molecularbioinformatics / sbml-mod-ws Goto Github PK

Around line 86 in src/sbmlmod/ModelEditor.py, there is a warning issued, that x species were not found. This warning is issued unconditioned and is also added, if all species were found.
Also, there is a space missing between the number and the warning.

I have not checked it, but can the backend work on Win or MAC platforms as well?

We should at least provide a short overview which function do what and how different parameters are handled. Otherwise nobody outside our team will have the chance to modify anything and it would thus actually make no sense to make the repository public.

Currently, the frontend offers the checkbox to use or not use case-sensitive mapping. Neither the wsdl nor the backend code (after skimming through it) offers this choice. Do we want the choice or not. I guess, not.
If not, the frontend needs to be changed (→ open an issue there, if the feature is not wanted).

This will be possible (and necessary), once the service migrated to UiT.

The example python script in testclient needs to be updated and further help and documentation is needed. If I understand it correctly it can not be run as it is out of the box. That should be the case. Define default values (if not done already)!

I get a processing failure with one of my data files for unknown reasons. Can send the file offline.

The Stocheometry.py file in src/sbmlmod is not used. Should it be deleted?

In ManipulateKineticParameters.py lines 372+ executeReplaceInitialConcentrationsOfSpecies does not properly handle batchmode if only one model file is given. Only the first data column is considered and hence, only one model file is generated. If several model files are given (from an earlier call of setting reaction parameters), it runs fine.

Repository does not seem up to date. Please commit and comment.

The test testWrongFileAsInputToValidateSBMLModelTriggersNotValidModelResponse fails with a newer libSBML version.

In the test, the number of errors is tested and is expected to be 2. With libSBML version 5.11.4, these errors were:

'Main XML content is empty.'
'An SBML document must contain a <model> element. The <model> element is optional in L3V2 and beyond. [...]'

However, with libSBML version 5.13.0, there is only one error, that says:

'XML content is not well-formed.'

To get libSBML version: pip show python-libsbml

What to do? Delete the test? Change the test?

We need to decide for some license (GPL (version?), MIT, etc.)

Provide the whole pipeline including the generation of cumulated concentration and flux-tables and visualisation as bash-script (I could do that) we could also stick to python scripts (Mathias or Sascha)?

The client code in src/pyclient seems to be private code with lots of hard coded file paths. To become useful it should be refactored to use files in src/testsbmlmod/resources. Good examples of clients would be a good thing. Alternatively the code should be deleted.

The testclient sbml-mod-ws/src/testClient/pyClientSample.py should work with a local wsdl.

Changing it to the local wsdl at ./src/sbmlmod/sbmlmod.wsdl, however, raises an error.
Using the one from http://cbu.bioinfo.no/wsdl/SBMLedit.wsdl locally resolves the issue, but still requires an online connection.

In the readme, it says to use ./bin/serverd --help to get help, but the only thing that happens is a comment saying Unknown command.
I'd just remove the hint to --help...

It would be good to give some basic info on how to run SBMLmod with this module only without the web frontend (sbml_mod_web). This information could easily be added to the existing readme, I think.

FAIL: testWrongFileAsInputToValidateSBMLModelTriggersNotValidModelResponse (testsbmlmod.TestSBMLmod.TestSBMLmod)

Traceback (most recent call last):
File "..../sbml-mod-ws/src/testsbmlmod/TestSBMLmod.py", line 57, in testWrongFileAsInputToValidateSBMLModelTriggersNotValidModelResponse
self.assertEquals(3, len(response.get_element_ErrorMessages()))
AssertionError: 3 != 2

Is this a problem with the code or the test?

Trying to run the pyClientSample script in the testClient folder does not finish.
Results in
urllib2.URLError: <urlopen error [Errno 113] No route to host>
error.

1. Can you replicate the error?
2. Suggestions?

• While compiling the wsdl assocaited documentation, I found two redundant entries in the wsdl file, which I deleted.
• I changed all esysbio addresses to uit sbmlmod addresses.
• I deleted two schema entries in the preamble, as I don't know where these are at the uit servers.
• I created a techn. Documentation, sensitive to these changes.

All of this is realised in the branch wsdl_modification.
Everything should still work fine, but I want to make sure: Please check that the application works fine. If so, we can merge into the master branch, as esysbio links should not be present anymore.
I will run checks as well.