LST1Analysis MVA

From my_wiki
Revision as of 12:10, 15 October 2021 by Monicava (talk | contribs) (→‎LST1 data formats)
Jump to: navigation, search

Location of LST1 data at PIC

Use CTA observatory LDAP account (eg: monica.vazquez)

Location of LST1 data at La Palma

  • Directory: /fefs/onsite/data
    Use cp01/cp02 to copy out data, never the login machine

LST1 DL2/DL3 Data

  • Crab DL3
    Location: /fefs/aswg/workspace/daniel.morcuende/data/real/DL3/20201120/v0.7.3/tailcut84/
    • dl3_LST-1.RunXXX.fits.gz
    • hdu-index.fits.gz
    • obs-index.fits.gz
  • Crab DL2 with dynamic cleaning & MC (tuned NSB & PSF - ICRC version)
    Location: /fefs/aswg/workspace/abelardo.moralejo/Crab_tuned_MC_20210924/20201120/merged/DL2/
    • dl2_LST-1.Run029*h5
    • dl2_merged.h5

DL2toDL3 conversion

General talk on the LST1 DL3 Tools by Chaitanya Priyadarshi: DL3_Tools_and_Spectral_analysis_of_Crab_data_with_gammapy.pdf

DL2toDL3 Tools

Tools are located in lstchain/tools:

  • lstchain_create_irf_file: create IRFs from the MC file
    • The IRFs can be generated using only the gamma MC file, to produce Effective Area, Energy Dispersion
      (and PSF for Full Enclosure IRFs, ie. without any theta cuts).
    • Proton and Electron MC are used to create the Background IRF.
    • For creating point-like IRFs, you have to pass
      the --point-like flag in the command-line, as by default the Tools generates a Full Enclosure IRF.
  • lstchain_create_dl3_file: create DL3 files
  • lstchain_create_dl3_index_files: generate HDU and OBS Index files of the DL3 files created.
    The DL3 files are indexed as per GADF and using ctapipe tools

Command options can be checked: lstchain_create_irf_file --help-all

DL2 Cuts

Fixed cut variables & default values (see here)

  • fixed_gh_cut [0.6]
  • fixed_theta_cut [0.2]
  • allowed_tels [1]

Cut Methods:

  1. --fixed-gh-cut 0.9
  2. --config /path/to/config.json
    Example json config files for the DL3 tools are here
    The DL3 JSON config file (dl3_tool_config.json) is a subset of the IRF config file (irf_tool_config.json).
    One can use the IRF config file for both tools to ensure that the same filter cuts for both MC and real data are used.

NOTE: the fixed_theta_cut option is not available in lstchain_create_dl3_file, so it is better to use the JSON option to be consistent with the IRF generation

Example DL2toDL3 commands

lstchain_create_irf_files
        -g /path/to/DL2_MC_gamma_file.h5
        -o /path/to/irf.fits.gz
        --point-like (Only for point_like IRFs)
        --fixed-gh-cut 0.9
        --fixed-theta-cut 0.2
        --irf-obs-time 50

lstchain_create_dl3_file
        -d /path/to/DL2_data_file.h5
        -o /path/to/DL3/file/
        --input-irf /path/to/irf.fits.gz
        --source-name Crab
        --source-ra 83.633deg
        --source-dec 22.01deg
        --fixed-gh-cut 0.9

lstchain_create_dl3_index_files
        -d /path/to/DL3/files/
        -o /path/to/DL3/index/files
        -p dl3*[run_1-run_n]*.fits.gz

