Re: [galaxy-dev] Functional Tests and ftype
On Fri, Apr 18, 2014 at 3:36 PM, Saket Choudhary wrote: > Thanks John, > > That really helped: > > ... > > | > /usr/local/lib/python2.7/dist-packages/setuptools-2.2-py2.7.egg/pkg_resources.py:991: > UserWarning: /home/travis/.python-eggs is writable by group/others and > vulnerable to attack when used with get_resource_filename. Consider a > more secure location (set with .set_extraction_path or the > PYTHON_EGG_CACHE environment variable). > > Fixed here: https://travis-ci.org/saketkc/galaxy_tools/jobs/23282261#L1402 > with: > > before_install: > - mkdir "$HOME/.python-eggs" > - chmod 700 "$HOME/.python-eggs" > > > > The test is failing, but that I expected it to anyway ;) > > Thanks everyone! Especially John and Peter! > > Saket Ah - that was hard to diagnose, perhaps there used to be a more visible warning because I already had that egg chmod trick in my TravsCI setup? Well done for solving that - and thank you John for the verbose setting, I think that will be useful in future too. Peter ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional Tests and ftype
Thanks John, That really helped: Exception: History in error state. >> begin captured stdout << - History with id adb5f5c93f827949 in error - summary of datasets in error below. -- | 1 - filter1_test5.tab (HID - NAME) | Dataset Blurb: | error | Dataset Info: | uploaded tabular file | /usr/local/lib/python2.7/dist-packages/setuptools-2.2-py2.7.egg/pkg_resources.py:991: UserWarning: /home/travis/.python-eggs is writable by group/others and vulnerable to attack when used with get_resource_filename. Consider a more | Dataset Job Standard Output: | *Standard output was empty.* | Dataset Job Standard Error: | /usr/local/lib/python2.7/dist-packages/setuptools-2.2-py2.7.egg/pkg_resources.py:991: UserWarning: /home/travis/.python-eggs is writable by group/others and vulnerable to attack when used with get_resource_filename. Consider a more secure location (set with .set_extraction_path or the PYTHON_EGG_CACHE environment variable). | -- - >> end captured stdout << -- Fixed here: https://travis-ci.org/saketkc/galaxy_tools/jobs/23282261#L1402 with: before_install: - mkdir "$HOME/.python-eggs" - chmod 700 "$HOME/.python-eggs" The test is failing, but that I expected it to anyway ;) Thanks everyone! Especially John and Peter! Saket On 18 April 2014 19:15, John Chilton wrote: > Pushed out this changeset > https://bitbucket.org/galaxy/galaxy-central/commits/183668ed911ca0ad937f3ffebf81557fb6567fd4 > which should allow better reporting of dataset errors in histories. > > Make sure export GALAXY_TEST_VERBOSE_ERRORS=True is executed before > your tests and then check the output of the API driven version - it > should have a breakdown of all datasets in error when it encounters a > history that is not ok. This should include the standard error and > standard output from the upload tool - which in turn should tell us > the exact problem (fingers crossed). > > > > On Fri, Apr 18, 2014 at 8:13 AM, Saket Choudhary wrote: >> * fork your repository. >> >> On 18 April 2014 18:42, Saket Choudhary wrote: >>> On 18 April 2014 18:06, Peter Cock wrote: On Fri, Apr 18, 2014 at 11:47 AM, Saket Choudhary wrote: > On 18 April 2014 14:59, Peter Cock wrote: >> Very strange. Adding "ls test-data" to the TravisCI list worked, >> and showed the input file condel_input.tabular was present: >> https://github.com/saketkc/galaxy_tools/commit/c04ff946db0182de9aca4f6dc36ff194404cae05 >> https://s3.amazonaws.com/archive.travis-ci.org/jobs/23239476/log.txt >> https://s3.amazonaws.com/archive.travis-ci.org/jobs/23239477/log.txt >> >> How about replacing that with "ls -l test-data/condel_input.tabular" >> to verify the ownership and permissions (although if that was >> the upload problem, Galaxy could give a clearer error). >> > > Thanks, tried it: > https://travis-ci.org/saketkc/galaxy_tools/jobs/23270454#L530 > I see, that looks fine: $ ls -l test-data/condel_input.tabular -rwxrwxr-x 1 travis travis 110 Apr 18 10:28 test-data/condel_input.tabular > This happens irrespective of using ftype. and the plce where ERROR > occurs is this: > > ... > > so my dataset state is set to ERROR, essentially upload failed. But > why, is what it doesn't tell me. If this were a systematic problem other people should have seen it - but I have tests using tabular input files which still work fine. I wonder where exactly in the process it is failing... perhaps in the metadata processing which would be consistent with it being some oddity in your sample file which is causing trouble? Can you upload this file condel_input.tabular into Galaxy via the UI without error? >>> >>> Yes, the upload works fine with 'Auto-Detect'. It sets the type to be >>> 'interval'. I am also able to upload if I manually set it to >>> 'tabular', which is anyway expected. >>> >>> Just to double check I tested it with 'filter1_test5.tab' that comes >>> with galaxy-central and that is failing too: >>> https://travis-ci.org/saketkc/galaxy_tools/jobs/23277155#L1283 >>> >>> I am going to clone your repository and try pushing in my tools to the >>> clone to see if that helps. >>> >>> Peter >> ___ >> Please keep all replies on the list by using "reply all" >> in your mail client. To manage your subscriptions to this >> and other Galaxy lists, please use the interface at: >> http://lists.bx.psu.edu/ >> >> To search Galaxy mailing lists use the unified search at: >> http://galaxyproject.org/search/mailinglists/ ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions
Re: [galaxy-dev] Functional Tests and ftype
Pushed out this changeset https://bitbucket.org/galaxy/galaxy-central/commits/183668ed911ca0ad937f3ffebf81557fb6567fd4 which should allow better reporting of dataset errors in histories. Make sure export GALAXY_TEST_VERBOSE_ERRORS=True is executed before your tests and then check the output of the API driven version - it should have a breakdown of all datasets in error when it encounters a history that is not ok. This should include the standard error and standard output from the upload tool - which in turn should tell us the exact problem (fingers crossed). On Fri, Apr 18, 2014 at 8:13 AM, Saket Choudhary wrote: > * fork your repository. > > On 18 April 2014 18:42, Saket Choudhary wrote: >> On 18 April 2014 18:06, Peter Cock wrote: >>> On Fri, Apr 18, 2014 at 11:47 AM, Saket Choudhary wrote: On 18 April 2014 14:59, Peter Cock wrote: > Very strange. Adding "ls test-data" to the TravisCI list worked, > and showed the input file condel_input.tabular was present: > https://github.com/saketkc/galaxy_tools/commit/c04ff946db0182de9aca4f6dc36ff194404cae05 > https://s3.amazonaws.com/archive.travis-ci.org/jobs/23239476/log.txt > https://s3.amazonaws.com/archive.travis-ci.org/jobs/23239477/log.txt > > How about replacing that with "ls -l test-data/condel_input.tabular" > to verify the ownership and permissions (although if that was > the upload problem, Galaxy could give a clearer error). > Thanks, tried it: https://travis-ci.org/saketkc/galaxy_tools/jobs/23270454#L530 >>> >>> I see, that looks fine: >>> >>> $ ls -l test-data/condel_input.tabular >>> -rwxrwxr-x 1 travis travis 110 Apr 18 10:28 test-data/condel_input.tabular >>> This happens irrespective of using ftype. and the plce where ERROR occurs is this: ... so my dataset state is set to ERROR, essentially upload failed. But why, is what it doesn't tell me. >>> >>> If this were a systematic problem other people should have seen >>> it - but I have tests using tabular input files which still work fine. >>> >>> I wonder where exactly in the process it is failing... perhaps in the >>> metadata processing which would be consistent with it being some >>> oddity in your sample file which is causing trouble? >>> >>> Can you upload this file condel_input.tabular into Galaxy via the >>> UI without error? >>> >> >> Yes, the upload works fine with 'Auto-Detect'. It sets the type to be >> 'interval'. I am also able to upload if I manually set it to >> 'tabular', which is anyway expected. >> >> Just to double check I tested it with 'filter1_test5.tab' that comes >> with galaxy-central and that is failing too: >> https://travis-ci.org/saketkc/galaxy_tools/jobs/23277155#L1283 >> >> I am going to clone your repository and try pushing in my tools to the >> clone to see if that helps. >> >> >>> Peter > ___ > Please keep all replies on the list by using "reply all" > in your mail client. To manage your subscriptions to this > and other Galaxy lists, please use the interface at: > http://lists.bx.psu.edu/ > > To search Galaxy mailing lists use the unified search at: > http://galaxyproject.org/search/mailinglists/ ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional Tests and ftype
* fork your repository. On 18 April 2014 18:42, Saket Choudhary wrote: > On 18 April 2014 18:06, Peter Cock wrote: >> On Fri, Apr 18, 2014 at 11:47 AM, Saket Choudhary wrote: >>> On 18 April 2014 14:59, Peter Cock wrote: Very strange. Adding "ls test-data" to the TravisCI list worked, and showed the input file condel_input.tabular was present: https://github.com/saketkc/galaxy_tools/commit/c04ff946db0182de9aca4f6dc36ff194404cae05 https://s3.amazonaws.com/archive.travis-ci.org/jobs/23239476/log.txt https://s3.amazonaws.com/archive.travis-ci.org/jobs/23239477/log.txt How about replacing that with "ls -l test-data/condel_input.tabular" to verify the ownership and permissions (although if that was the upload problem, Galaxy could give a clearer error). >>> >>> Thanks, tried it: >>> https://travis-ci.org/saketkc/galaxy_tools/jobs/23270454#L530 >>> >> >> I see, that looks fine: >> >> $ ls -l test-data/condel_input.tabular >> -rwxrwxr-x 1 travis travis 110 Apr 18 10:28 test-data/condel_input.tabular >> >>> This happens irrespective of using ftype. and the plce where ERROR >>> occurs is this: >>> >>> ... >>> >>> so my dataset state is set to ERROR, essentially upload failed. But >>> why, is what it doesn't tell me. >> >> If this were a systematic problem other people should have seen >> it - but I have tests using tabular input files which still work fine. >> >> I wonder where exactly in the process it is failing... perhaps in the >> metadata processing which would be consistent with it being some >> oddity in your sample file which is causing trouble? >> >> Can you upload this file condel_input.tabular into Galaxy via the >> UI without error? >> > > Yes, the upload works fine with 'Auto-Detect'. It sets the type to be > 'interval'. I am also able to upload if I manually set it to > 'tabular', which is anyway expected. > > Just to double check I tested it with 'filter1_test5.tab' that comes > with galaxy-central and that is failing too: > https://travis-ci.org/saketkc/galaxy_tools/jobs/23277155#L1283 > > I am going to clone your repository and try pushing in my tools to the > clone to see if that helps. > > >> Peter ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional Tests and ftype
