Plotting time evolution of magnetic field

import pandas as pd
import km3compass as kc
import os
import matplotlib.pyplot as plt
import numpy as np

Getting some data

In this example we are downloading 3 runs to look at the magnetic

field time evolution. You can notice the field kwargs_query used in the reader. Here, we are using it to only focus on the data that come from the DU 11, reducing by a factor 6 the amount of data we download. Removing this criteria will just lead to the whole detector being download.

In the first code block, we also introduce a bit of intelligence to

check if the data were already downloaded, avoiding to re-download them when already available.

reader = None
filename = "sea_sample.h5"
filekey = "test_sample"
detoid = "D_ORCA006"

if not os.path.isfile(filename):
    reader = kc.readerOnline(
        detoid, minrun=9000, maxrun=9002, kwargs_query={"DUID": 11}
    )
    reader.save_df(filename, filekey)
else:
    reader = kc.readerOnline(filename=filename, filekey=filekey)

Loading data from DB ...
        detoid = D_ORCA006
        minrun = 9000
        maxrun = 9002
        Getting run 9000... - Done in 0 seconds
        Getting run 9001... - Done in 8 seconds
        Getting run 9002... - Done in 25 seconds
Data stored in 'sea_sample.h5', under key 'test_sample'

Calibrate the data

In this code block, we are now using the detector_calibration object to automatically calibrate all the data coming from our detector. This object also contains some summary function, to show which DOMs get calibrated, with which version of the calibraiton.

det_calib = kc.detector_calibration(reader)

det_calib.apply_calibration()
det_calib.print_calibration_summary()
det_calib.plot_calibration_summary()

DOM mac address : 08:00:30:11:bf:fe
DOM mac address : 08:00:30:11:ca:49
DOM mac address : 08:00:30:30:59:b4
DOM mac address : 08:00:30:30:60:1d
No calibration expected, 808476701 is a base module.
DOM mac address : 08:00:30:37:a3:79
DOM mac address : 08:00:30:37:cc:4d
DOM mac address : 08:00:30:37:ee:24
DOM mac address : 08:00:30:37:ff:2c
DOM mac address : 08:00:30:37:ff:5b
DOM mac address : 08:00:30:38:02:25
DOM mac address : 08:00:30:38:04:a4
DOM mac address : 08:00:30:38:09:06
DOM mac address : 08:00:30:38:09:0f
DOM mac address : 08:00:30:38:0a:f1
DOM mac address : 08:00:30:38:29:88
DOM mac address : 08:00:30:38:29:e3
DOM mac address : 08:00:30:38:2e:a7
DOM mac address : 08:00:30:38:30:7a
DOM mac address : 08:00:30:38:51:7b
DOM/BM before calibration: 18/1
DOM after calibration: 18
Details about calibration:

---------- Summary per DU ----------
DUID  TESTNAME                Variant  FIRMWARE_VERSION
11    AHRS-CALIBRATION-v3     AHRS     4.1                  1
                              LSM303   CLB                 17
      No calib (Base module)  AHRS     -                    1

---------- Summary full detector ----------
TESTNAME                Variant  FIRMWARE_VERSION
AHRS-CALIBRATION-v3     AHRS     4.1                  1
                        LSM303   CLB                 17
No calib (Base module)  AHRS     -                    1

<Figure size 640x480 with 1 Axes>

Resampling the dataset

The compass data are saved every 10 seconds for these runs. For most of the application, where we are looking at effect that takes hours to happen, this unnecessary. Here, we will average the value per 10 minutes slot, in order to reduce the measure fluctuation as well as getting a lighter data set.

This done with the pandas.DataFrame method resample. If you don’t know about, you should read the doc: this is awesome !

df = None
origin = np.min(det_calib.df["datetime"])

# Will resample separately each DOM, using the same time origin each time.
for modID in np.unique(det_calib.df["DOMID"]):
    df_tmp = det_calib.df[det_calib.df["DOMID"] == modID]
    df_tmp = df_tmp.resample("10min", on="datetime", origin=origin).mean(
        numeric_only=True
    )

    # df_tmp contains the resample value for dom modID
    # Now merging the results in df
    df = pd.concat((df, df_tmp))

Convert to spherical coordinates

Now that our dataset is calibrated and resampled, we can convert the x, y & z coordinates of the magnetic field in spherical coordinate, more usefull when it comes to look at the DOM positionning

coord_sphere = kc.cart2spherical(df[["AHRS_H0", "AHRS_H1", "AHRS_H2"]].values)
df["r"] = coord_sphere[:, 0]
df["phi"] = kc.rad2deg(coord_sphere[:, 1])
df["theta"] = kc.rad2deg(coord_sphere[:, 2])

# We also all rows that contains nan, just in case.
df = df.dropna(axis=0)

Plotting the time evolution

Here we are plotting 3 differents parameters in function of time : r, phi and theta. We define a coloscale that goes from 0 to 18 and that will represent the different DOMs along the DU.

for duid in np.unique(df["DUID"]):
    fig, axes = plt.subplots(3, 1, figsize=[6, 6], sharex=True)

    fig.suptitle("DU {}".format(int(duid)))

    axes[0].set_ylabel("phi (z) [deg]")
    axes[1].set_ylabel("theta (x:y) [deg]")
    axes[2].set_ylabel("Norm [G]")

    ind = df["DUID"] == duid

    kwargs = {
        "marker": ".",
        "c": df["FLOORID"][ind],
        "vmin": 0,
        "vmax": 18,
        "linestyle": "-",
    }

    sc = axes[0].scatter(df.index[ind], df["phi"][ind], **kwargs)
    fig.colorbar(sc, ax=axes[0])
    sc = axes[1].scatter(df.index[ind], df["theta"][ind], **kwargs)
    fig.colorbar(sc, ax=axes[1])
    sc = axes[2].scatter(df.index[ind], df["r"][ind], **kwargs)
    fig.colorbar(sc, ax=axes[2])

    plt.xticks(rotation=45)
    plt.tight_layout()

plt.show()