load_all function gives error when trying to run qc

[arjunsridhar12345]

When trying to run qc / load all of the data streams, get the following error:

File c:\Users\arjun.sridhar\source\repos\dynamic-foraging-processing\.venv\Lib\site-packages\aind_behavior_dynamic_foraging\data_qc\suite.py:32, in make_qc_runner(dataset)
     30 def make_qc_runner(dataset: contract.Dataset) -> qc.Runner:
     31     _runner = qc.Runner()
---> [32](file:///C:/Users/arjun.sridhar/source/repos/dynamic-foraging-processing/.venv/Lib/site-packages/aind_behavior_dynamic_foraging/data_qc/suite.py:32)     dataset.load_all(strict=False)
     33     exclude: list[contract.DataStream] = []
     34     rig: AindDynamicForagingRig = dataset["Behavior"]["InputSchemas"]["Rig"].data

File c:\Users\arjun.sridhar\source\repos\dynamic-foraging-processing\.venv\Lib\site-packages\contraqctor\contract\base.py:424, in DataStream.load_all(self, strict)
    422 if stream is None:
    423     continue
--> [424](file:///C:/Users/arjun.sridhar/source/repos/dynamic-foraging-processing/.venv/Lib/site-packages/contraqctor/contract/base.py:424) stream.load_all(strict=strict)
    425 if stream.has_error and strict:
    426     cast(_typing.ErrorOnLoad, stream.data).raise_from_error()

File c:\Users\arjun.sridhar\source\repos\dynamic-foraging-processing\.venv\Lib\site-packages\contraqctor\contract\base.py:424, in DataStream.load_all(self, strict)
    422 if stream is None:
    423     continue
--> [424](file:///C:/Users/arjun.sridhar/source/repos/dynamic-foraging-processing/.venv/Lib/site-packages/contraqctor/contract/base.py:424) stream.load_all(strict=strict)
...
--> [551](file:///C:/Users/arjun.sridhar/source/repos/dynamic-foraging-processing/.venv/Lib/site-packages/contraqctor/contract/base.py:551)     raise ValueError("Data must be a list of DataStreams.")
    552 self._update_data_stream_mapping()
    553 return self

ValueError: Data must be a list of DataStreams.

Here's an example session (this is uploaded to code ocean), but seem to get the same error running locally and in code ocean: https://codeocean.allenneuraldynamics.org/data-assets/89e9efd0-c14f-4417-bdc1-fd76d1cad797/behavior_845023_2026-06-17_19-41-02

Very uneducated guess is that there might be some streams defined in the contract that are not present in the current datasets (mainly the HarpLickometerLeft and HarpLickometerRight) but not really sure if that would cause this error.

I'm on contraqctor version 0.5.8, but could be some dependency thing I have wrong.

[bruno-f-cruz]

Can you add reference to the code you are running?

[arjunsridhar12345]

yeah I was trying this:

from aind_behavior_dynamic_foraging.data_qc.suite import make_qc_runner
from aind_behavior_dynamic_foraging.data_contract import dataset
from pathlib import Path

df_dataset = dataset(Path("/data/behavior_845023_2026-06-17_19-41-02"))
make_qc_runner(df_dataset)

[bruno-f-cruz]

Ok there are a couple of issues that I need to fix upstream that I didn't see before.

1. The error you are getting is solved by this Honor strict=False when a collection's own load fails contraqctor#107
2. I think you will have other errors when making the test suites tho. I gotta move all the IO logic to the setup_suite and teardown_suite methods so they crash on the runner context, not at instantiation time.

I will update this issue once i have a solution

[bruno-f-cruz]

Ok I have made updates to the upstream package and fixed some edge cases on the qc in this package. That being said, the contraqctor is not really aware of the notion of "optional devices". That logic, if important, will have to be defined in this qc script. Here is the qc report for this session in case you need a reference. There are some tests here we should def go over since they indicate some stuff at the rig is not correctly connected or we have some bugs in the acquisition code.

here.html

specifically:

• Stage is not responding to all messages
• Stage doesn't seem to be aligned to the clock synch
• EndSession file doesn't exist

[arjunsridhar12345]

thanks for taking a look. tagging @alexpiet and @micahwoodard for visibility.