On 18 April 2014 18:06, Peter Cock wrote: > On Fri, Apr 18, 2014 at 11:47 AM, Saket Choudhary wrote: >> On 18 April 2014 14:59, Peter Cock wrote: >>> Very strange. Adding "ls test-data" to the TravisCI list worked, >>> and showed the input file condel_input.tabular was present: >>> https://github.com/saketkc/galaxy_tools/commit/c04ff946db0182de9aca4f6dc36ff194404cae05 >>> https://s3.amazonaws.com/archive.travis-ci.org/jobs/23239476/log.txt >>> https://s3.amazonaws.com/archive.travis-ci.org/jobs/23239477/log.txt >>> >>> How about replacing that with "ls -l test-data/condel_input.tabular" >>> to verify the ownership and permissions (although if that was >>> the upload problem, Galaxy could give a clearer error). >>> >> >> Thanks, tried it: >> https://travis-ci.org/saketkc/galaxy_tools/jobs/23270454#L530 >> > > I see, that looks fine: > > $ ls -l test-data/condel_input.tabular > -rwxrwxr-x 1 travis travis 110 Apr 18 10:28 test-data/condel_input.tabular > >> This happens irrespective of using ftype. and the plce where ERROR >> occurs is this: >> >> ... >> >> so my dataset state is set to ERROR, essentially upload failed. But >> why, is what it doesn't tell me. > > If this were a systematic problem other people should have seen > it - but I have tests using tabular input files which still work fine. > > I wonder where exactly in the process it is failing... perhaps in the > metadata processing which would be consistent with it being some > oddity in your sample file which is causing trouble? > > Can you upload this file condel_input.tabular into Galaxy via the > UI without error? > Yes, the upload works fine with 'Auto-Detect'. It sets the type to be 'interval'. I am also able to upload if I manually set it to 'tabular', which is anyway expected. Just to double check I tested it with 'filter1_test5.tab' that comes with galaxy-central and that is failing too: https://travis-ci.org/saketkc/galaxy_tools/jobs/23277155#L1283 I am going to clone your repository and try pushing in my tools to the clone to see if that helps. > Peter ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional Tests and ftype
On Fri, Apr 18, 2014 at 11:47 AM, Saket Choudhary wrote: > On 18 April 2014 14:59, Peter Cock wrote: >> Very strange. Adding "ls test-data" to the TravisCI list worked, >> and showed the input file condel_input.tabular was present: >> https://github.com/saketkc/galaxy_tools/commit/c04ff946db0182de9aca4f6dc36ff194404cae05 >> https://s3.amazonaws.com/archive.travis-ci.org/jobs/23239476/log.txt >> https://s3.amazonaws.com/archive.travis-ci.org/jobs/23239477/log.txt >> >> How about replacing that with "ls -l test-data/condel_input.tabular" >> to verify the ownership and permissions (although if that was >> the upload problem, Galaxy could give a clearer error). >> > > Thanks, tried it: > https://travis-ci.org/saketkc/galaxy_tools/jobs/23270454#L530 > I see, that looks fine: $ ls -l test-data/condel_input.tabular -rwxrwxr-x 1 travis travis 110 Apr 18 10:28 test-data/condel_input.tabular > This happens irrespective of using ftype. and the plce where ERROR > occurs is this: > > ... > > so my dataset state is set to ERROR, essentially upload failed. But > why, is what it doesn't tell me. If this were a systematic problem other people should have seen it - but I have tests using tabular input files which still work fine. I wonder where exactly in the process it is failing... perhaps in the metadata processing which would be consistent with it being some oddity in your sample file which is causing trouble? Can you upload this file condel_input.tabular into Galaxy via the UI without error? Peter ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional Tests and ftype
On 18 April 2014 14:59, Peter Cock wrote: > On Thu, Apr 17, 2014 at 10:39 PM, Saket Choudhary wrote: >> John pointed out that the error its the upload of the test file itself >> which is failing. I can locate the input files >> in test-data/ >> >> https://travis-ci.org/saketkc/galaxy_tools/jobs/23238748#L1723 >> >> Any clues as to where exactly am I going wrong? > > Very strange. Adding "ls test-data" to the TravisCI list worked, > and showed the input file condel_input.tabular was present: > https://github.com/saketkc/galaxy_tools/commit/c04ff946db0182de9aca4f6dc36ff194404cae05 > https://s3.amazonaws.com/archive.travis-ci.org/jobs/23239476/log.txt > https://s3.amazonaws.com/archive.travis-ci.org/jobs/23239477/log.txt > > How about replacing that with "ls -l test-data/condel_input.tabular" > to verify the ownership and permissions (although if that was > the upload problem, Galaxy could give a clearer error). > Thanks, tried it: https://travis-ci.org/saketkc/galaxy_tools/jobs/23270454#L530 This happens irrespective of using ftype. and the plce where ERROR occurs is this: galaxy.jobs.runners.local DEBUG 2014-04-17 21:25:53,808 execution finished: GALAXY_SLOTS="1"; export GALAXY_SLOTS; python /home/travis/build/saketkc/galaxy_tools/galaxy-central-master/tools/data_source/upload.py /home/travis/build/saketkc/galaxy_tools/galaxy-central-master /tmp/tmpNenHnc/tmp83XrmN/new_files_path_4s4ndH/tmpvroRLA /tmp/tmpNenHnc/tmp83XrmN/new_files_path_4s4ndH/tmpv9tzOU 1:/tmp/tmpNenHnc/tmp83XrmN/job_working_directory_AsN2Fs/000/1/dataset_1_files:/tmp/tmpNenHnc/tmp83XrmN/database/files/000/dataset_1.dat galaxy.datatypes.metadata INFO 2014-04-17 21:25:55,902 /tmp/tmpNenHnc/tmp83XrmN/job_working_directory_AsN2Fs/000/1/metadata_in_HistoryDatasetAssociation_1_bVl30h,/tmp/tmpNenHnc/tmp83XrmN/job_working_directory_AsN2Fs/000/1/metadata_kwds_HistoryDatasetAssociation_1_jiQBR3,/tmp/tmpNenHnc/tmp83XrmN/job_working_directory_AsN2Fs/000/1/metadata_out_HistoryDatasetAssociation_1_aoG7yJ,/tmp/tmpNenHnc/tmp83XrmN/job_working_directory_AsN2Fs/000/1/metadata_results_HistoryDatasetAssociation_1_JM1wqM,,/tmp/tmpNenHnc/tmp83XrmN/job_working_directory_AsN2Fs/000/1/metadata_override_HistoryDatasetAssociation_1_ZUFWdn galaxy.jobs.runners DEBUG 2014-04-17 21:25:55,903 executing external set_meta script for job 1: /home/travis/build/saketkc/galaxy_tools/galaxy-central-master/set_metadata.sh /tmp/tmpNenHnc/tmp83XrmN/database/files /tmp/tmpNenHnc/tmp83XrmN/job_working_directory_AsN2Fs/000/1 . /tmp/tmpNenHnc/functional_tests_wsgi.ini /tmp/tmpNenHnc/tmp83XrmN/new_files_path_4s4ndH/tmpvroRLA /tmp/tmpNenHnc/tmp83XrmN/job_working_directory_AsN2Fs/000/1/galaxy.json /tmp/tmpNenHnc/tmp83XrmN/job_working_directory_AsN2Fs/000/1/metadata_in_HistoryDatasetAssociation_1_bVl30h,/tmp/tmpNenHnc/tmp83XrmN/job_working_directory_AsN2Fs/000/1/metadata_kwds_HistoryDatasetAssociation_1_jiQBR3,/tmp/tmpNenHnc/tmp83XrmN/job_working_directory_AsN2Fs/000/1/metadata_out_HistoryDatasetAssociation_1_aoG7yJ,/tmp/tmpNenHnc/tmp83XrmN/job_working_directory_AsN2Fs/000/1/metadata_results_HistoryDatasetAssociation_1_JM1wqM,,/tmp/tmpNenHnc/tmp83XrmN/job_working_directory_AsN2Fs/000/1/metadata_override_HistoryDatasetAssociation_1_ZUFWdn galaxy.jobs.runners DEBUG 2014-04-17 21:25:57,620 execution of external set_meta for job 1 finished galaxy.jobs DEBUG 2014-04-17 21:25:57,727 setting dataset state to ERROR so my dataset state is set to ERROR, essentially upload failed. But why, is what it doesn't tell me. > Likely unrelated, but I don't think I have bothered to set the ftype > when uploading tabular files / using them as test inputs. Did you > ever try this without the ftype? > > Peter ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional Tests and ftype
On Thu, Apr 17, 2014 at 10:39 PM, Saket Choudhary wrote: > John pointed out that the error its the upload of the test file itself > which is failing. I can locate the input files > in test-data/ > > https://travis-ci.org/saketkc/galaxy_tools/jobs/23238748#L1723 > > Any clues as to where exactly am I going wrong? Very strange. Adding "ls test-data" to the TravisCI list worked, and showed the input file condel_input.tabular was present: https://github.com/saketkc/galaxy_tools/commit/c04ff946db0182de9aca4f6dc36ff194404cae05 https://s3.amazonaws.com/archive.travis-ci.org/jobs/23239476/log.txt https://s3.amazonaws.com/archive.travis-ci.org/jobs/23239477/log.txt How about replacing that with "ls -l test-data/condel_input.tabular" to verify the ownership and permissions (although if that was the upload problem, Galaxy could give a clearer error). Likely unrelated, but I don't think I have bothered to set the ftype when uploading tabular files / using them as test inputs. Did you ever try this without the ftype? Peter ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional Tests and ftype
Can I throw in a bounty for this one? ;-) On 18 April 2014 03:09, Saket Choudhary wrote: > John pointed out that the error its the upload of the test file itself > which is failing. I can locate the input files > in test-data/ > > https://travis-ci.org/saketkc/galaxy_tools/jobs/23238748#L1723 > > Any clues as to where exactly am I going wrong? > > On 17 April 2014 19:40, Saket Choudhary wrote: >> Apparently my tool itself does a web API call, so the test is bound to >> fail locally. My best bet is to rely on travis for now, however I >> still see this failing: >> https://travis-ci.org/saketkc/galaxy_tools/jobs/23206181#L1721 >> >> The input file exists: >> https://travis-ci.org/saketkc/galaxy_tools/jobs/23206181#L1721 >> >> On testtoolshed: http://testtoolshed.g2.bx.psu.edu/view/saketkc/condel >> >> Installation fails, because pycurl fails to install which I shall >> report in other thread. >> >> Saket >> >> On 17 April 2014 18:45, Saket Choudhary wrote: >>> Culprit: System Proxy. >>> >>> I am behind a proxy and have my http_proxy, https_proxy variables set. >>> The trick is to unset them. >>> Because I am relying on John's template[1] for testing, I had to just >>> unset http_proxy ONLY. >>> >>> Maybe I should add this to the wiki? >>> >>> [1] >>> https://github.com/jmchilton/galaxy-downloads/raw/master/db_gx_rev_0117.sqlite >>> >>> On 17 April 2014 18:18, Ross wrote: Hi, Saket - 436 tests in 19.153 seconds means something odd (or fast) with your setup? What errors do you see? Ross Lazarus MBBS MPH; Head, Medical Bioinformatics, BakerIDI; Tel: +61 385321444 http://scholar.google.com/citations?hl=en&user=UCUuEM4J On Thu, Apr 17, 2014 at 10:43 PM, Saket Choudhary wrote: > > Just to add, that is the case with all the tests: > > $ sh run_functional_tests.sh > Ran 436 tests in 19.153s > > FAILED (errors=434) > > All tests fail locally and hence I decided to rely on travis. > https://wiki.galaxyproject.org/Admin/RunningTests#Functional_Tests > says it might be due to conflicting version of paste in my PYTHONPATH. > But it happens even if I run these under a virtualenv. > > Saket > > On 17 April 2014 18:03, Saket