Example DL2toDL3 commands - MVA specific

lstchain_create_irf_files -g /Users/monicava/LST1_DATA/Abelardo/dl2_merged.h5 -o /Users/monicava/LST1_DATA/Abelardo/irf.gh01.fits.gz --point-like --config irf_tool_config_gh01.json --irf-obs-time 50
lstchain_create_dl3_file -d /Users/monicava/LST1_DATA/Abelardo/dl2_LST-1.Run02967.h5 -o /Users/monicava/LST1_DATA/Abelardo/DL3_gh01 --input-irf /Users/monicava/LST1_DATA/Abelardo/irf.gh01.fits.gz --source-name Crab --source-ra 83.633deg --source-dec 22.01deg --config irf_tool_config_gh01.json
lstchain_create_dl3_index_files -d /Users/monicava/LST1_DATA/Abelardo/DL3_gh01

Run summary info/data-check

LST1 data formats

File content can be obtained as follows: 

from lstchain.io import get_dataset_keys
from lstchain.io.io import dl1_params_lstcam_key, dl2_params_lstcam_key, dl1_images_lstcam_key
dl2_file = "/fefs/aswg/data/real/DL2/20210903/v0.7.3/tailcut84/dl2_LST-1.Run05998.0000.h5"
import pandas as pd
get_dataset_keys(dl2_file)

Out[]: 
['configuration/instrument/subarray/layout',
 'configuration/instrument/subarray/layout.__table_column_meta__',
 'configuration/instrument/telescope/camera/geometry_LSTCam',
 'configuration/instrument/telescope/camera/geometry_LSTCam.__table_column_meta__',
 'configuration/instrument/telescope/camera/readout_LSTCam',
 'configuration/instrument/telescope/camera/readout_LSTCam.__table_column_meta__',
 'configuration/instrument/telescope/optics',
 'configuration/instrument/telescope/optics.__table_column_meta__',
 'dl1/event/telescope/monitoring/calibration',
 'dl1/event/telescope/monitoring/flatfield',
 'dl1/event/telescope/monitoring/pedestal',
 'dl2/event/telescope/parameters/LST_LSTCam']

dl1_file = "/fefs/aswg/data/real/DL1/20210903/v0.7.3/tailcut84/dl1_LST-1.Run05998.0000.h5"
get_dataset_keys(dl1_file)

Out[]: 
['configuration/instrument/subarray/layout',
 'configuration/instrument/subarray/layout.__table_column_meta__',
 'configuration/instrument/telescope/camera/geometry_LSTCam',
 'configuration/instrument/telescope/camera/geometry_LSTCam.__table_column_meta__',
 'configuration/instrument/telescope/camera/readout_LSTCam',
 'configuration/instrument/telescope/camera/readout_LSTCam.__table_column_meta__',
 'configuration/instrument/telescope/optics',
 'configuration/instrument/telescope/optics.__table_column_meta__',
 'dl1/event/telescope/image/LST_LSTCam',
 'dl1/event/telescope/monitoring/calibration',
 'dl1/event/telescope/monitoring/flatfield',
 'dl1/event/telescope/monitoring/pedestal',
 'dl1/event/telescope/parameters/LST_LSTCam']

Data is in the tables:

  • 'dl1/event/telescope/image/LST_LSTCam' (camera calibrated images)
  • 'dl1/event/telescope/parameters/LST_LSTCam' (Hillas parameters, timestamps, pointing, etc)

In case of DL2 only the table parameters 'dl2/event/telescope/parameters/LST_LSTCam' are available, which is a copy of the DL1 table with the reconstructed parameters added

To access the parameter tables in DL1 and DL2: 

