fractal-analytics-platform / fractal-tasks-core

Main tasks for the Fractal analytics platform
https://fractal-analytics-platform.github.io/fractal-tasks-core/
BSD 3-Clause "New" or "Revised" License
14 stars 6 forks source link

Get gridless test set to pass #109

Closed jluethi closed 2 years ago

jluethi commented 2 years ago

When trying to run the gridless test case UZH path: /data/active/fractal/Liberali/FractalTesting20220124/gridless_Yokogawa_recording-FMI/20220316_sec_FOCM_test-R1_E2/day3/TIF (number 4 of search first here: https://github.com/fractal-analytics-platform/fractal/issues/213), the metadata parsing fails with the error message: Exception: When parsing the Yokogawa mlf file, some sites had varying step size in Z. That is not supported for the OME-Zarr parsing

Let's check whether this is really the problem or whether something in the metadata parsing is failing for this file. Could be an issue when the metadata was created for this test set, could be a parsing problem.

jluethi commented 2 years ago

Digging into this, the problem again is float precision. In this case, the steps are of size 0.6 um. Thus, represented as floats. That means that sometimes they are 0.5999999999999979, sometimes they are 0.6000000000000014. I'll need to adapt the parsing to be robust against this.

jluethi commented 2 years ago

Test again now that https://github.com/fractal-analytics-platform/fractal-tasks-core/pull/136 fixes the precision issue.

jluethi commented 2 years ago

With the metadata fix, the gridless search first now works as expected! 🎉 See 3f5c8b884683e881a6b07e058ef3b03348f9772e for the example

Bildschirmfoto 2022-10-14 um 19 36 15

Overexposed (same image, but shows borders of FOVs):

Bildschirmfoto 2022-10-14 um 19 33 34