Choudhary wrote: > > On 17 April 2014 15:34, Peter Cock wrote: > >> On Thu, Apr 17, 2014 at 12:59 AM, Saket Choudhary > >> wrote: > >>> I happened to copy Peter's .travis.yml for running tests for some of > >>> my packages. > >> > >> Background information: > >> > >> http://blastedbio.blogspot.co.uk/2013/09/using-travis-ci-for-testing-galaxy-tools.html > >> > >>> However the process has not been smooth at all. > >>> See : https://travis-ci.org/saketkc/galaxy_tools/jobs/23166678#L908 > >>> > >>> AssertionError: Attempting to set field 'input' to value > >>> '['condel_input.tabular']' in form 'tool_form' threw exception: cannot > >>> find value/label "condel_input.tabular" in list control > >>> > >>> I am already passing it the correct ftype : > >>> > >>> https://github.com/saketkc/galaxy_tools/blob/travis-test/condel_web/condel_web.xml#L22 > >>> > >>> Is there something I am missing? > >> > >> This is similar to (but different) to the error message I recently > >> reported as being unclear when a datatype was not defined: > >> http://lists.bx.psu.edu/pipermail/galaxy-dev/2014-April/019090.html > >> > >> In your case, this is an input file which seems to be missing, > >> test-data/condel_input.tabular - but I see the line copying the > >> tool's test-data into the main Galaxy test-data folder is there > >> in the .travis.yml file. > >> > >> I am puzzled too. > >> > >> The key question I would ask you is: do your tests work locally? > >> If so, are they installed via a (local) Tool Shed, or manually along > >> the same lines of the attempted TravisCI tool installation? > >> > > > > Unfortunately No. > > > > TwillAssertionError: code is 500 != 200 > > > > I am running Galaxy under virtualenv. Unfortunately I get the same > > error irrespective of whther my input/output files exits inside > > test-data/ directory. I forgot to link them to test-data path earlier, > > but I get the same error even otherwise. > > > > > >> Regards, > >> > >> Peter > ___ > Please keep all replies on the list by using "reply all" > in your mail client. To manage your subscriptions to this > and other Galaxy lists, please use the interface at: > http://lists.bx.psu.edu/ > > To search Galaxy mailing lists use the unified search at: > http://galaxyproject.org/search/mailinglists/ ___ Please keep all replies on the list by using "reply all"
Re: [galaxy-dev] Functional Tests and ftype
John pointed out that the error its the upload of the test file itself which is failing. I can locate the input files in test-data/ https://travis-ci.org/saketkc/galaxy_tools/jobs/23238748#L1723 Any clues as to where exactly am I going wrong? On 17 April 2014 19:40, Saket Choudhary wrote: > Apparently my tool itself does a web API call, so the test is bound to > fail locally. My best bet is to rely on travis for now, however I > still see this failing: > https://travis-ci.org/saketkc/galaxy_tools/jobs/23206181#L1721 > > The input file exists: > https://travis-ci.org/saketkc/galaxy_tools/jobs/23206181#L1721 > > On testtoolshed: http://testtoolshed.g2.bx.psu.edu/view/saketkc/condel > > Installation fails, because pycurl fails to install which I shall > report in other thread. > > Saket > > On 17 April 2014 18:45, Saket Choudhary wrote: >> Culprit: System Proxy. >> >> I am behind a proxy and have my http_proxy, https_proxy variables set. >> The trick is to unset them. >> Because I am relying on John's template[1] for testing, I had to just >> unset http_proxy ONLY. >> >> Maybe I should add this to the wiki? >> >> [1] >> https://github.com/jmchilton/galaxy-downloads/raw/master/db_gx_rev_0117.sqlite >> >> On 17 April 2014 18:18, Ross wrote: >>> Hi, Saket - 436 tests in 19.153 seconds means something odd (or fast) with >>> your setup? >>> What errors do you see? >>> >>> Ross Lazarus MBBS MPH; >>> Head, Medical Bioinformatics, BakerIDI; Tel: +61 385321444 >>> http://scholar.google.com/citations?hl=en&user=UCUuEM4J >>> >>> >>> On Thu, Apr 17, 2014 at 10:43 PM, Saket Choudhary wrote: Just to add, that is the case with all the tests: $ sh run_functional_tests.sh Ran 436 tests in 19.153s FAILED (errors=434) All tests fail locally and hence I decided to rely on travis. https://wiki.galaxyproject.org/Admin/RunningTests#Functional_Tests says it might be due to conflicting version of paste in my PYTHONPATH. But it happens even if I run these under a virtualenv. Saket On 17 April 2014 18:03, Saket Choudhary wrote: > On 17 April 2014 15:34, Peter Cock wrote: >> On Thu, Apr 17, 2014 at 12:59 AM, Saket Choudhary >> wrote: >>> I happened to copy Peter's .travis.yml for running tests for some of >>> my packages. >> >> Background information: >> >> http://blastedbio.blogspot.co.uk/2013/09/using-travis-ci-for-testing-galaxy-tools.html >> >>> However the process has not been smooth at all. >>> See : https://travis-ci.org/saketkc/galaxy_tools/jobs/23166678#L908 >>> >>> AssertionError: Attempting to set field 'input' to value >>> '['condel_input.tabular']' in form 'tool_form' threw exception: cannot >>> find value/label "condel_input.tabular" in list control >>> >>> I am already passing it the correct ftype : >>> >>> https://github.com/saketkc/galaxy_tools/blob/travis-test/condel_web/condel_web.xml#L22 >>> >>> Is there something I am missing? >> >> This is similar to (but different) to the error message I recently >> reported as being unclear when a datatype was not defined: >> http://lists.bx.psu.edu/pipermail/galaxy-dev/2014-April/019090.html >> >> In your case, this is an input file which seems to be missing, >> test-data/condel_input.tabular - but I see the line copying the >> tool's test-data into the main Galaxy test-data folder is there >> in the .travis.yml file. >> >> I am puzzled too. >> >> The key question I would ask you is: do your tests work locally? >> If so, are they installed via a (local) Tool Shed, or manually along >> the same lines of the attempted TravisCI tool installation? >> > > Unfortunately No. > > TwillAssertionError: code is 500 != 200 > > I am running Galaxy under virtualenv. Unfortunately I get the same > error irrespective of whther my input/output files exits inside > test-data/ directory. I forgot to link them to test-data path earlier, > but I get the same error even otherwise. > > >> Regards, >> >> Peter ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/ >>> >>> ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/sea
Re: [galaxy-dev] Functional Tests and ftype
Apparently my tool itself does a web API call, so the test is bound to fail locally. My best bet is to rely on travis for now, however I still see this failing: https://travis-ci.org/saketkc/galaxy_tools/jobs/23206181#L1721 The input file exists: https://travis-ci.org/saketkc/galaxy_tools/jobs/23206181#L1721 On testtoolshed: http://testtoolshed.g2.bx.psu.edu/view/saketkc/condel Installation fails, because pycurl fails to install which I shall report in other thread. Saket On 17 April 2014 18:45, Saket Choudhary wrote: > Culprit: System Proxy. > > I am behind a proxy and have my http_proxy, https_proxy variables set. > The trick is to unset them. > Because I am relying on John's template[1] for testing, I had to just > unset http_proxy ONLY. > > Maybe I should add this to the wiki? > > [1] > https://github.com/jmchilton/galaxy-downloads/raw/master/db_gx_rev_0117.sqlite > > On 17 April 2014 18:18, Ross wrote: >> Hi, Saket - 436 tests in 19.153 seconds means something odd (or fast) with >> your setup? >> What errors do you see? >> >> Ross Lazarus MBBS MPH; >> Head, Medical Bioinformatics, BakerIDI; Tel: +61 385321444 >> http://scholar.google.com/citations?hl=en&user=UCUuEM4J >> >> >> On Thu, Apr 17, 2014 at 10:43 PM, Saket Choudhary wrote: >>> >>> Just to add, that is the case with all the tests: >>> >>> $ sh run_functional_tests.sh >>> Ran 436 tests in 19.153s >>> >>> FAILED (errors=434) >>> >>> All tests fail locally and hence I decided to rely on travis. >>> https://wiki.galaxyproject.org/Admin/RunningTests#Functional_Tests >>> says it might be due to conflicting version of paste in my PYTHONPATH. >>> But it happens even if I run these under a virtualenv. >>> >>> Saket >>> >>> On 17 April 2014 18:03, Saket Choudhary wrote: >>> > On 17 April 2014 15:34, Peter Cock wrote: >>> >> On Thu, Apr 17, 2014 at 12:59 AM, Saket Choudhary >>> >> wrote: >>> >>> I happened to copy Peter's .travis.yml for running tests for some of >>> >>> my packages. >>> >> >>> >> Background information: >>> >> >>> >> http://blastedbio.blogspot.co.uk/2013/09/using-travis-ci-for-testing-galaxy-tools.html >>> >> >>> >>> However the process has not been smooth at all. >>> >>> See : https://travis-ci.org/saketkc/galaxy_tools/jobs/23166678#L908 >>> >>> >>> >>> AssertionError: Attempting to set field 'input' to value >>> >>> '['condel_input.tabular']' in form 'tool_form' threw exception: cannot >>> >>> find value/label "condel_input.tabular" in list control >>> >>> >>> >>> I am already passing it the correct ftype : >>> >>> >>> >>> https://github.com/saketkc/galaxy_tools/blob/travis-test/condel_web/condel_web.xml#L22 >>> >>> >>> >>> Is there something I am missing? >>> >> >>> >> This is similar to (but different) to the error message I recently >>> >> reported as being unclear when a datatype was not defined: >>> >> http://lists.bx.psu.edu/pipermail/galaxy-dev/2014-April/019090.html >>> >> >>> >> In your case, this is an input file which seems to be missing, >>> >> test-data/condel_input.tabular - but I see the line copying the >>> >> tool's test-data into the main Galaxy test-data folder is there >>> >> in the .travis.yml file. >>> >> >>> >> I am puzzled too. >>> >> >>> >> The key question I would ask you is: do your tests work locally? >>> >> If so, are they installed via a (local) Tool Shed, or manually along >>> >> the same lines of the attempted TravisCI tool installation? >>> >> >>> > >>> > Unfortunately No. >>> > >>> > TwillAssertionError: code is 500 != 200 >>> > >>> > I am running Galaxy under virtualenv. Unfortunately I get the same >>> > error irrespective of whther my input/output files exits inside >>> > test-data/ directory. I forgot to link them to test-data path earlier, >>> > but I get the same error even otherwise. >>> > >>> > >>> >> Regards, >>> >> >>> >> Peter >>> ___ >>> Please keep all replies on the list by using "reply all" >>> in your mail client. To manage your subscriptions to this >>> and other Galaxy lists, please use the interface at: >>> http://lists.bx.psu.edu/ >>> >>> To search Galaxy mailing lists use the unified search at: >>> http://galaxyproject.org/search/mailinglists/ >> >> ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional Tests and ftype
