Refactor#2
Conversation
|
Can you elaborate on exactly what testing you've done? You must have supplied the --skip-gracedb-upload option, correct? For instance, your call to writeMessage on is not defined. Instead, writeLogMessage is defined, but does not actually return anything. I hesitate to merge anything into the production repo with incomplete development like this. Also, can you please specify which GraceId you used for testing? |
|
I modified the ini file to not point to directories I did not have access to, then this command was run: Twice. Once unmodified and once with the segdb route manually forced. Your note on the return value is absolutely correct (there will be another commit coming to fix that). I should note that this PR will be updated as time goes on to address issues, and when you feel comfortable merging, then we can merge. I wanted substantive feedback on the way the change was happening currently. I'm happy to do more thorough tests on a live gracedb instance, I just need a good test event to do so. |
|
You can (and should) test to your heart's desire using gracedb-test. For that, you'll need to modify the "gracedb_url" option in the config file to point to https://gracedb-test.ligo.org/api/ and pick any GraceId from there. Note, the dmt files are cleaned up periodically at CIT so you can't use an ancient event. Hopefully I'll be able to test this a bit myself in the next few weeks and give you more concrete development suggestions. |
I've refactored the
ligolw_segment_query_segdbcall to use the python API. I've also written out the call toligolw_segment_query_segdb --dmtto use a simplified file location and retrieval setup. A brief test shows that both work for a test gracedb event. Please review for clarity, efficiency, and any subtler bugs. I have not tested that the logging to gracedb is identical to the previous incarnation.