Irf maker and cut optimiser #2473
Conversation
ctapipe/irf/optimise.py
Outdated
| def __init__(self, gh_cuts=None, offset_lim=None): | ||
| self.gh_cuts = gh_cuts | ||
| if gh_cuts: | ||
| self.gh_cuts.meta["extname"] = "GH_CUTS" |
There was a problem hiding this comment.
extname is the name of a FITS extension. You should't set this in the metadata of the table, but only when writing out, it's the name argument of write.
There was a problem hiding this comment.
I got
TypeError: write_table_fits() got an unexpected keyword argument 'name'
when I tried self.gh_cuts.write(out_name, name="GH_CUTS", format="fits", overwrite=overwrite) and I don't find anything about a name argument to table.write, do you mean I should manage the fits file more directly and just cast the tables to hdus?
LukasBeiske
force-pushed
the
irf-maker
branch
from
643a841 to
7c2efc7
Compare
LukasBeiske
force-pushed
the
irf-maker
branch
from
d8f1268 to
b127e30
Compare
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
… possibility to save an OptimizationResult with only gh cuts for now.
Analysis Details9 Issues
Coverage and Duplications
Project ID: cta-observatory_ctapipe_AY52EYhuvuGcMFidNyUs |
| ).tag(config=True) | ||
|
|
||
| energy_migration_linear_bins = Bool( | ||
| help="Bin energy migration ratio using linear bins", |
There was a problem hiding this comment.
Would be useful to mention here what is the default if this is set to false. Looking at the code, it is geometric (logarithmic) bins. This could also be clearer as:
energy_migration_binning = CaselessStrEnum(
['linear', 'logarithmic'],
help="How energy bins are distributed between energy_migration_min and energy_migration_max",
default_value="logarithmic"
).tag(config=True)| ).tag(config=True) | ||
|
|
||
| energy_migration_n_bins = Integer( | ||
| help="Number of bins in log scale for energy migration ratio", |
There was a problem hiding this comment.
log scale is incorrect if the user chooses energy_migration_linear_bins=True
|
|
||
| @abstractmethod | ||
| def make_edisp_hdu( | ||
| self, events: QTable, point_like: bool, extname: str = "ENERGY MIGRATION" |
There was a problem hiding this comment.
general comment for here and other places: since we have options point_like and source_is_pointlike, this is quite confusing. Maybe rename point_like to direction_cut_applied or spatial_selection_applied. That will make it much easier to follow. Also, this is more in line with the fact that we don't use the term "point-like IRFs" anymore, and instead use "SpatialSelection=None | RAD_MAX"
There was a problem hiding this comment.
I will change it to spatial_selection_applied. I guess we should also change it accordingly in pyirf then, right?
| "Do not compute background rate.", | ||
| ), | ||
| **flag( | ||
| "point-like", |
There was a problem hiding this comment.
see earlier comment: "point-like/full-enclosure" terminology should be replaced to avoid confusion between point-like cuts vs point-like simulated sources. The flag could be spatial_selection_applied or similar
| super().__init__(parent=parent, **kwargs) | ||
|
|
||
| @abstractmethod | ||
| def make_bias_resolution_hdu( |
There was a problem hiding this comment.
Any reason to not just use __call__ like other Components? (same comment for all Components in this file). That would make the API similar to all other parts of ctapipe
There was a problem hiding this comment.
No reason. I think __call__ would be better 👍 I will also change the benchmark and cut optimization components accordingly.
| self, | ||
| signal_events: QTable, | ||
| background_events: QTable, | ||
| theta_cut: QTable, |
There was a problem hiding this comment.
maybe to harmonize terminology, use spatial_selection_table?
There was a problem hiding this comment.
Then we should probably not use "theta cut" anywhere anymore, right? So also rename it in the cut calculation components etc.
| __all__ = ["EventLoader", "EventPreProcessor"] | ||
|
|
||
|
|
||
| class EventPreProcessor(QualityQuery): |
There was a problem hiding this comment.
preprocess is a word, so this should be EventPreprocessor,
Also generally we do not inherit from QualityQuery for in other components, but rather have the component contain a QualityQuery. That makes more semantic sense, as this class does much more than a QualityQuery.
Then you would have a config file much like other classes in ctapipe:
EventPreprocessor:
EventQualityQuery:
quality_critieria:
...
energy_reconstructor: ...| from enum import Enum | ||
|
|
||
| import astropy.units as u | ||
| from pyirf.spectral import CRAB_HEGRA, IRFDOC_ELECTRON_SPECTRUM, IRFDOC_PROTON_SPECTRUM |
There was a problem hiding this comment.
wasn't this moved to a new package? Or is that just handled inside pyirf?
| target_spectrum=self.gamma_target_spectrum, | ||
| ) | ||
| ] | ||
| if self.optimization_algorithm != "PercentileCuts": |
There was a problem hiding this comment.
Might be safer to use isinstance() here and elsewhere instead of a string compare to the class name, so that if we ever re-factor or rename classes, it gets updated correctly. But that would mean importing those sub-classes or the module, so not clear if that makes it messier.
| if include_bkg: | ||
| assert isinstance(hdul["BACKGROUND"], fits.BinTableHDU) | ||
|
|
||
| os.remove(output_path) # Delete output file |
There was a problem hiding this comment.
use output_path.unlink() instead, so no need to import os.
Thanks! I can also add an example config file (maybe containing all config options and their default values?) at |
As we are planning to produce IRFs on events that were not used to select the GH cut, it makes sense to split the procedure generating IRFs into two tools. Because this splitting results in a large restructuring of the the previous PR on generating IRFs (#2315) that was basically working properly, I open a new PR to preserve the discussion on the old attempt and let its more or less functional code serve as easy reference.