HI John, Saket, Looking at Skaet's .travis.yml he doesn't mess about with the default datatypes_conf.xml(.sample) like I do, so his TravisCI should have all the default Galaxy datatypes (including tabular). In my more complex .travis.yml I modify this to two reasons, first to add new datatypes, and second for a slight speedup I remove most of the unused datatypes. Peter On Thu, Apr 17, 2014 at 1:45 PM, Saket Choudhary wrote: > Hi John, > > My tool's input datatype is tabular. I do not have any custom datatypes > defined. > > Saket > > On 17 April 2014 18:10, John Chilton wrote: >> I don't see anywhere in your Travis YAML file where you are modifying >> Galaxy's datatypes (like Peter does for instance on these lines >> https://github.com/peterjc/pico_galaxy/blob/master/.travis.yml#L166). >> If those datatypes are not defined you cannot upload a tool with a >> specific type. Am I missing something? >> >> -John ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional Tests and ftype
Culprit: System Proxy. I am behind a proxy and have my http_proxy, https_proxy variables set. The trick is to unset them. Because I am relying on John's template[1] for testing, I had to just unset http_proxy ONLY. Maybe I should add this to the wiki? [1] https://github.com/jmchilton/galaxy-downloads/raw/master/db_gx_rev_0117.sqlite On 17 April 2014 18:18, Ross wrote: > Hi, Saket - 436 tests in 19.153 seconds means something odd (or fast) with > your setup? > What errors do you see? > > Ross Lazarus MBBS MPH; > Head, Medical Bioinformatics, BakerIDI; Tel: +61 385321444 > http://scholar.google.com/citations?hl=en&user=UCUuEM4J > > > On Thu, Apr 17, 2014 at 10:43 PM, Saket Choudhary wrote: >> >> Just to add, that is the case with all the tests: >> >> $ sh run_functional_tests.sh >> Ran 436 tests in 19.153s >> >> FAILED (errors=434) >> >> All tests fail locally and hence I decided to rely on travis. >> https://wiki.galaxyproject.org/Admin/RunningTests#Functional_Tests >> says it might be due to conflicting version of paste in my PYTHONPATH. >> But it happens even if I run these under a virtualenv. >> >> Saket >> >> On 17 April 2014 18:03, Saket Choudhary wrote: >> > On 17 April 2014 15:34, Peter Cock wrote: >> >> On Thu, Apr 17, 2014 at 12:59 AM, Saket Choudhary >> >> wrote: >> >>> I happened to copy Peter's .travis.yml for running tests for some of >> >>> my packages. >> >> >> >> Background information: >> >> >> >> http://blastedbio.blogspot.co.uk/2013/09/using-travis-ci-for-testing-galaxy-tools.html >> >> >> >>> However the process has not been smooth at all. >> >>> See : https://travis-ci.org/saketkc/galaxy_tools/jobs/23166678#L908 >> >>> >> >>> AssertionError: Attempting to set field 'input' to value >> >>> '['condel_input.tabular']' in form 'tool_form' threw exception: cannot >> >>> find value/label "condel_input.tabular" in list control >> >>> >> >>> I am already passing it the correct ftype : >> >>> >> >>> https://github.com/saketkc/galaxy_tools/blob/travis-test/condel_web/condel_web.xml#L22 >> >>> >> >>> Is there something I am missing? >> >> >> >> This is similar to (but different) to the error message I recently >> >> reported as being unclear when a datatype was not defined: >> >> http://lists.bx.psu.edu/pipermail/galaxy-dev/2014-April/019090.html >> >> >> >> In your case, this is an input file which seems to be missing, >> >> test-data/condel_input.tabular - but I see the line copying the >> >> tool's test-data into the main Galaxy test-data folder is there >> >> in the .travis.yml file. >> >> >> >> I am puzzled too. >> >> >> >> The key question I would ask you is: do your tests work locally? >> >> If so, are they installed via a (local) Tool Shed, or manually along >> >> the same lines of the attempted TravisCI tool installation? >> >> >> > >> > Unfortunately No. >> > >> > TwillAssertionError: code is 500 != 200 >> > >> > I am running Galaxy under virtualenv. Unfortunately I get the same >> > error irrespective of whther my input/output files exits inside >> > test-data/ directory. I forgot to link them to test-data path earlier, >> > but I get the same error even otherwise. >> > >> > >> >> Regards, >> >> >> >> Peter >> ___ >> Please keep all replies on the list by using "reply all" >> in your mail client. To manage your subscriptions to this >> and other Galaxy lists, please use the interface at: >> http://lists.bx.psu.edu/ >> >> To search Galaxy mailing lists use the unified search at: >> http://galaxyproject.org/search/mailinglists/ > > ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional Tests and ftype
Hi, Saket - 436 tests in 19.153 seconds means something odd (or fast) with your setup? What errors do you see? Ross Lazarus MBBS MPH; Head, Medical Bioinformatics, BakerIDI; Tel: +61 385321444 http://scholar.google.com/citations?hl=en&user=UCUuEM4J On Thu, Apr 17, 2014 at 10:43 PM, Saket Choudhary wrote: > Just to add, that is the case with all the tests: > > $ sh run_functional_tests.sh > Ran 436 tests in 19.153s > > FAILED (errors=434) > > All tests fail locally and hence I decided to rely on travis. > https://wiki.galaxyproject.org/Admin/RunningTests#Functional_Tests > says it might be due to conflicting version of paste in my PYTHONPATH. > But it happens even if I run these under a virtualenv. > > Saket > > On 17 April 2014 18:03, Saket Choudhary wrote: > > On 17 April 2014 15:34, Peter Cock wrote: > >> On Thu, Apr 17, 2014 at 12:59 AM, Saket Choudhary > wrote: > >>> I happened to copy Peter's .travis.yml for running tests for some of > >>> my packages. > >> > >> Background information: > >> > http://blastedbio.blogspot.co.uk/2013/09/using-travis-ci-for-testing-galaxy-tools.html > >> > >>> However the process has not been smooth at all. > >>> See : https://travis-ci.org/saketkc/galaxy_tools/jobs/23166678#L908 > >>> > >>> AssertionError: Attempting to set field 'input' to value > >>> '['condel_input.tabular']' in form 'tool_form' threw exception: cannot > >>> find value/label "condel_input.tabular" in list control > >>> > >>> I am already passing it the correct ftype : > >>> > https://github.com/saketkc/galaxy_tools/blob/travis-test/condel_web/condel_web.xml#L22 > >>> > >>> Is there something I am missing? > >> > >> This is similar to (but different) to the error message I recently > >> reported as being unclear when a datatype was not defined: > >> http://lists.bx.psu.edu/pipermail/galaxy-dev/2014-April/019090.html > >> > >> In your case, this is an input file which seems to be missing, > >> test-data/condel_input.tabular - but I see the line copying the > >> tool's test-data into the main Galaxy test-data folder is there > >> in the .travis.yml file. > >> > >> I am puzzled too. > >> > >> The key question I would ask you is: do your tests work locally? > >> If so, are they installed via a (local) Tool Shed, or manually along > >> the same lines of the attempted TravisCI tool installation? > >> > > > > Unfortunately No. > > > > TwillAssertionError: code is 500 != 200 > > > > I am running Galaxy under virtualenv. Unfortunately I get the same > > error irrespective of whther my input/output files exits inside > > test-data/ directory. I forgot to link them to test-data path earlier, > > but I get the same error even otherwise. > > > > > >> Regards, > >> > >> Peter > ___ > Please keep all replies on the list by using "reply all" > in your mail client. To manage your subscriptions to this > and other Galaxy lists, please use the interface at: > http://lists.bx.psu.edu/ > > To search Galaxy mailing lists use the unified search at: > http://galaxyproject.org/search/mailinglists/ > ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional Tests and ftype
Oppps, duh sorry I thought condel_input.tabular was a datatype not a file. That makes sense. I'll look some more. -John On Thu, Apr 17, 2014 at 7:45 AM, Saket Choudhary wrote: > Hi John, > > My tool's input datatype is tabular. I do not have any custom datatypes > defined. > > Saket > > On 17 April 2014 18:10, John Chilton wrote: >> I don't see anywhere in your Travis YAML file where you are modifying >> Galaxy's datatypes (like Peter does for instance on these lines >> https://github.com/peterjc/pico_galaxy/blob/master/.travis.yml#L166). >> If those datatypes are not defined you cannot upload a tool with a >> specific type. Am I missing something? >> >> -John >> >> On Thu, Apr 17, 2014 at 7:33 AM, Saket Choudhary wrote: >>> On 17 April 2014 15:34, Peter Cock wrote: On Thu, Apr 17, 2014 at 12:59 AM, Saket Choudhary wrote: > I happened to copy Peter's .travis.yml for running tests for some of > my packages. Background information: http://blastedbio.blogspot.co.uk/2013/09/using-travis-ci-for-testing-galaxy-tools.html > However the process has not been smooth at all. > See : https://travis-ci.org/saketkc/galaxy_tools/jobs/23166678#L908 > > AssertionError: Attempting to set field 'input' to value > '['condel_input.tabular']' in form 'tool_form' threw exception: cannot > find value/label "condel_input.tabular" in list control > > I am already passing it the correct ftype : > https://github.com/saketkc/galaxy_tools/blob/travis-test/condel_web/condel_web.xml#L22 > > Is there something I am missing? This is similar to (but different) to the error message I recently reported as being unclear when a datatype was not defined: http://lists.bx.psu.edu/pipermail/galaxy-dev/2014-April/019090.html In your case, this is an input file which seems to be missing, test-data/condel_input.tabular - but I see the line copying the tool's test-data into the main Galaxy test-data folder is there in the .travis.yml file. I am puzzled too. The key question I would ask you is: do your tests work locally? If so, are they installed via a (local) Tool Shed, or manually along the same lines of the attempted TravisCI tool installation? >>> >>> Unfortunately No. >>> >>> TwillAssertionError: code is 500 != 200 >>> >>> I am running Galaxy under virtualenv. Unfortunately I get the same >>> error irrespective of whther my input/output files exits inside >>> test-data/ directory. I forgot to link them to test-data path earlier, >>> but I get the same error even otherwise. >>> >>> Regards, Peter >>> ___ >>> Please keep all replies on the list by using "reply all" >>> in your mail client. To manage your subscriptions to this >>> and other Galaxy lists, please use the interface at: >>> http://lists.bx.psu.edu/ >>> >>> To search Galaxy mailing lists use the unified search at: >>> http://galaxyproject.org/search/mailinglists/ ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional Tests and ftype
