Split aqmon hits issue #1061
Split aqmon hits issue #1061
Conversation
There was a problem hiding this comment.
Thanks Robin, I think this is exactly what we need. What might help is moving the AqmonDtype class to the AqmonHits plugin - where you can put it in a function infer_dtype-function which is how we usually handle logic like this in strax(en). Then, you can re-use self.dtype and self.dtype.names (to equal the wanted_fields).
|
Hi @JoranAngevaare, |
| # in busy runs veto starts get split at chunk edges into two. we introduce hitlets: | ||
| aqmon_hitlets = strax.create_hitlets_from_hits(aqmon_hits, (0,0), [0, 1000]) | ||
| to_pe = np.ones(808) # stay in ADC units: these are NIM signals | ||
| aqmon_hitlets = strax.get_hitlets_data(aqmon_hitlets, records, to_pe=to_pe) |
There was a problem hiding this comment.
This is where the current tests are failing. It should actually work like this. When strax.get_hitlets_data is called it lastly complains at https://github.com/AxFoundation/strax/blob/156254287c2037876a7040460b3551d590bf5589/strax/processing/hitlets.py#L214 . My hunch is that we pass something that look's a little different to what https://github.com/AxFoundation/strax/blob/156254287c2037876a7040460b3551d590bf5589/strax/processing/hitlets.py#L222 usually sees and the cached compiled function doesn't like it and seg faults. Any idea on how to fix this? If we delete the cache, we might have the same problem the other way around, no? Maybe we find out what's the difference it doesn't like and provide what it likes.
There was a problem hiding this comment.
We're back to this issue now:
[...]
Fatal Python error: Segmentation fault
Current thread 0x00007fd8cd53f700 (most recent call first):
File "/opt/hostedtoolcache/Python/3.8.13/x64/lib/python3.8/site-packages/strax/processing/hitlets.py", line 214 in get_hitlets_data
File "/home/runner/work///***/plugins/acqmon_processing.py", line 117 in compute
File "/opt/hostedtoolcache/Python/3.8.13/x64/lib/python3.8/site-packages/strax/plugin.py", line 596 in do_compute
File "/opt/hostedtoolcache/Python/3.8.13/x64/lib/python3.8/site-packages/strax/plugin.py", line 488 in iter
File "/opt/hostedtoolcache/Python/3.8.13/x64/lib/python3.8/site-packages/strax/mailbox.py", line 281 in _send_from
File "/opt/hostedtoolcache/Python/3.8.13/x64/lib/python3.8/threading.py", line 870 in run
File "/opt/hostedtoolcache/Python/3.8.13/x64/lib/python3.8/threading.py", line 932 in _bootstrap_inner
File "/opt/hostedtoolcache/Python/3.8.13/x64/lib/python3.8/threading.py", line 890 in _bootstrap
[...]
/home/runner/work/_temp/a65694df-fcc9-4870-821e-663ce8843c36.sh: line 1: 2489 Segmentation fault (core dumped) coverage run --source=*** -m pytest --durations 0
tests/test_daq_reader.py
Error: Process completed with exit code 139.
straxen/plugins/acqmon_processing.py
Outdated
| @staticmethod | ||
| def fake_hit(start, dt=1, length=1): | ||
| hit = np.zeros(1, strax.hit_dtype) | ||
| aqmon_hit_dtype = strax.unpack_dtype(self.deps['aqmon_hits'].dtype_for('aqmon_hits')) | ||
| hit = np.zeros(1, aqmon_hit_dtype) |
There was a problem hiding this comment.
I think you need to add self and remove the static method
|
@RoBGlaBe did you manage to fix it? |
What does the code in this PR do / what does it improve?
Counting veto starts correctly instead of double counting some.
Can you briefly describe how it works?
Veto start signals were counted twice if they started in one data package (for the lack of a better word of "raw_record level chunks") and ended in another due to the usage of hits. Creating hitlets in the same Plugin solves this problem.
Can you give a minimal working example (or illustrate with a figure)?
See note: https://xe1t-wiki.lngs.infn.it/doku.php?id=xenon:xenonnt:robingb:aqmon_hit_splitting
Please include the following if applicable:
veto_intervalsreports an error for039466#979