm
>>> collecting 'Biskit'
>>> collecting 'Biskit.Dock'
>>> collecting 'Biskit.Mod'
>>> collecting 'Biskit.PVM'
>>> collecting 'Biskit.Statistics'
>>> AmberRstParser.getXyz test ... ok
>>> AmberRstParser.getModel test ... ok
>>> AmberRstParser.getXyz test ... ok
>>> ChainSeparator test ... ok
>>> ModelList test ... ok
>>> Errors test (empty test) ... ok
>>> PDBParseModel test ... ok
>>> ChainWriter test ... ok
>>> msms test ... ok
>>> SettingsManager test ... ok
>>> surfaceRacerTools test ... ok
>>> Trajectory test ... ok
>>> PCRModel test ... ok
>>> rmsFit test ... ok
>>> gnuplot.parallelAxesPlot test (interactive only) ... ok
>>> gnuplot.plot test ... ok
>>> gnuplot.plot to file test ... ok
>>> gnuplot.scatter test (interactive only) ... ok
>>> ChainCleaner test ... ok
>>> AmberParmBuilder.parmMirror test ... /home/oak/tools/amber9/exe/teLeap:
>>> Exec format error. Binary file not executable.
>>> ERROR
>>> AmberParmBuilder.parmSolvated test ... 1HPT: Removing multiple
>>> occupancies of atoms ...
>>> No atoms with multiple occupancies to remove
>>> 1HPT: Looking for non-standard residue names...
>>> Found 0 atoms with non-standard residue names.
>>> Checking content of standard amino-acids...
>>> Removed 0 atoms because they were non-standard or followed a missing
>>> atom.
>>> /home/oak/tools/amber9/exe/teLeap: Exec format error. Binary file not
>>> executable.
>>> ERROR
>>> MatrixPlot test ... ok
>>> FuzzyCluster test ... ok
>>> Blast2Seq test ... ok
>>> HmmerProfile test ... ok
>>> HmmSearch test ... ok
>>> Hmmer test ... ok
>>> match2seq test ... ok
>>> SparseArray test ... ok
>>> decorators test ... ok
>>> PDBModel chain break handling and writing test ... ok
>>> PDBModel chain methods test ... ok
>>> PDBModel single residue chain test ... ok
>>> PDBModel.removeRes test ... ok
>>> PDBModel renameAmberRes tests ... ok
>>> PDBModel report&plot test ... ok
>>> PDBModel sorting test ... ok
>>> PDBModel source / disconnection tests ... ok
>>> PDBModel xplor2amber test ... ok
>>> SurfaceRacer test ... ok
>>> EnsembleTraj.fit/fitMembers/plotMembers test ... ok
>>> EnsembleTraj.outliers/concat test ... ok
>>> molUtils test ... ok
>>> CommandLine.reportArguments test ... ok
>>> CommandLine.parse test ... ok
>>> CommandLine.makeTest test ... ok
>>> Xplorer test ... ok
>>> Fold_X test ... ERROR
>>> Pymoler test ...
>>>
>>> and then hangs up testing pymol (on a debian system with pymol
>>> installed) - don't care about that fold_x and amber error.
>>>
>>> greets
>>> Oliver
>>>
>>> On Tue, 2008-07-15 at 14:52 +0200, Raik Gruenberg wrote:
>>>> Hi Oliver,
>>>>
>>>> Before doing anything else, could you please (1) update Biskit to the latest SVN
>>>> version (if you are working from the older released version):
>>>>
>>>> cd biskit
>>>> svn update
>>>>
>>>> and then (2) run the test suite?
>>>>
>>>> cd biskit/Biskit
>>>> python test.py -i Biskit.Mod -e pvm
>>>>
>>>> ... will test all homology modelling classes but not the pvm-related ones.
>>>>
>>>> Then let's take it from there.
>>>>
>>>> Oliver Kuhn wrote:
>>>>> Dear Raik,
>>>>>
>>>>> I have problems with both functions:
>>>>>
>>>>> First, search_templates.py doesn't write me a chain_index.txt.
>>>>>
>>>>> BLAST give the following deprecated warning:
>>>>> /var/lib/python-support/python2.5/Bio/Blast/NCBIXML.py:153:
>>>>> DeprecationWarning: Bio.Blast.NCBIXML.BlastParser.parse has been
>>>>> deprecated; please use Bio.Blast.NCBIXML.parse instead
>>>>> warnings.warn("Bio.Blast.NCBIXML.BlastParser.parse has been
>>>>> deprecated; please use Bio.Blast.NCBIXML.parse instead",
>>>>> DeprecationWarning)
>>>>> I think thats not the problem (the pdb fetch works)!
>>>> Ah well, the BioPython parsers are constantly changing to chase after the
>>>> constantly changing blast output... I haven't worked with this BioPython version
>>>> yet. Looks like the code has to be adapted once more but that seems indeed not
>>>> to be the problem.
>>>>
>>>>> Then comes the clustering with the following error:
>>>>> - Command: /usr/bin/blastclust
>>>>> -i /home/oak/works/biskit/1OHR/templates/all.fasta -S 1.750000 -L
>>>>> 0.900000 -a 1
>>>>> Traceback (most recent call last):
>>>>> File "/home/oak/tools/biskit/scripts/Mod/search_templates.py", line
>>>>> 151, in <module>
>>>>> searcher.clusterFastaIterative( simCut=simCut, lenCut=simLen,
>>>>> ncpu=nCpu )
>>>>> File "/home/oak/tools/biskit/Biskit/Mod/SequenceSearcher.py", line
>>>>> 710, in clusterFastaIterative
>>>>> self.clusterFasta( fastaIn, simCut, lenCut, ncpu )
>>>>> File "/home/oak/tools/biskit/Biskit/Mod/SequenceSearcher.py", line
>>>>> 668, in clusterFasta
>>>>> raise BlastError( err )
>>>>> Biskit.Mod.SequenceSearcher.BlastError: 256
>>>>>
>>>> Let's first check whether the unit tests fail or whether this bug is specific to
>>>> your target.
>>>>
>>>> Greetings,
>>>> Raik
>>>>
>>>>> Second, I wrote the chain_index.txt manually and started
>>>>> clean_templates.py which raises the following error:
>>>>> Error cleaning /home/oak/works/biskit/1OHR/templates/nr/1K6C.pdb
>>>>> <type 'exceptions.ValueError'>
>>>>> in /home/oak/tools/biskit/Biskit/Mod/TemplateCleaner.py line 270:
>>>>> ('Input argument must be 1d',).
>>>>>
>>>>> Would be great if you can help me with that.
>>>>>
>>>>> Best regards
>>>>> Oliver
>>>>>
>>>>>
>>>>>
--
________________________________
Dr. Raik Gruenberg
http://www.raiks.de/contact.html
________________________________
|