Hi John, My tool's input datatype is tabular. I do not have any custom datatypes defined. Saket On 17 April 2014 18:10, John Chilton wrote: > I don't see anywhere in your Travis YAML file where you are modifying > Galaxy's datatypes (like Peter does for instance on these lines > https://github.com/peterjc/pico_galaxy/blob/master/.travis.yml#L166). > If those datatypes are not defined you cannot upload a tool with a > specific type. Am I missing something? > > -John > > On Thu, Apr 17, 2014 at 7:33 AM, Saket Choudhary wrote: >> On 17 April 2014 15:34, Peter Cock wrote: >>> On Thu, Apr 17, 2014 at 12:59 AM, Saket Choudhary wrote: I happened to copy Peter's .travis.yml for running tests for some of my packages. >>> >>> Background information: >>> http://blastedbio.blogspot.co.uk/2013/09/using-travis-ci-for-testing-galaxy-tools.html >>> However the process has not been smooth at all. See : https://travis-ci.org/saketkc/galaxy_tools/jobs/23166678#L908 AssertionError: Attempting to set field 'input' to value '['condel_input.tabular']' in form 'tool_form' threw exception: cannot find value/label "condel_input.tabular" in list control I am already passing it the correct ftype : https://github.com/saketkc/galaxy_tools/blob/travis-test/condel_web/condel_web.xml#L22 Is there something I am missing? >>> >>> This is similar to (but different) to the error message I recently >>> reported as being unclear when a datatype was not defined: >>> http://lists.bx.psu.edu/pipermail/galaxy-dev/2014-April/019090.html >>> >>> In your case, this is an input file which seems to be missing, >>> test-data/condel_input.tabular - but I see the line copying the >>> tool's test-data into the main Galaxy test-data folder is there >>> in the .travis.yml file. >>> >>> I am puzzled too. >>> >>> The key question I would ask you is: do your tests work locally? >>> If so, are they installed via a (local) Tool Shed, or manually along >>> the same lines of the attempted TravisCI tool installation? >>> >> >> Unfortunately No. >> >> TwillAssertionError: code is 500 != 200 >> >> I am running Galaxy under virtualenv. Unfortunately I get the same >> error irrespective of whther my input/output files exits inside >> test-data/ directory. I forgot to link them to test-data path earlier, >> but I get the same error even otherwise. >> >> >>> Regards, >>> >>> Peter >> ___ >> Please keep all replies on the list by using "reply all" >> in your mail client. To manage your subscriptions to this >> and other Galaxy lists, please use the interface at: >> http://lists.bx.psu.edu/ >> >> To search Galaxy mailing lists use the unified search at: >> http://galaxyproject.org/search/mailinglists/ ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional Tests and ftype
Just to add, that is the case with all the tests: $ sh run_functional_tests.sh Ran 436 tests in 19.153s FAILED (errors=434) All tests fail locally and hence I decided to rely on travis. https://wiki.galaxyproject.org/Admin/RunningTests#Functional_Tests says it might be due to conflicting version of paste in my PYTHONPATH. But it happens even if I run these under a virtualenv. Saket On 17 April 2014 18:03, Saket Choudhary wrote: > On 17 April 2014 15:34, Peter Cock wrote: >> On Thu, Apr 17, 2014 at 12:59 AM, Saket Choudhary wrote: >>> I happened to copy Peter's .travis.yml for running tests for some of >>> my packages. >> >> Background information: >> http://blastedbio.blogspot.co.uk/2013/09/using-travis-ci-for-testing-galaxy-tools.html >> >>> However the process has not been smooth at all. >>> See : https://travis-ci.org/saketkc/galaxy_tools/jobs/23166678#L908 >>> >>> AssertionError: Attempting to set field 'input' to value >>> '['condel_input.tabular']' in form 'tool_form' threw exception: cannot >>> find value/label "condel_input.tabular" in list control >>> >>> I am already passing it the correct ftype : >>> https://github.com/saketkc/galaxy_tools/blob/travis-test/condel_web/condel_web.xml#L22 >>> >>> Is there something I am missing? >> >> This is similar to (but different) to the error message I recently >> reported as being unclear when a datatype was not defined: >> http://lists.bx.psu.edu/pipermail/galaxy-dev/2014-April/019090.html >> >> In your case, this is an input file which seems to be missing, >> test-data/condel_input.tabular - but I see the line copying the >> tool's test-data into the main Galaxy test-data folder is there >> in the .travis.yml file. >> >> I am puzzled too. >> >> The key question I would ask you is: do your tests work locally? >> If so, are they installed via a (local) Tool Shed, or manually along >> the same lines of the attempted TravisCI tool installation? >> > > Unfortunately No. > > TwillAssertionError: code is 500 != 200 > > I am running Galaxy under virtualenv. Unfortunately I get the same > error irrespective of whther my input/output files exits inside > test-data/ directory. I forgot to link them to test-data path earlier, > but I get the same error even otherwise. > > >> Regards, >> >> Peter ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional Tests and ftype
I don't see anywhere in your Travis YAML file where you are modifying Galaxy's datatypes (like Peter does for instance on these lines https://github.com/peterjc/pico_galaxy/blob/master/.travis.yml#L166). If those datatypes are not defined you cannot upload a tool with a specific type. Am I missing something? -John On Thu, Apr 17, 2014 at 7:33 AM, Saket Choudhary wrote: > On 17 April 2014 15:34, Peter Cock wrote: >> On Thu, Apr 17, 2014 at 12:59 AM, Saket Choudhary wrote: >>> I happened to copy Peter's .travis.yml for running tests for some of >>> my packages. >> >> Background information: >> http://blastedbio.blogspot.co.uk/2013/09/using-travis-ci-for-testing-galaxy-tools.html >> >>> However the process has not been smooth at all. >>> See : https://travis-ci.org/saketkc/galaxy_tools/jobs/23166678#L908 >>> >>> AssertionError: Attempting to set field 'input' to value >>> '['condel_input.tabular']' in form 'tool_form' threw exception: cannot >>> find value/label "condel_input.tabular" in list control >>> >>> I am already passing it the correct ftype : >>> https://github.com/saketkc/galaxy_tools/blob/travis-test/condel_web/condel_web.xml#L22 >>> >>> Is there something I am missing? >> >> This is similar to (but different) to the error message I recently >> reported as being unclear when a datatype was not defined: >> http://lists.bx.psu.edu/pipermail/galaxy-dev/2014-April/019090.html >> >> In your case, this is an input file which seems to be missing, >> test-data/condel_input.tabular - but I see the line copying the >> tool's test-data into the main Galaxy test-data folder is there >> in the .travis.yml file. >> >> I am puzzled too. >> >> The key question I would ask you is: do your tests work locally? >> If so, are they installed via a (local) Tool Shed, or manually along >> the same lines of the attempted TravisCI tool installation? >> > > Unfortunately No. > > TwillAssertionError: code is 500 != 200 > > I am running Galaxy under virtualenv. Unfortunately I get the same > error irrespective of whther my input/output files exits inside > test-data/ directory. I forgot to link them to test-data path earlier, > but I get the same error even otherwise. > > >> Regards, >> >> Peter > ___ > Please keep all replies on the list by using "reply all" > in your mail client. To manage your subscriptions to this > and other Galaxy lists, please use the interface at: > http://lists.bx.psu.edu/ > > To search Galaxy mailing lists use the unified search at: > http://galaxyproject.org/search/mailinglists/ ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional Tests and ftype
On 17 April 2014 15:34, Peter Cock wrote: > On Thu, Apr 17, 2014 at 12:59 AM, Saket Choudhary wrote: >> I happened to copy Peter's .travis.yml for running tests for some of >> my packages. > > Background information: > http://blastedbio.blogspot.co.uk/2013/09/using-travis-ci-for-testing-galaxy-tools.html > >> However the process has not been smooth at all. >> See : https://travis-ci.org/saketkc/galaxy_tools/jobs/23166678#L908 >> >> AssertionError: Attempting to set field 'input' to value >> '['condel_input.tabular']' in form 'tool_form' threw exception: cannot >> find value/label "condel_input.tabular" in list control >> >> I am already passing it the correct ftype : >> https://github.com/saketkc/galaxy_tools/blob/travis-test/condel_web/condel_web.xml#L22 >> >> Is there something I am missing? > > This is similar to (but different) to the error message I recently > reported as being unclear when a datatype was not defined: > http://lists.bx.psu.edu/pipermail/galaxy-dev/2014-April/019090.html > > In your case, this is an input file which seems to be missing, > test-data/condel_input.tabular - but I see the line copying the > tool's test-data into the main Galaxy test-data folder is there > in the .travis.yml file. > > I am puzzled too. > > The key question I would ask you is: do your tests work locally? > If so, are they installed via a (local) Tool Shed, or manually along > the same lines of the attempted TravisCI tool installation? > Unfortunately No. TwillAssertionError: code is 500 != 200 I am running Galaxy under virtualenv. Unfortunately I get the same error irrespective of whther my input/output files exits inside test-data/ directory. I forgot to link them to test-data path earlier, but I get the same error even otherwise. > Regards, > > Peter ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional Tests and ftype
On Thu, Apr 17, 2014 at 12:59 AM, Saket Choudhary wrote: > I happened to copy Peter's .travis.yml for running tests for some of > my packages. Background information: http://blastedbio.blogspot.co.uk/2013/09/using-travis-ci-for-testing-galaxy-tools.html > However the process has not been smooth at all. > See : https://travis-ci.org/saketkc/galaxy_tools/jobs/23166678#L908 > > AssertionError: Attempting to set field 'input' to value > '['condel_input.tabular']' in form 'tool_form' threw exception: cannot > find value/label "condel_input.tabular" in list control > > I am already passing it the correct ftype : > https://github.com/saketkc/galaxy_tools/blob/travis-test/condel_web/condel_web.xml#L22 > > Is there something I am missing? This is similar to (but different) to the error message I recently reported as being unclear when a datatype was not defined: http://lists.bx.psu.edu/pipermail/galaxy-dev/2014-April/019090.html In your case, this is an input file which seems to be missing, test-data/condel_input.tabular - but I see the line copying the tool's test-data into the main Galaxy test-data folder is there in the .travis.yml file. I am puzzled too. The key question I would ask you is: do your tests work locally? If so, are they installed via a (local) Tool Shed, or manually along the same lines of the attempted TravisCI tool installation? Regards, Peter ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional tests results in Invalid HTTP return code 404/500, allowed codes: 200
Hello,
I still haven't figured out why Galaxy's Twill is returning HTTP code
404 when running the functional tests script on my personal Tool Shed. I
just did another test to make sure that it is indeed returning the wrong
HTTP code.
