Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Aug 16 21:24
    Midnighter commented #626
  • Aug 16 15:11
    JonathanRob commented #626
  • Aug 16 15:10
    JonathanRob commented #626
  • Aug 13 09:34
    Midnighter commented #670
  • Aug 06 08:00
    ChristianLieven synchronize #670
  • Aug 06 08:00

    ChristianLieven on register_marks

    fix: flake8 chore: add entry to changelog (compare)

  • Aug 05 13:00
    ChristianLieven edited #670
  • Aug 05 12:50
    ChristianLieven edited #670
  • Aug 05 12:50
    ChristianLieven opened #670
  • Aug 05 12:34

    ChristianLieven on register_marks

    chore: register custom pytest m… (compare)

  • Jul 18 03:30
    mjenior commented #669
  • Jul 17 14:49
    Midnighter commented #669
  • Jul 17 12:55
    mjenior closed #669
  • Jul 17 12:53
    mjenior commented #669
  • Jul 16 20:34
    Midnighter commented #669
  • Jul 16 18:29
    mjenior opened #669
  • Jun 24 10:38

    Midnighter on chore-upgrade

    (compare)

  • Jun 24 10:38

    Midnighter on develop

    chore: upgrade pandas chore: unpin click chore: avoid deprecated `read_t… and 4 more (compare)

  • Jun 24 10:37
    Midnighter closed #668
  • Jun 24 10:37
    Midnighter closed #665
Moritz E. Beber
@Midnighter
With regard to the history report, it should work for a repo that is properly set up but we will actually spend time starting next week to bring our demo repositories up to speed. So we will probably find small bugs and can soon give you detailed information for how to set up the repo.
Snorre Sulheim
@sulheim
Okey! Another, maybe stupid question: I’ve almost managed to get it to work, but now travis complaints about the GITHUB_TOKEN not being defined. What am I doing wrong?
Snorre Sulheim
@sulheim
Never mind :)
Moritz E. Beber
@Midnighter
:+1: :)
Snorre Sulheim
@sulheim
When I run memote history report memote checkout gh-pages branch, and then just goes "Model was not modified in commit ..." forever. Seems like it is running in a loop because it repeats checking the same commits with som interval. Any idea of whats going on?
Moritz E. Beber
@Midnighter
Hmm, not immediately. Is that repo public so I could take a look?
Snorre Sulheim
@sulheim
Yes, https://github.com/SysBioChalmers/sco-GEM. However, it finished after 10 minutes, but the history report doesn't look good (index.html) in gh-pages branch. Seems like I need to create result-files (json.gz for old commits), see the log-file from travis
Snorre Sulheim
@sulheim
I'm trying to run memote history to re-create the results from previous commits, but then I get this weird error message:
Traceback (most recent call last):
  File "c:\users\snorres\appdata\local\continuum\anaconda3\envs\memote\lib\runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "c:\users\snorres\appdata\local\continuum\anaconda3\envs\memote\lib\runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "C:\Users\snorres\AppData\Local\Continuum\Anaconda3\envs\memote\Scripts\memote.exe\__main__.py", line 9, in <module>
  File "c:\users\snorres\appdata\local\continuum\anaconda3\envs\memote\lib\site-packages\click\core.py", line 722, in __call__
    return self.main(*args, **kwargs)
  File "c:\users\snorres\appdata\local\continuum\anaconda3\envs\memote\lib\site-packages\click\core.py", line 697, in main
    rv = self.invoke(ctx)
  File "c:\users\snorres\appdata\local\continuum\anaconda3\envs\memote\lib\site-packages\click\core.py", line 1066, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "c:\users\snorres\appdata\local\continuum\anaconda3\envs\memote\lib\site-packages\click\core.py", line 895, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "c:\users\snorres\appdata\local\continuum\anaconda3\envs\memote\lib\site-packages\click\core.py", line 535, in invoke
    return callback(*args, **kwargs)
  File "c:\users\snorres\appdata\local\continuum\anaconda3\envs\memote\lib\site-packages\memote\suite\cli\runner.py", line 309, in history
    callbacks.validate_path(model)
AttributeError: module 'memote.suite.cli.callbacks' has no attribute 'validate_path'
Moritz E. Beber
@Midnighter
Sorry, I'm at a conference these days, so I'll only be able to look at this later in the week. Just give me until the end of the week to sort things out.
Snorre Sulheim
@sulheim
Thanks, no rush Moritz. Another, hopefully simpler question: What should the custom config file for custom tests look like?
Christian Lieven
@ChristianLieven
@sulheim did you run this with the latest release? Could you try with the branch cli_tests?
Moritz E. Beber
@Midnighter

What should the custom config file for custom tests look like?