dl2_dataframe = pd.read_hdf(file, key=dl2_params_lstcam_key)
dl2_dataframe

Out[]: 
       obs_id  event_id      intensity  log_intensity         x         y         r       phi    length  ...  reco_energy  reco_disp_dx  reco_disp_dy  reco_src_x  reco_src_y  reco_alt   reco_az  reco_type  gammaness
0        5998         1    5529.521235       3.742688 -0.032178 -0.020007  0.037890 -2.585328  1.168153  ...    12.028581     -1.395368     -0.322723   -1.427546   -0.342730  1.084904  0.228667        101   0.025000
1        5998         4      61.694788       1.790248  0.093973  0.365420  0.377309  1.319085  0.045765  ...     0.021868      0.033724      0.019880    0.127698    0.385300  1.140388  0.287830        101   0.056333
2        5998         5     131.579536       2.119188  0.242024  0.147060  0.283200  0.546006  0.123974  ...     0.114741      0.246228     -0.281001    0.488253   -0.133941  1.153445  0.243046        101   0.188333
3        5998         6      37.598133       1.575166  0.405366 -0.546641  0.680542 -0.932721  0.072361  ...     0.062491      0.011559     -0.049816    0.416924   -0.596457  1.150419  0.202634        101   0.290833
4        5998         7      75.891508       1.880193 -0.129859 -0.524562  0.540397 -1.813474  0.078325  ...     0.033458      0.010284      0.027173   -0.119575   -0.497389  1.131426  0.213075        101   0.165000
...       ...       ...            ...            ...       ...       ...       ...       ...       ...  ...          ...           ...           ...         ...         ...       ...       ...        ...        ...
28271    5998     52989      89.536526       1.952000 -0.775412  0.466715  0.905034  2.599783  0.061538  ...     0.043203     -0.004094      0.009934   -0.779506    0.476649  1.108024  0.292202        101   0.165000
28272    5998     52990     113.433699       2.054742 -0.015224 -0.558476  0.558683 -1.598049  0.072026  ...     0.034507     -0.150657      0.088593   -0.165881   -0.469883  1.129934  0.214743        101   0.091000
28273    5998     52991    3053.237697       3.484761  0.621141  0.404424  0.741198  0.577148  0.319016  ...     0.218918      0.064154     -0.829353    0.685294   -0.424929  1.160371  0.216046        101   0.010000
28274    5998     52992  132592.328588       5.122518  0.000555  0.000571  0.000796  0.799391  1.163453  ...    30.287504     -0.286239      0.399654   -0.285684    0.400225  1.125740  0.287283        101   0.178036
28275    5998     52997      74.290200       1.870932 -0.595926  0.781502  0.982788  2.222275  0.097937  ...     0.292780      0.711934      0.310702    0.116007    1.092203  1.138651  0.347324        101   0.442667

[28276 rows x 51 columns]

Accesing images is more complicated and requires the tables module 

from astropy.table import Table
images = Table.read(dl1_file, path=dl1_images_lstcam_key)['image']
images

Out[]: 
<Column name='image' dtype='float32' shape=(1855,) length=53000>
    8.701056 .. 7.1084614
   3.7857776 .. 1.2068052
   3.8285189 .. 0.9846253
   4.5408783 .. 5.1921587
     4.754586 .. 1.026284
  0.7226335 .. 0.09590531
   3.4738295 .. 2.0677526
 2.8597107 .. -0.43177217
 -0.20343336 .. 3.8313065
   1.3372487 .. 1.5678478
    3.2301373 .. 2.428795
   3.4438455 .. 4.2895527
    3.871757 .. 1.4984165
  -0.5311186 .. 1.8316865
    2.717239 .. 4.3192797
  1.3362556 .. 0.48472026
 0.52366954 .. 0.26254028
 1.7769248 .. -0.06901258
   2.6032615 .. 3.2203112
   1.7484305 .. 1.8331614
   4.7545853 .. 1.7067102
   2.2185874 .. 1.2590411
-0.43089166 .. 0.29031274
   2.7479038 .. 1.5956203
                      ...

for each of the 53000 events, a calibrated image of the camera is obtained (array of 1855 pixels)

The rest of the tables in the files are metadata (info about the camera, optics, calibration, etc)