$ python
Python 2.7.3 (default, Sep 26 2013, 20:08:41)
[GCC 4.6.3] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import twill.commands as tc
>>> tc.go('http://toolshed.galaxy.domain.org/')
==> at http://shed.galaxy.byobu.info/
>>> print tc.browser.get_code()
200
As you can see, now Twill returns HTTP code 200 as expected. This test
was done with python-twill (0.9-3) on Linux. Galaxy also comes with
Twill 0.9.
To test this further, I made some changes to the Python script
responsible for testing the HTTP return code. Here is the diff:
diff -r 6822f41bc9bb
test/install_and_test_tool_shed_repositories/base/twilltestcase.py
---
a/test/install_and_test_tool_shed_repositories/base/twilltestcase.py
Mon Aug 19 13:06:17 2013 -0400
+++
b/test/install_and_test_tool_shed_repositories/base/twilltestcase.py
Sat Oct 19 14:53:26 2013 +0200
@@ -67,7 +67,7 @@
# install_repositories_by_revision redirect, so we have to
include 403 in the allowed HTTP
# status codes and log in again.
url = '%s/repository/install_repositories_by_revision?%s' % (
tool_shed_url, install_params )
-self.visit_url( url, allowed_codes=[ 200, 403 ] )
+self.visit_url( url, allowed_codes=[ 200, 403, 404 ] )
self.logout()
self.login( email='t...@bx.psu.edu', username='test' )
install_params = urllib.urlencode( dict(
repository_ids=encoded_repository_id,
@@ -105,7 +105,7 @@
self.check_for_strings( post_submit_strings_displayed,
strings_not_displayed )
repository_ids = self.initiate_installation_process(
new_tool_panel_section=new_tool_panel_section )
self.wait_for_repository_installation( repository_ids )
-def visit_url( self, url, allowed_codes=[ 200 ] ):
+def visit_url( self, url, allowed_codes=[ 200, 404 ] ):
new_url = tc.go( url )
return_code = tc.browser.get_code()
assert return_code in allowed_codes, 'Invalid HTTP return code
%s, allowed codes: %s' % \
So I made the script accept the 404 return code, to see if it can now
successfully run the functional tests of my tool, and it did!
# 2013-10-19 14:38:58 - repository installation and testing script
completed.
# Repository revisions tested: 1
#
--
# 1 repositories passed all tests:
#
# usearch owned by serrano, changeset revision 095baaf5d877
There are two tests in this repository and it passed both. I think this
proves that Twill indeed returns the wrong HTTP code. It returns 404
(and sometimes 500) when it should return 200. I have no idea why this
happens in Galaxy and I'm hoping that someone else can have a look at it.
Regards,
Serrano
___
Please keep all replies on the list by using "reply all"
in your mail client. To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
http://lists.bx.psu.edu/
To search Galaxy mailing lists use the unified search at:
http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional tests broken, NameError: global name 'db_path' is not defined
Hello all, I'm having this same issue with Functional Test being broken. I've switched between 'stable' and 'default' and it doesn't seem to fix the issue. Its possible that I'm not switching properly? This is my first local install of Galaxy so I'm new to the structure, detailed help would be very appreciated. Cheers Ian $ sudo hg update default 0 files updated, 0 files merged, 0 files removed, 0 files unresolved $ hg heads | more changeset: 10421:a477486bf18e branch: stable tag: tip user:Nate Coraor date:Thu Sep 26 11:02:58 2013 -0400 summary: Bugfix for tool-to-destination mapping, tool ids are lowercased but the mapping id was not lowercased. changeset: 10411:c42567f43aa7 user:greg date:Mon Aug 19 13:19:56 2013 -0400 summary: Filter invalid objects when generating the list of repository_dependencies objects that are associated with a tool shed repository installed into Galaxy. $ sudo hg update stable 194 files updated, 0 files merged, 48 files removed, 0 files unresolved $ hg heads | more changeset: 10421:a477486bf18e branch: stable tag: tip user:Nate Coraor date:Thu Sep 26 11:02:58 2013 -0400 summary: Bugfix for tool-to-destination mapping, tool ids are lowercased but the mapping id was not lowercased. changeset: 10411:c42567f43aa7 user:greg date:Mon Aug 19 13:19:56 2013 -0400 summary: Filter invalid objects when generating the list of repository_dependencies objects that are associated with a tool shed repository installed into Galaxy. > Message: 16 > Date: Thu, 3 Oct 2013 12:06:16 +0100 > From: Peter Cock > To: Galaxy Dev > Subject: [galaxy-dev] Functional tests broken,NameError: global name > 'db_path' is not defined > Message-ID: > > Content-Type: text/plain; charset=ISO-8859-1 > > Hi all, > > Something has broken on the functional test framework. > > Attempting to run the tests fails, as does just listing the IDs: > > $ ./run_functional_tests.sh --list > /mnt/galaxy/galaxy-central/eggs/twill-0.9-py2.6.egg/twill/other_packages/_mechanize_dist/_auth.py:14: > DeprecationWarning: the md5 module is deprecated; use hashlib instead > import re, base64, urlparse, posixpath, md5, sha, sys, copy > /mnt/galaxy/galaxy-central/eggs/twill-0.9-py2.6.egg/twill/other_packages/_mechanize_dist/_auth.py:14: > DeprecationWarning: the sha module is deprecated; use the hashlib > module instead > import re, base64, urlparse, posixpath, md5, sha, sys, copy > /mnt/galaxy/galaxy-central/eggs/pysam-0.4.2_kanwei_b10f6e722e9a-py2.6-linux-x86_64-ucs4.egg/pysam/__init__.py:1: > RuntimeWarning: __builtin__.file size changed, may indicate binary > incompatibility > from csamtools import * > Traceback (most recent call last): > File "./scripts/functional_tests.py", line 457, in >sys.exit( main() ) > File "./scripts/functional_tests.py", line 246, in main >database_connection = 'sqlite:///' + os.path.join( db_path, > 'universe.sqlite' ) > NameError: global name 'db_path' is not defined > 'run_functional_tests.sh help' for help > > > $ hg branch > stable > [galaxy@ppserver galaxy-central]$ hg heads | more > changeset: 11805:a3c018044538 > tag: tip > user:greg > date:Wed Oct 02 16:44:48 2013 -0400 > summary: Eliminate the problematic "Select one" option from select fields. > > changeset: 11761:a477486bf18e > branch: stable > parent: 11689:4b1c7ad464ea > user:Nate Coraor > date:Thu Sep 26 11:02:58 2013 -0400 > summary: Bugfix for tool-to-destination mapping, tool ids are > lowercased but the mapping id was not lowercased. > > ... > > > Regards, > > Peter > > > -- > > Message: 17 > Date: Thu, 03 Oct 2013 08:53:47 -0400 > From: Dave Bouvier > To: Peter Cock , Galaxy Dev > > Subject: Re: [galaxy-dev] Functional tests broken, NameError: global > name 'db_path' is not defined > Message-ID: <524d68db.5060...@bx.psu.edu> > Content-Type: text/plain; charset=ISO-8859-1; format=flowed > > Peter, > > This bug was fixed on the default branch in 10427:33092bc258d7, or > whichever changeset that corresponds to in your repository. The relevant > changelog entry, if you want to find and apply it to your repository, is > as follows: > > changeset: 10427:33092bc258d7 > user:Carl Eberhard > date:Wed Aug 21 10:00:09 2013 -0700 > summary: Functional tests: fix sqlite db_path var; fix api test for > hda update with nonsense ke
Re: [galaxy-dev] Functional tests broken, NameError: global name 'db_path' is not defined
On Thu, Oct 3, 2013 at 1:53 PM, Dave Bouvier wrote: > Peter, > > This bug was fixed on the default branch in 10427:33092bc258d7, or whichever > changeset that corresponds to in your repository. The relevant changelog > entry, if you want to find and apply it to your repository, is as follows: > > changeset: 10427:33092bc258d7 > user:Carl Eberhard > date:Wed Aug 21 10:00:09 2013 -0700 > summary: Functional tests: fix sqlite db_path var; fix api test for hda > update with nonsense key > Switching to the default branch did indeed fix it - so this needs to be added to the stable branch (not there yet): https://bitbucket.org/galaxy/galaxy-central/src/a477486bf18eafdd14dd7ba1e91e17f1b05e8121/scripts/functional_tests.py?at=stable Thanks, Peter ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional tests broken, NameError: global name 'db_path' is not defined
On Thu, Oct 3, 2013 at 1:53 PM, Dave Bouvier wrote: > Peter, > > This bug was fixed on the default branch in 10427:33092bc258d7, or whichever > changeset that corresponds to in your repository. The relevant changelog > entry, if you want to find and apply it to your repository, is as follows: > > changeset: 10427:33092bc258d7 > user:Carl Eberhard > date:Wed Aug 21 10:00:09 2013 -0700 > summary: Functional tests: fix sqlite db_path var; fix api test for hda > update with nonsense key > > >--Dave B. Thanks - from the date that ought to be on my development system already, but I can check more specifically now. Peter ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional tests broken, NameError: global name 'db_path' is not defined
Peter, This bug was fixed on the default branch in 10427:33092bc258d7, or whichever changeset that corresponds to in your repository. The relevant changelog entry, if you want to find and apply it to your repository, is as follows: changeset: 10427:33092bc258d7 user:Carl Eberhard date:Wed Aug 21 10:00:09 2013 -0700 summary: Functional tests: fix sqlite db_path var; fix api test for hda update with nonsense key --Dave B. On 10/03/2013 07:06 AM, Peter Cock wrote: Hi all, Something has broken on the functional test framework. Attempting to run the tests fails, as does just listing the IDs: $ ./run_functional_tests.sh --list /mnt/galaxy/galaxy-central/eggs/twill-0.9-py2.6.egg/twill/other_packages/_mechanize_dist/_auth.py:14: DeprecationWarning: the md5 module is deprecated; use hashlib instead import re, base64, urlparse, posixpath, md5, sha, sys, copy /mnt/galaxy/galaxy-central/eggs/twill-0.9-py2.6.egg/twill/other_packages/_mechanize_dist/_auth.py:14: DeprecationWarning: the sha module is deprecated; use the hashlib module instead import re, base64, urlparse, posixpath, md5, sha, sys, copy /mnt/galaxy/galaxy-central/eggs/pysam-0.4.2_kanwei_b10f6e722e9a-py2.6-linux-x86_64-ucs4.egg/pysam/__init__.py:1: RuntimeWarning: __builtin__.file size changed, may indicate binary incompatibility from csamtools import * Traceback (most recent call last): File "./scripts/functional_tests.py", line 457, in sys.exit( main() ) File "./scripts/functional_tests.py", line 246, in main database_connection = 'sqlite:///' + os.path.join( db_path, 'universe.sqlite' ) NameError: global name 'db_path' is not defined 'run_functional_tests.sh help' for help $ hg branch stable [galaxy@ppserver galaxy-central]$ hg heads | more changeset: 11805:a3c018044538 tag: tip user:greg date:Wed Oct 02 16:44:48 2013 -0400 summary: Eliminate the problematic "Select one" option from select fields. changeset: 11761:a477486bf18e branch: stable parent: 11689:4b1c7ad464ea user:Nate Coraor date:Thu Sep 26 11:02:58 2013 -0400 summary: Bugfix for tool-to-destination mapping, tool ids are lowercased but the mapping id was not lowercased. ... Regards, Peter ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/ ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional tests results in Invalid HTTP return code 404/500, allowed codes: 200
