Test_merfish.py takes forever to run

[suestring7]

Hi there,

I installed MERlin following the instructions in the documentation and implemented some workarounds:

1. I rolled back certain packages' versions to resolve some pytest errors.
2. I modified the pandas concat() by adding "axis=".
3. I enabled long path support in Windows to make the path exceeding the 260 character limits.

However, I am still encountering an issue with test_merfish.py running extremely slow (over 16 hours and not finished yet). I came across a previous issue posted by @HazenBabcock, where it seems that his computer took less than 2 minutes to complete all the tests, including passing the test_merfish.

Is this extended test run time expected? Should I continue waiting, or do you have any suggestions on how to monitor the progress and identify the slow-running processes?

Additionally, could I request information on your current working environment's package versions in case it's still an environment-related issue?

PS: My processor information

Caption                              DeviceID  MaxClockSpeed  Name                                            NumberOfCores  Status
AMD64 Family 23 Model 49 Stepping 0  CPU0      3800           AMD Ryzen Threadripper 3960X 24-Core Processor  24             OK

Best,
Yuan

[HazenBabcock]

I'm pretty sure that test_merfish.py is not expected to take more than a few minutes so it sounds like something has gone wrong. What that might be however is hard to say, as usually when something goes wrong MERlin exits with an error message. Do you have a small dataset that you could try to run MERlin on to verify that it isn't some issue with the test environment?

[suestring7]

I'm pretty sure that test_merfish.py is not expected to take more than a few minutes so it sounds like something has gone wrong. What that might be however is hard to say, as usually when something goes wrong MERlin exits with an error message. Do you have a small dataset that you could try to run MERlin on to verify that it isn't some issue with the test environment?

Thank you for the quick reply. I am planning to start with the public dataset from https://zhuang.harvard.edu/merfish.html . However, I found them slightly different from the required inputs which may need some extra effort to sort the data. Thus we plan to run the test first and then move on to try the dataset.