IdXMLFile.load() raises "can not handle type"

[yueqixuan]

Describe the problem you encountered

An exception is raised when calling IdXMLFile().load() with a valid idXML file.

File pyopenms/_pyopenms_3.pyx:3368, in pyopenms._pyopenms_3.IdXMLFile.load()
Exception: can not handle type of ('resources/quantms/LFQ_PXD007683/a05191_consensus.idXML', [], [])

This error did not occur in versions ≤ 3.4.0.

To Reproduce
Steps to reproduce the behavior:

1. Data
   pmultiqc/example/LFQ or pmultiqc/example/LFQ

2. Code

from pyopenms import IdXMLFile

raw_id = "resources/quantms/LFQ_PXD007683/a05191_consensus.idXML"

protein_ids: list = []
peptide_ids: list = []
IdXMLFile().load(raw_id, protein_ids, peptide_ids)

3. Error

---------------------------------------------------------------------------
Exception                                 Traceback (most recent call last)
Cell In[1], line 10
      8 protein_ids: list = []
      9 peptide_ids: list = []
---> 10 IdXMLFile().load(raw_id, protein_ids, peptide_ids)

File pyopenms/_pyopenms_3.pyx:3368, in pyopenms._pyopenms_3.IdXMLFile.load()

Exception: can not handle type of ('resources/quantms/LFQ_PXD007683/a05191_consensus.idXML', [], [])

Screenshots

System information:

• OS: Ubuntu
• OS Version: 24.04.2 LTS
• Python version: 3.11.0
• pyopenms version: 3.5.0
• How did you install pyopenms? Please cross a box with an X.
  • conda::bioconda: Specify your conda/mamba command and/or your environment with conda list.
  • conda::openms: Specify your conda/mamba command and/or your environment with conda list.
  • pip
  • nightly wheel
  • built from source

Additional context

[coderabbitai]

📝 CodeRabbit Plan Mode

Generate an implementation plan and prompts that you can use with your favorite coding agent.

• Create Plan

Examples

• Example 1
• Example 2

---

🔗 Similar Issues

Related Issues

• TOPPView fails to load mzML it stored using "save visible area" OpenMS#8007
• FeatureMap annotation can not handle FeatureMapDF OpenMS#8139
• Crash when using MzMLFile().load() function OpenMS#8290
• 'pyopenms._pyopenms_6.MSChromatogram' object has no attribute 'get_df' #458
• OPENMS_DATA_PATH issue in pyopenms 3.4 on windows OpenMS#8356

🔗 Related PRs

#471 - add SpectrumLookup example [merged]
#472 - Tutorial extended: Map Alignment, Feature Linking and IDMapper [merged]

👤 Suggested Assignees

• timosachsenberg
• cbielow
• martingarridorc
• iuliia-nikolskaia
• vdaevoloop

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord or schedule a call!

[jpfeuffer]

Hi!
Yes, the API changed. We introduced a list wrapper for PeptideIdentifications that should speed up a lot of applications because you do not need to copy it into a python list for most of the cases.

You can always check the most recent API in our API docs: https://pyopenms.readthedocs.io/en/latest/apidocs/_autosummary/pyopenms/pyopenms.IdXMLFile.html#pyopenms.IdXMLFile

you need to change
pep_ids = [] to pep_ids = oms.PeptideIdentificationList[]

Expect more changes downsrtream in your application because of this. Or fix pyopenms to 3.4.0.

@timosachsenberg I hope this was expected. If not, one should provide more backwards compatible interface in the future.

[jpfeuffer]

By the way, does your IDE not show the type hints we generate for pyopenms??

[yueqixuan]

@jpfeuffer Thank you very much for your reply. After switching to PeptideIdentificationList, it works as expected. For now, to maintain compatibility for users, I plan to pin the pyopenms version to <= 3.4.0.

In my IDE (VS Code with Pylance), I see the following:

(method) def load(
    filename: bytes | str | Any,
    protein_ids: List[ProteinIdentification],
    peptide_ids: Any
) -> None

Cython signature: void load(String filename, libcpp_vector[ProteinIdentification] & protein_ids, PeptideIdentificationList & peptide_ids) Loads the identifications of an idXML file without identifier

However, when I inspect the pyopenms source code directly, everything looks correct. It may be an issue with VS Code / Pylance not fully picking up the generated type hints, or possibly related to caching in my local environment.

Thanks again for your help.

[jpfeuffer]

Thanks for reporting back. The type hints indeed look strange. I'll double check to be sure.

[jpfeuffer]

Thanks for reporting back. The type hints indeed look strange. I'll double check to be sure.

[jpfeuffer]

Thanks for reporting back. The type hints indeed look strange. I'll double check to be sure.

[jpfeuffer]

Actually now I am pretty sure this was not intended since the function definition is duplicated. I assume someone wanted to have the old interface in one of those cases:

https://github.com/OpenMS/OpenMS/blob/0044887f67ab18b5f6ce014cf3b30bfd15548a28/src/pyOpenMS/pxds/IdXMLFile.pxd#L14

@timosachsenberg

[timosachsenberg]

sorry about this inconvinience. The python API is still evolving but changes required are usually pretty small.
In the future we will require the PeptideIdentificationList. That's why I think switching to it could make sense.
@jpfeuffer is there a simple way to keep a backwards compatible interface? e.g. with a custom conversion provider?

[jpfeuffer]

@timosachsenberg if the underlying function for vector in c++ is gone, you can only add an addon function in pyx that converts before. But you must do it for every occurrence of PeptideIdentificationList.
There is currently no way to produce multiple python functions from a single declaration (e.g. one for list and one for the peptidelist object).

In any case the duplicated declaration in Idxmlfile must go. I think it is the reason for wrong type hints.