Dave, The space isn't there in my actual XML file. I must have accidentally put the space there while constructing the email. The URL that fails with the 404 error is actually correct (as stated in my first post). This means that it successfully obtained the URL from the XML file. So I suspect the problem is somewhere else. Serrano On 09/16/2013 09:30 PM, Dave Bouvier wrote: > It looks like you might have a space in your tool_sheds_conf.xml after > domain.org/ but before the " character, is that the case? If so, I > would suggest removing it and trying again. > >--Dave B. ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] Functional tests results in Invalid HTTP return code 404/500, allowed codes: 200
It looks like you might have a space in your tool_sheds_conf.xml after domain.org/ but before the " character, is that the case? If so, I would suggest removing it and trying again. --Dave B. On 09/16/2013 03:26 PM, Serrano Pereira wrote: Dave, Thanks for your reply. I already had my tool shed specified in tool_sheds_conf.xml, which is why I omitted the GALAXY_INSTALL_TEST_TOOL_SHEDS_CONF environment variable. I just tried it with, GALAXY_INSTALL_TEST_TOOL_SHEDS_CONF=install_and_test_tool_sheds_conf.xml The contents of the install_and_test_tool_sheds_conf.xml is as follows: http://toolshed.galaxy.domain.org/ "/> But it results in the exact same error: AssertionError: Invalid HTTP return code 404, allowed codes: 200 Do you have any other ideas? Regards, Serrano On 09/16/2013 08:37 PM, Dave Bouvier wrote: Serrano, The GALAXY_INSTALL_TEST_TOOL_SHEDS_CONF environment variable should be set to an XML file that specifies the tool shed you're installing from. I've updated the wiki page to be clearer on that point. --Dave B. On 09/16/2013 12:32 PM, Serrano Pereira wrote: Dear Galaxy hackers, I'm having trouble with the tool test framework. I have a local installation of a Galaxy Tool Shed (latest stable from the repository) on my Linux server. I have an Apache proxy to my tool shed (per http://wiki.galaxyproject.org/HostingALocalToolShed). The Tool Shed seems to work fine. I'm able to create and maintain repositories and I'm able to install tools from the Tool Shed from remote Galaxy instances. Now I want to run the functional tests of the tools in the Tool Shed. I do that by following the instructions on the wiki (http://wiki.galaxyproject.org/AutomatedToolTests). It comes down to the following commands: -- cd /opt/galaxy/galaxy-dist/ export GALAXY_INSTALL_TEST_TOOL_SHED_API_KEY=2aeecfe2d9f28d0c5b6959a41886b5fd export GALAXY_INSTALL_TEST_TOOL_SHED_URL=http://toolshed.galaxy.domain.org/ python lib/tool_shed/scripts/check_repositories_for_functional_tests.py tool_shed_wsgi.ini sh install_and_test_tool_shed_repositories.sh -- The second script (install_and_test_tool_shed_repositories.sh) fails with the following error: == FAIL: install_repository_usearch (install_and_test_tool_shed_repositories.functional.test_install_repositories.TestInstallRepository_usearch) Install the repository usearch from http://toolshed.galaxy.domain.org/. -- Traceback (most recent call last): File "/opt/galaxy/galaxy-dist/test/install_and_test_tool_shed_repositories/functional/test_install_repositories.py", line 48, in test_install_repository self.do_installation( repository_dict ) File "/opt/galaxy/galaxy-dist/test/install_and_test_tool_shed_repositories/functional/test_install_repositories.py", line 17, in do_installation self.install_repository( repository_info_dict ) File "/opt/galaxy/galaxy-dist/test/install_and_test_tool_shed_repositories/base/twilltestcase.py", line 61, in install_repository self.visit_url( '%s/repository/preview_tools_in_changeset?%s' % ( tool_shed_url, preview_params ) ) File "/opt/galaxy/galaxy-dist/test/install_and_test_tool_shed_repositories/base/twilltestcase.py", line 112, in visit_url ( return_code, url, ', '.join( str( code ) for code in allowed_codes ) ) AssertionError: Invalid HTTP return code 404 for http://toolshed.galaxy.domain.org//repository/preview_tools_in_changeset?repository_id=3651c288c4818b12&changeset_revision=095baaf5d877, allowed codes: 200 -- (I've changed twilltestcase.py to also print the URL that is being browsed.) This is were it get's weird. The HTTP 404 response code indicates that the page was not found. However, when I browse to that URL, it works fine! What's more, when I try to access that page with `wget` from the same machine the HTTP response code is 200: -- wget http://toolshed.galaxy.domain.org//repository/preview_tools_in_changeset?repository_id=3651c288c4818b12&changeset_revision=095baaf5d877 ... HTTP request sent, awaiting response... 200 OK ... -- When I use the Tool Shed URL http://localhost:9009/ instead, I get a similar error, only the return code is 500 instead of 404. I suspect that the HTTP return code returned by the Python module that returns it (Mechanize?) is invalid. I found a note about this error on the wiki (http://wiki.galaxyproject.org/Admin/Running%20Tests): "Note: If you have another version of paste installed in your PYTHONPATH, it can cause problems running the functional tests. Many tests will e
Re: [galaxy-dev] Functional tests results in Invalid HTTP return code 404/500, allowed codes: 200
Dave, Thanks for your reply. I already had my tool shed specified in tool_sheds_conf.xml, which is why I omitted the GALAXY_INSTALL_TEST_TOOL_SHEDS_CONF environment variable. I just tried it with, GALAXY_INSTALL_TEST_TOOL_SHEDS_CONF=install_and_test_tool_sheds_conf.xml The contents of the install_and_test_tool_sheds_conf.xml is as follows: http://toolshed.galaxy.domain.org/ "/> But it results in the exact same error: AssertionError: Invalid HTTP return code 404, allowed codes: 200 Do you have any other ideas? Regards, Serrano On 09/16/2013 08:37 PM, Dave Bouvier wrote: > Serrano, > > The GALAXY_INSTALL_TEST_TOOL_SHEDS_CONF environment variable should be > set to an XML file that specifies the tool shed you're installing > from. I've updated the wiki page to be clearer on that point. > >--Dave B. > > On 09/16/2013 12:32 PM, Serrano Pereira wrote: >> Dear Galaxy hackers, >> >> I'm having trouble with the tool test framework. I have a local >> installation of a Galaxy Tool Shed (latest stable from the repository) >> on my Linux server. I have an Apache proxy to my tool shed (per >> http://wiki.galaxyproject.org/HostingALocalToolShed). The Tool Shed >> seems to work fine. I'm able to create and maintain repositories and I'm >> able to install tools from the Tool Shed from remote Galaxy instances. >> >> Now I want to run the functional tests of the tools in the Tool Shed. I >> do that by following the instructions on the wiki >> (http://wiki.galaxyproject.org/AutomatedToolTests). It comes down to the >> following commands: >> >> -- >> cd /opt/galaxy/galaxy-dist/ >> export >> GALAXY_INSTALL_TEST_TOOL_SHED_API_KEY=2aeecfe2d9f28d0c5b6959a41886b5fd >> export >> GALAXY_INSTALL_TEST_TOOL_SHED_URL=http://toolshed.galaxy.domain.org/ >> python lib/tool_shed/scripts/check_repositories_for_functional_tests.py >> tool_shed_wsgi.ini >> sh install_and_test_tool_shed_repositories.sh >> -- >> >> The second script (install_and_test_tool_shed_repositories.sh) fails >> with the following error: >> >> == >> FAIL: install_repository_usearch >> (install_and_test_tool_shed_repositories.functional.test_install_repositories.TestInstallRepository_usearch) >> >> Install the repository usearch from http://toolshed.galaxy.domain.org/. >> -- >> Traceback (most recent call last): >>File >> "/opt/galaxy/galaxy-dist/test/install_and_test_tool_shed_repositories/functional/test_install_repositories.py", >> >> line 48, in test_install_repository >> self.do_installation( repository_dict ) >>File >> "/opt/galaxy/galaxy-dist/test/install_and_test_tool_shed_repositories/functional/test_install_repositories.py", >> >> line 17, in do_installation >> self.install_repository( repository_info_dict ) >>File >> "/opt/galaxy/galaxy-dist/test/install_and_test_tool_shed_repositories/base/twilltestcase.py", >> >> line 61, in install_repository >> self.visit_url( '%s/repository/preview_tools_in_changeset?%s' % ( >> tool_shed_url, preview_params ) ) >>File >> "/opt/galaxy/galaxy-dist/test/install_and_test_tool_shed_repositories/base/twilltestcase.py", >> >> line 112, in visit_url >> ( return_code, url, ', '.join( str( code ) for code in >> allowed_codes ) ) >> AssertionError: Invalid HTTP return code 404 for >> http://toolshed.galaxy.domain.org//repository/preview_tools_in_changeset?repository_id=3651c288c4818b12&changeset_revision=095baaf5d877, >> >> allowed codes: 200 >> -- >> >> (I've changed twilltestcase.py to also print the URL that is being >> browsed.) This is were it get's weird. The HTTP 404 response code >> indicates that the page was not found. However, when I browse to that >> URL, it works fine! What's more, when I try to access that page with >> `wget` from the same machine the HTTP response code is 200: >> >> -- >> wget >> http://toolshed.galaxy.domain.org//repository/preview_tools_in_changeset?repository_id=3651c288c4818b12&changeset_revision=095baaf5d877 >> >> ... >> HTTP request sent, awaiting response... 200 OK >> ... >> -- >> >> When I use the Tool Shed URL http://localhost:9009/ instead, I get a >> similar error, only the return code is 500 instead of 404. I suspect >> that the HTTP return code returned by the Python module that returns it >> (Mechanize?) is invalid. I found a note about this error on the wiki >> (http://wiki.galaxyproject.org/Admin/Running%20Tests): >> >> "Note: If you have another version of paste installed in your >> PYTHONPATH, it can cause problems running the functional tests. Many >> tests will e
