Flat Field coefficients computation + the rise and decay times #221
Conversation
…requered changes in containers files. And the script to compute raise and decay times. Cleaned branch.
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #221 +/- ##
==========================================
+ Coverage 51.79% 51.82% +0.03%
==========================================
Files 78 78
Lines 6505 6510 +5
==========================================
+ Hits 3369 3374 +5
Misses 3136 3136 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
| (self.gainLG * mask, gainLG_err, gainLG_err) | ||
| ).T | ||
| self._results.is_valid = mask | ||
| log.info("Trying to write charges") |
There was a problem hiding this comment.
Maybe we should make it clear everywhere that these 2 new fields are for the high-gain channel ? E.g. by renaming them charge_hg and charge_hg_std everywhere ?
| ] | ||
|
|
||
| print("[DEBUG] Running command:", " ".join(cmd)) | ||
| subprocess.run(cmd, check=True) |
There was a problem hiding this comment.
We may try to find a more pythonic way, e.g. by invoking directly the API found in the script gain_PhotoStat_computation.py ?
| print(f"[INFO] {filename_ps} not found, running gain_PhotoStat_computation.py...") | ||
|
|
||
| gain_script = os.path.expanduser( | ||
| "~/local/src/nectarchain/src/nectarchain/" |
There was a problem hiding this comment.
We should find another way, and avoid using such an absolute path.
| else: | ||
| print(f"[INFO] File {run_path} already exists, skipping computation.") | ||
|
|
||
| print(f"[INFO] Starting analysis on {filename_ps}") |
There was a problem hiding this comment.
print -> logging, to be implemented.
|
|
||
|
|
||
| def optimize_with_outlier_rejection(sigma, data): | ||
| def define_delete_out(sigma, data): |
|
|
||
| # Define the least-squares function | ||
| def LSQ_wrap(a0, a1, a2, a3): | ||
| a4 = a3 # changed |
There was a problem hiding this comment.
It may be good to add a comment here, explicitly stating that a4 = a3 means that we assume a symmetric 2D-Gaussian (is that correct, by the way ? Seems so, looking at https://ctapipe.readthedocs.io/en/latest/api/ctapipe.image.toymodel.Gaussian.html#ctapipe.image.toymodel.Gaussian)
| print(f"Max residual: {max_residual*100:.2f}%") | ||
|
|
||
| dict_missing_pix["rejected_outliers"] = ( | ||
| 1855 |
There was a problem hiding this comment.
Let's use ctapipe_io_nectarcam.constants.N_PIXELS instead.
|
|
||
| # main | ||
| if __name__ == "__main__": | ||
| camera = CameraGeometry.from_name("NectarCam-003").transform_to( |
There was a problem hiding this comment.
Let's discover the camera geometry from the data, e.g. using EventSource.subarray.tel[0].camera.geometry instead.
| ) | ||
| low_gain = Field(type=np.ndarray, dtype=np.float64, ndim=2, description="low gain") | ||
| pixels_id = Field(type=np.ndarray, dtype=np.uint16, ndim=1, description="pixel ids") | ||
| charge = Field(type=np.ndarray, dtype=np.float64, ndim=1, description="charge") |
There was a problem hiding this comment.
I would suggest to create a PhotostatContainer to let the gain container in a high abstraction level, the charge and charge_std fileds are not needed in the SPEfitContainer which inherits from the gaincontainer
| high_gain=np.zeros((self.npixels, 3)), | ||
| low_gain=np.zeros((self.npixels, 3)), | ||
| pixels_id=self._pixels_id, | ||
| charge=np.zeros((self.npixels, 1)), |
There was a problem hiding this comment.
and here return the PhotoStatContainer
|
Hi @anmikhno, thanks a lot for your work, I just quickly let a comment about the abstraction of |
| # ctapipe modules | ||
| from ctapipe.visualization import CameraDisplay | ||
| from iminuit import Minuit | ||
| from jacobi import propagate |
There was a problem hiding this comment.
This creates a new dependency for nectarchain. Is it strictly required ?
There was a problem hiding this comment.
If I understand scipy/scipy#17059 correctly, this has been merged into scipy directly since version 1.15.0. A recent installation of nectarchain comes with at least scipy v1.16.0. Could we try to use the implementation from scipy instead of adding another dependency ?
…requered changes in containers files. And the script to compute raise and decay times. Cleaned branch.
3 participants
The flat field coefficients are computed using the Gaussian fit, on top pf that the outliers are removed to improve the fit. The script generated pdf with the plots and a table with flat-field coefficients computed per pixel using the fit (model method) and flat field coefficients computed using the simple formula (n_pe_per_pix/mean(n_pe_over_camera))
Cleaned branch.