Software tool installation

LSTCHAIN

Donwload LST analysis software cta-lstchain (wrapper that uses centrally supported ctapipe routines) 

git clone https://github.com/cta-observatory/cta-lstchain.git
cd cta-lstchain

LSTCHAIN_VER=0.7.2
wget https://raw.githubusercontent.com/cta-observatory/cta-lstchain/v$LSTCHAIN_VER/environment.yml
conda env create -n lst -f environment.yml
conda activate lst
pip install lstchain==$LSTCHAIN_VER
rm environment.yml

In case of problems with conda env try mamba:
conda install -c conda-forge -n base mamba
mamba env create -f environment.yml

Files needed to analyze golden Run 442 (20190527):

  • Pedestal run: pedestal_file_run446_0000.fits
  • Low level DRS4 calibration: calibration.hdf5

Run directly jupyter notebook

cd cta-lstchain/notebooks
jupyter notebook
-> select Analyze_real_muon_data.ipynb

Convert example jupyter notebook to python

cd cta-lstchain/notebooks
jupyter nbconvert --to python Analyze_real_muon_data.ipynb

Fix path of data, pedestal and calibration and run! 

Lines to fix:
source = event_source(input_url="../data/LST-1.4.Run00442.0001.fits.fz",max_events=None)
pedestal_path="../pedestal/pedestal_file_run446_0000.fits",
with HDF5TableReader('../calibration/calibration.hdf5') as h5_table:

python Analyze_real_muon_data.py

Gammapy

FIX UBUNTU
conda create -n gammapyenv python=3.7
conda activate gammapyenv
conda install gammapy=0.18.2 sherpa=4.12.0
conda install jupyter ipython jupyterlab pandas healpy iminuit naima emcee corner parfice
gammapy info


curl -O https://gammapy.org/download/install/gammapy-0.18.2-environment.yml
conda env create -f gammapy-0.18.2-environment.yml
conda activate gammapy-0.18.2
gammapy info

Ctools

Installation MAC 

conda config --append channels conda-forge
conda config --append channels cta-observatory
conda create -n ctathai python=3.8
conda activate ctathai
conda install ctools
conda install matplotlib
conda install jupyterlab
export CALDB=/path/where/you/installed/IRF/caldb
export HESSDATA=/path/where/you/installed/HESS/data

Installation UBUNTU 

conda create -n ctathai python=3.7
conda activate ctathai
conda install -c cta-observatory ctools=1.7.4
conda install matplotlib
conda install jupyterlab
export CALDB=/path/where/you/installed/IRF/caldb
export HESSDATA=/path/where/you/installed/HESS/data


Commands: 

ctobssim  edisp=yes
fv events.fits
ctskymap 
ds9 skymap.fits 
ctbin
fv cntcube.fits 
ctexpcube
ctpsfcube
ctbkgcube
ctlike
csresmap
csresspec components=yes
$CTOOLS/share/examples/python/show_residuals.py resspec.fits

Fermipy

conda config --append channels fermi
conda create -n fermiset -c conda-forge -c fermi fermitools=2.0.8 python=3.7 clhep=2.4.4.1
conda activate fermiset
conda install fermipy

CTApipe

git clone https://github.com/cta-observatory/ctapipe.git
cd ctapipe

CTAPIPE_VER=0.10.5
wget https://raw.githubusercontent.com/cta-observatory/ctapipe/v$CTAPIPE_VER/environment.yml
conda env create -n cta -f environment.yml
conda activate cta
conda install -c conda-forge ctapipe=$CTAPIPE_VER

Ubuntu 20.04 issues

  • Mars/root does not work so gcc needs to be downgrade to version 5: gcc downgrade
  • System Program Problem Detected at reboot:
sudo gedit /etc/default/apport 
enabled=0
  • Problem installing ctools with conda, modify ~/.condarc
channel_priority: flexible


Extra

MAGIC DL3 data

https://github.com/open-gamma-ray-astro/joint-crab/tree/master/data/magic

Retrieved from "https://www.lst1.iac.es/wiki/index.php?title=LST1Analysis_MVA&oldid=3387"