starting coding the svom ft tom filters

fink_filters/filter_svom_ft_tom/filter.py

@@ -0,0 +1,150 @@
+import pandas as pd
+
+from fink_filters.classification import extract_fink_classification_
+from fink_utils.xmatch.simbad import return_list_of_eg_host
+
+from astropy.coordinates import SkyCoord
+from pyspark.sql.functions import pandas_udf
+from pyspark.sql.types import BooleanType
+from fink_filters.tester import spark_unit_tests
+
+
+@pandas_udf(BooleanType())
+def gvom_filter(
+    rb: pd.Series,
+    magpsf: pd.Series,
+    mag_rate: pd.Series,
+    ra: pd.Series,
+    dec: pd.Series,
+    cdsxmatch: pd.Series,
+    roid: pd.Series,
+    mulens: pd.Series,
+    snn_snia_vs_nonia: pd.Series,
+    snn_sn_vs_all: pd.Series,
+    rf_snia_vs_nonia: pd.Series,
+    ndethist: pd.Series,
+    drb: pd.Series,
+    classtar: pd.Series,
+    jd: pd.Series,
+    jdstarthist: pd.Series,
+    rf_kn_vs_nonkn: pd.Series,
+    tracklet: pd.Series,
+) -> pd.Series:
+    """
+    Science filter of the gvom network.
+    Select alerts that are likely to be fast transient.
+
+    Parameters
+    ----------
+    rb : pd.Series
+        real bogus
+    magpsf : pd.Series
+        magnitude
+    ra : pd.Series
+        right ascension
+    dec : pd.Series
+        declination
+    cdsxmatch : pd.Series
+        cds tag
+    roid : pd.Series
+        asteroid tag
+    mulens : pd.Series
+        micro-lensing score
+    snn_snia_vs_nonia : pd.Series
+        supernovae Ia score
+    snn_sn_vs_all : pd.Series
+        supernovae all score
+    rf_snia_vs_nonia : pd.Series
+        supernovae Ia score
+    ndethist : pd.Series
+        number of detection in the history
+    drb : pd.Series
+        deep learning real bogus
+    classtar : pd.Series
+        Sextractor score
+    jd : pd.Series
+        julian date of the alerts
+    jdstarthist : pd.Series
+        start history julian date
+    rf_kn_vs_nonkn : pd.Series
+        kilonovae score
+    tracklet : pd.Series
+        satellite identification
+
+    Returns
+    -------
+    pd.Series
+        if True, the alert match the gvom science filter
+
+    Example
+    -------
+    >>> from fink_utils.spark.utils import apply_user_defined_filter
+    >>> df = spark.read.format('parquet').load('datatest')
+    >>> f = 'fink_filters.filter_svom_ft_tom.filter.gvom_filter'
+    >>> df = apply_user_defined_filter(df, f)
+    >>> print(df.count())
+    0
+    """
+
+    classification = extract_fink_classification_(
+        cdsxmatch,
+        roid,
+        mulens,
+        snn_snia_vs_nonia,
+        snn_sn_vs_all,
+        rf_snia_vs_nonia,
+        ndethist,
+        drb,
+        classtar,
+        jd,
+        jdstarthist,
+        rf_kn_vs_nonkn,
+        tracklet,
+    )
+
+    # non-buggy alerts
+    f_bogus = rb >= 0.9
+
+    # extra-galactic alerts
+    base_extragalactic = return_list_of_eg_host()
+    fink_extragalactic = [
+        "KN candidate",
+        "SN candidate",
+        "Early SN Ia candidate",
+        "Ambiguous",
+    ]
+    extragalactic = list(base_extragalactic) + list(fink_extragalactic)
+    f_class = classification.isin(extragalactic)
+
+    coord = SkyCoord(ra, dec, unit="deg")
+    # alerts not in the milky way
+    gal_latitude = coord.galactic.b.value
+    mask_south_gal = gal_latitude < -15
+    mask_north_gal = gal_latitude > 15
+    f_gal = mask_north_gal | mask_south_gal
+
+    # alerts not in the ecliptic
+    ecl_latitude = coord.transform_to("geocentricmeanecliptic").lat.value
+    mask_south_ecl = ecl_latitude < -15
+    mask_north_ecl = ecl_latitude > 15
+    f_ecl = mask_north_ecl | mask_south_ecl
+
+    # bright alerts
+    f_brightness = magpsf <= 18.0
+
+    # short living transient
+    f_short = jd - jdstarthist <= 5
+
+    # fast transient
+    f_fast = mag_rate.abs() > 0.3
+
+    f_gvom = f_bogus & f_class & f_short & f_brightness & f_gal & f_ecl & f_fast
+    return f_gvom
+
+
+if __name__ == "__main__":
+    """Execute the test suite"""
+
+    # Run the test suite
+    globs = globals()
+    spark_unit_tests(globs)

requirements.txt

@@ -16,3 +16,5 @@ pandas
 pyarrow
 
 h5py
+
+py4j>=0.10.9.5