For the tests themselves you just need a directory with Python test modules. No configuration required other than making memote aware of the directory. https://memote.readthedocs.io/en/latest/custom_tests.html
You only need a configuration if you want to create separate sections for the new custom tests. If you don't, they will all show up at the end of the report anyhow.

Snorre Sulheim
@sulheim
For the snapshot it works with just pointing to a directory with the --custom-testsflag, however for the history it seems like --custom-configis the only option
@ChristianLieven Yes, ran it with 0.9.6, but I can give it a try on the cli_testsbranch¨'
Christian Lieven
@ChristianLieven

For the snapshot it works with just pointing to a directory with the --custom-testsflag, however for the history it seems like --custom-configis the only option

Okay I’ll look into that!

Snorre Sulheim
@sulheim
Thanks @ChristianLieven!
Joshua Chan
@shjchan

Hi @ChristianLieven , was just wandering around and saw this:

Oops. I misspoke! This is the test that previously needed loopless FVA find_stoichiometrically_balanced_cycles. It’s now running regular FVA but still quite slowly.

Is it still doing FVA in the latest version? We do have an algorithm way faster than FVA to find out reactions in cycles. See section 2.5 in this paper:
https://academic.oup.com/bioinformatics/article/34/24/4248/5026660
Another method still way faster than FVA, but probably slower than above (compared in Fig. 4 in the paper above):
https://academic.oup.com/bioinformatics/article/32/24/3807/2525667

Christian Lieven
@ChristianLieven

Hi @shjchan, Thanks for making us aware! Yes indeed it is still running the FVA method implemented in cobra.flux_analysis. flux_variability_analysis. I should point out that with memote we rely on a pure Python stack, essentially making use of whatever exists in the cobrapy & optlang ecosystem (https://puu.sh/D9ox4/e8840cbef4.png). Have you considered implementing your algorithm there? Moreover, do you happen to know how it compares against ExtraFastCC published here: https://link.springer.com/article/10.1186/s12859-019-2615-x#Sec7?

Would you mind perhaps cross-posting your comment to the cobrapy gitter channel too? @Midnighter, @cdiener and @pstjohn may be able to appreciate the algorithmic details a little more ;). Possibly they would also know where and how that is best implemented such that memote can benefit from it.

Although my experience with implementing algorithms and my capacity for side-projects is quite limited, I’m happy to help if you have any questions with regards to re-implementation in Python.

Moritz E. Beber
@Midnighter
Definitely sounds interesting @shjchan. Indeed, a comparison with ExtraFastCC would be cool.
Joshua Chan
@shjchan
Hi @Midnighter, do you know where I can find the implementation of ExtraFastCC alone? Seems like it is bundled in the ModelExplorer executable. And I could not find the math formulation or algorithmic procedure. @ChristianLieven , I can probably try that.
Moritz E. Beber
@Midnighter

So we are kept in suspense.

As this paper is focused on the visualization tool ModelExplorer, we leave the precise details of the algorithm to another paper.

Christian Lieven
@ChristianLieven
I may have jumped the gun there. I know that the ExtraFastCC paper was due being published from last years COBRA conference but I had thought this would be it.
Moritz E. Beber
@Midnighter
I must admit that I'm really disappointed by what was allowed to happen there. ModelExplorer is distributed under a CC-BY-NO$ license which the creative commons recommend themselves to not apply to software (https://creativecommons.org/faq/#can-i-apply-a-creative-commons-license-to-software) and no access to any of the sources either.
Christian Lieven
@ChristianLieven

Oh no.. soo I guess they disqualified themselves from being used by memote then?

Additionally, our licenses are currently not compatible with the major software licenses, so it would be difficult to integrate CC-licensed work with other free software.

Moritz E. Beber
@Midnighter
Well, if they publish the algorithm we can still re-implement it around cobrapy. From this manuscript it's not clear to me, however, if the algorithm really has a better complexity or just makes good use of GPU calculations and thus has speed gains from that.
Matthias König
@matthiaskoenig
Hi all, how can I get detailed information on the failed tests when creating a report for a model?
For instance I am running: memote report snapshot --filename tiny_example_11_memote.html tiny_example_11.xml
But I only get
../../../../../../memote/memote/suite/tests/test_biomass.py Fsssssssss [ 67%]
i.e. an overview over the failed tests. But I want to see the stacktrace why the test failed
Is there an easy way to run all tests on a model and get access to all the stack traces of the failed tests?
This is especially interesting for my tests which errored, because I want to see what went wrong.
Christian Lieven
@ChristianLieven
Hey Matthias, try memote --verbosity DEBUG run tiny_example_11.xml, that should give you the full traceback for each test!
Moritz E. Beber
@Midnighter
Actually, in this case you need to adjust the pytest information level. You can use -a in order to add arbitrary pytest arguments.
memote report snapshot -a "-vv --tb=long" --filename tiny_example_11_memote.html tiny_example_11.xml
Matthias König
@matthiaskoenig
@ChristianLieven and @Li
@ChristianLieven and @Midnighter thanks for the information
Joshua Chan
@shjchan

@ChristianLieven @Midnighter

Hi @shjchan, Thanks for making us aware! Yes indeed it is still running the FVA method implemented in cobra.flux_analysis. flux_variability_analysis. I should point out that with memote we rely on a pure Python stack, essentially making use of whatever exists in the cobrapy & optlang ecosystem (https://puu.sh/D9ox4/e8840cbef4.png). Have you considered implementing your algorithm there? Moreover, do you happen to know how it compares against ExtraFastCC published here: https://link.springer.com/article/10.1186/s12859-019-2615-x#Sec7?

