ebi-ait / ingest-graph-validator

HCA Ingest Service Graph Validation Suite
MIT License
1 stars 0 forks source link

False errors by 10x_has_more_than_2_files.adoc #13

Closed rays22 closed 3 years ago

rays22 commented 4 years ago

Describe the bug The rule has to take into account both the lane and the process id. i.e. there should be 2-4 files per lane per process, otherwise it produces false errors. Also, give a more informative error message than an array of num_files.

To Reproduce Steps to reproduce the behavior:

  1. Use submission 85e72912-9f91-4489-8169-3b43cc65a16a. https://github.com/HumanCellAtlas/hca-data-wrangling/issues/356
  2. Hydrate and run a validation test.
  3. Check the validation test output for errors.
  4. See error

    [ingest_graph_validator.actions.test_action] - ERROR: test [10x_has_more_than_2_files.adoc] failed: non-empty result. [ingest_graph_validator.actions.test_action] - ERROR: result: [{'num_files': 6}, {'num_files': 6}, {'num_files': 6}, {'num_files': 6}, {'num_files': 6}, {'num_files': 6}, {'num_files': 6}, {'num_files': 6}, {'num_files': 6}, {'num_files': 6}, {'num_files': 6}, {'num_files': 6}, ...

Expected behavior No errors for dataset 85e72912-9f91-4489-8169-3b43cc65a16a.

Screenshots NA Desktop (please complete the following information):

mshadbolt commented 3 years ago

I think this needs to account for the lane index in the test in order to be more accurate. As in

ESapenaVentura commented 3 years ago

This is solved by changing the test to check for only 2+ files, and expecting if we get more than expected (e.g. 5) will be captured by "ensure_lane_index"