search

EdinburghGenomics / smrtino

Run pipeline for PacBio Sequel SMRT cells
Other
0 stars 0 forks source link

Detect and deal with test runs that the Sequel generates #4

Open tbooth opened 1 year ago

tbooth commented 1 year ago

Test runs are synthetic runs where the system is run dry - there is no actual cell present. Test cells are pretty easy to spot from looking at the XML. But possibly the pipeline will see the run before it sees the cell info, so we can't always immediately see if a run is a "test run".

We can probably assume that all these runs will have a single cell, but I don't think I need to. Maybe just regard any run where all cells are test cells as a "test run" and close the ticket automatically when the final report is made. The interim e-mails will clearly say that the cell is a test cell.

Once we are producing a single report per cell then I think this logic will be a tad cleaner, so address that (#2) first.

tbooth commented 1 year ago

Example of such a run - r64175e_20230220_093234

tbooth commented 1 year ago

Added an is_testrun.sh script that currently looks for InstrumentId="Inst1234" in run.metadata.xml (but could easily grep for various other strings).

And a testrun state that is analogous to the aborted state.

TODO

1) need to ensure that runs left in the testrun state are cleaned up by the data deletion scripts. 1) need to ensure that reports on egcloud are cleaned up (both ones on there now, and also no more being added)

tbooth commented 1 year ago

Reports are cleaned up, but I still need to fix and check the deletion scripts.