Would you mind perhaps cross-posting your comment to the cobrapy gitter channel too? @Midnighter, @cdiener and @pstjohn may be able to appreciate the algorithmic details a little more ;). Possibly they would also know where and how that is best implemented such that memote can benefit from it.

Although my experience with implementing algorithms and my capacity for side-projects is quite limited, I’m happy to help if you have any questions with regards to re-implementation in Python.

Hi Christian and Moritz,

I finally have time to take a look at this. Actually I see that cobrapy already has fastcc. So an easy way to speed up is just to change memote.support.consistency.find_stoichiometrically_balanced_cycles.
Instead of doing FVA there, you can just do

model_int = cobra.flux_analysis.fastcc(model)
return [r.id for r in model_int.reactions]

I have tested a little bit just using the iJO1366 model. Using cobrapy, FVA took 41 sec and Fastcc took 14 sec, a pretty significant improvement already. The results were exactly the same because fastcc was designed for finding all active reactions. When you shut down all exchange reactions, it will get all reactions lying in loops.
I haven't had chance to implement my algorithm in cobrapy, though in COBRAtoolbox it was just a blink (0.3 sec) while if I use the fastcc in the COBRAtoolbox, it took a rather long time (39 sec).

Moritz E. Beber
@Midnighter
Hi Joshua, I believe fastcc has some issues at the moment opencobra/cobrapy#806
And I'm guessing that you were using multiple cores for FVA since 41 seconds is quite fast. Many workflows will run only on one core so getting your algorithm implemented would be very helpful indeed.
Sanzhar Naizabekov
@ensakz
Screenshot from 2019-05-23 11-43-56.png

Hi everyone,

I tried to run memote report snapshot function to assess my draft GEM. The report seems to be working ok, however, after several seconds, the numbers in the Independent Section of the report disappear behind Specific Section of the report. It seems to me that there is some problems with html parsing when memote creates snapshot report. I tried different browsers to load my html report (Chrome, Mozilla, Microsoft Edge) but the problems seem to appear in every case. I am attaching screenshot of the report problem as well as index.html file for my report. I am writing this message just to ask - is there any suggestions regarding this problem?

Best regards,
Sanzhar

Moritz E. Beber
@Midnighter
Hi @ensakz, Thank you for the report. That's definitely concerning. I have not experienced this issue myself. Have you tried reloading the page in the same browser? Does the issue occur again every time?
Sanzhar Naizabekov
@ensakz
@Midnighter yes, I tried. It seems the problem is consistent
Moritz E. Beber
@Midnighter
@ChristianLieven have you seen this before? I'm not sure how to debug that right now since it doesn't occur for me. @ensakz I assume you're on Windows 10?
Sanzhar Naizabekov
@ensakz
@Midnighter i checked report on Mozilla on Linux, and on Chrome and Microsoft Edge on windows 10
Melclic
@Melclic
Hello. I am a bioinformatician working on a pan-European project called IBISBA (https://www.ibisba.eu). We are developing a platform for metabolic engineering tools using the Galaxy workflow. One of the major feature of the platform is that all communication between the tools is done using the SBML format. I am very interested in your tool because of the report feature. My questions are the following. Do you know of projects or people that integrated your tool into Galaxy? I am also interested in the version control of the model. Does it work only on GIT version control? Do you have any additional information on that aspect of the project. Thanks a lot.
Christian Lieven
@ChristianLieven
@ensakz what is your screen's resolution? I’ve checked for scaling some but not all resolutions consistently. Could you try the same report on a different screen, or send us the generated HTML for debugging?
Christian Lieven
@ChristianLieven
Hi @Melclic thank you for your interest! Sorry, I am not aware of anyone integrating memote into Galaxy yet. Yes, memote so far is tightly integrated with git. For more information on that you could take a look at the biorxiv publication: https://www.biorxiv.org/content/10.1101/350991v1, git repositories that use memote: https://github.com/ChristianLieven/memote-m-capsulatus or our documentation that describes the commands to set up the integration: https://memote.readthedocs.io/en/latest/getting_started.html#ci-tested-online-and-public-workflow