Re: [galaxy-dev] Functional tests results in Invalid HTTP return code 404/500, allowed codes: 200
Serrano, The GALAXY_INSTALL_TEST_TOOL_SHEDS_CONF environment variable should be set to an XML file that specifies the tool shed you're installing from. I've updated the wiki page to be clearer on that point. --Dave B. On 09/16/2013 12:32 PM, Serrano Pereira wrote: Dear Galaxy hackers, I'm having trouble with the tool test framework. I have a local installation of a Galaxy Tool Shed (latest stable from the repository) on my Linux server. I have an Apache proxy to my tool shed (per http://wiki.galaxyproject.org/HostingALocalToolShed). The Tool Shed seems to work fine. I'm able to create and maintain repositories and I'm able to install tools from the Tool Shed from remote Galaxy instances. Now I want to run the functional tests of the tools in the Tool Shed. I do that by following the instructions on the wiki (http://wiki.galaxyproject.org/AutomatedToolTests). It comes down to the following commands: -- cd /opt/galaxy/galaxy-dist/ export GALAXY_INSTALL_TEST_TOOL_SHED_API_KEY=2aeecfe2d9f28d0c5b6959a41886b5fd export GALAXY_INSTALL_TEST_TOOL_SHED_URL=http://toolshed.galaxy.domain.org/ python lib/tool_shed/scripts/check_repositories_for_functional_tests.py tool_shed_wsgi.ini sh install_and_test_tool_shed_repositories.sh -- The second script (install_and_test_tool_shed_repositories.sh) fails with the following error: == FAIL: install_repository_usearch (install_and_test_tool_shed_repositories.functional.test_install_repositories.TestInstallRepository_usearch) Install the repository usearch from http://toolshed.galaxy.domain.org/. -- Traceback (most recent call last): File "/opt/galaxy/galaxy-dist/test/install_and_test_tool_shed_repositories/functional/test_install_repositories.py", line 48, in test_install_repository self.do_installation( repository_dict ) File "/opt/galaxy/galaxy-dist/test/install_and_test_tool_shed_repositories/functional/test_install_repositories.py", line 17, in do_installation self.install_repository( repository_info_dict ) File "/opt/galaxy/galaxy-dist/test/install_and_test_tool_shed_repositories/base/twilltestcase.py", line 61, in install_repository self.visit_url( '%s/repository/preview_tools_in_changeset?%s' % ( tool_shed_url, preview_params ) ) File "/opt/galaxy/galaxy-dist/test/install_and_test_tool_shed_repositories/base/twilltestcase.py", line 112, in visit_url ( return_code, url, ', '.join( str( code ) for code in allowed_codes ) ) AssertionError: Invalid HTTP return code 404 for http://toolshed.galaxy.domain.org//repository/preview_tools_in_changeset?repository_id=3651c288c4818b12&changeset_revision=095baaf5d877, allowed codes: 200 -- (I've changed twilltestcase.py to also print the URL that is being browsed.) This is were it get's weird. The HTTP 404 response code indicates that the page was not found. However, when I browse to that URL, it works fine! What's more, when I try to access that page with `wget` from the same machine the HTTP response code is 200: -- wget http://toolshed.galaxy.domain.org//repository/preview_tools_in_changeset?repository_id=3651c288c4818b12&changeset_revision=095baaf5d877 ... HTTP request sent, awaiting response... 200 OK ... -- When I use the Tool Shed URL http://localhost:9009/ instead, I get a similar error, only the return code is 500 instead of 404. I suspect that the HTTP return code returned by the Python module that returns it (Mechanize?) is invalid. I found a note about this error on the wiki (http://wiki.galaxyproject.org/Admin/Running%20Tests): "Note: If you have another version of paste installed in your PYTHONPATH, it can cause problems running the functional tests. Many tests will error (instead of passing/failing) because of an HTML code 500 where a 200 was expected. There are two workarounds for this: uninstall paste or use virtualenv. " The thing is, I do not have another version of Paste installed. What can I do to fix this error? Regards, Serrano Pereira ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/ ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, pl
Re: [galaxy-dev] Functional tests
Frank, I can't see anything that would cause this behavior by looking at your logs, so I cannot answer what is causing this problem in your environment. Basically it seems that nose is not finding any of the Python test scripts in the test subdirectory in your Galaxy installation directory, but I am not able to tell you why (I assume you haven't moved any directories). Sorry for not being of any help here. Greg Von Kuster On Feb 21, 2012, at 2:55 AM, Frank Sørensen wrote: > Hi Greg, > > Den 20-02-2012 15:06, Greg Von Kuster skrev: >> Hello Frank, see my inline comments. >> >> On Feb 20, 2012, at 6:56 AM, Frank Sørensen wrote: >> >>> Hi Guys, >>> >>> Sorry if this is a stupid question, but in order to find out which external >>> reference files (.loc - list files etc.) we need for our newly installed >>> Galaxy server, and to get an overview of which tools are in working order, >>> I have tried for a couple of days now, to run the script >>> "run_functional_tests.sh". >> What is the exact command line that you are using to attempt to execute the >> functional tests? Are you using the following? >> >> %sh run_functional_test.sh > Yes, exactly - from .../galaxy-dist: ./run_functional_tests.sh >> >> >>> I have copied tool_conf.xml over tool_conf.xml.sample, >> The functional tests use tool_conf.xml.sample, so what are the contents of >> that file in your environment? > It is the same as tool_conf.xml. The complete list of all the tools that come > with the Galaxy installation, plus our own. >>> and I've tried to run all scripts mentioned in >>> http://wiki.g2.bx.psu.edu/Main/Tests/Framework. >> I assume by the above you mean this link: >> >> http://wiki.g2.bx.psu.edu/Main/Tests%20Framework > Yes ofcourse... >> >> Here is another link that provide additional information for testing: >> >> http://wiki.g2.bx.psu.edu/Admin/Running%20Tests > Thanks, I'll check it out. >> >>> I have tried to stop the server and run run_functional_tests.sh and all I >>> get is the message saying "Ran 0 tests in 0.000s", and an empty >>> run_functional_tests.html - file. >>> I haven't been able to find any documentation to clarify this matter. Can >>> anyone give me a pointer please? >> >> We'll need more information about your environment, but we'll get this >> figured out. > %uname -a > Linux moma 2.6.32-21-server #32-Ubuntu SMP Fri Apr 16 09:17:34 UTC 2010 > x86_64 GNU/Linux > > I have attached the result of %lshw and the output from > run_functional_tests.sh and our global settings file. > >>> Our Galaxy server is installed using Mercurial on an Ubuntu server. I have >>> written tool-wrappers for some of our own programs, and they all work fine. >>> >> > Kind regards > > - Frank > > -- > Frank Sørensen, B.Sc., Programmer > Molecular Diagnostic Laboratory (MDL) > Molekylær Medicinsk Afdeling (MOMA) > Århus Universitetshospital Skejby, Brendstrupgårdsvej, 8200 Århus N > Tlf. +45 7845 5363 > ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/
Re: [galaxy-dev] Functional tests
Hello Frank, see my inline comments. On Feb 20, 2012, at 6:56 AM, Frank Sørensen wrote: > Hi Guys, > > Sorry if this is a stupid question, but in order to find out which external > reference files (.loc - list files etc.) we need for our newly installed > Galaxy server, and to get an overview of which tools are in working order, I > have tried for a couple of days now, to run the script > "run_functional_tests.sh". What is the exact command line that you are using to attempt to execute the functional tests? Are you using the following? %sh run_functional_test.sh > > I have copied tool_conf.xml over tool_conf.xml.sample, The functional tests use tool_conf.xml.sample, so what are the contents of that file in your environment? > and I've tried to run all scripts mentioned in > http://wiki.g2.bx.psu.edu/Main/Tests/Framework. I assume by the above you mean this link: http://wiki.g2.bx.psu.edu/Main/Tests%20Framework Here is another link that provide additional information for testing: http://wiki.g2.bx.psu.edu/Admin/Running%20Tests > I have tried to stop the server and run run_functional_tests.sh and all I get > is the message saying "Ran 0 tests in 0.000s", and an empty > run_functional_tests.html - file. > I haven't been able to find any documentation to clarify this matter. Can > anyone give me a pointer please? We'll need more information about your environment, but we'll get this figured out. > > Our Galaxy server is installed using Mercurial on an Ubuntu server. I have > written tool-wrappers for some of our own programs, and they all work fine. > > Regards > > - Frank > > -- > Frank Sørensen, B.Sc., Programmer > Molecular Diagnostic Laboratory (MDL) > Molekylær Medicinsk Afdeling (MOMA) > Århus Universitetshospital Skejby, Brendstrupgårdsvej, 8200 Århus N > Tlf. +45 7845 5363 > ___ > Please keep all replies on the list by using "reply all" > in your mail client. To manage your subscriptions to this > and other Galaxy lists, please use the interface at: > > http://lists.bx.psu.edu/ > ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/
