add format data

scripts/format_data.py

@@ -0,0 +1,48 @@
+from __future__ import annotations
+from traffic.data import (  # noqa: F401
+    aixm_airspaces,
+    aixm_navaids,
+    aixm_airways,
+)
+from traffic.core import Traffic
+import pandas as pd
+
+# here, extent is the Bordeaux centre, altitude_min is FL200
+extent = "LFBBBDX"
+altitude_min = 20000
+
+# for each flight, metadata contains icao24, callsign, flight id and flight plan
+metadata = pd.read_parquet("A2207_old.parquet")
+metadata_simple = (
+    metadata.groupby("flight_id", as_index=False)
+    .last()
+    .eval("icao24 = icao24.str.lower()")
+)
+
+# directory contains the trajectories as parquet files
+directory = "../../LFBB_A2207"
+concatenated_df = pd.read_parquet(directory)
+
+# first filter on the data
+# we originally the "onground" value, but we drop it here as it is not useful
+t = (
+    Traffic(concatenated_df)
+    .drop(columns=["onground"])
+    .clip(aixm_airspaces[extent])
+    .query(f"altitude>{altitude_min}")
+    .eval(max_workers=4)
+)
+
+# in t2, we assign flight ids using metadata_simple
+t2 = (
+    t.iterate_lazy(iterate_kw=dict(by=metadata_simple))
+    .resample("1s")
+    .eval(desc="", max_workers=4)
+)
+
+# filter and resample
+assert t2 is not None
+t2 = t2.filter().resample("1s").eval(max_workers=4)  # pour retest
+assert t2 is not None
+
+t2.to_parquet("test_format_data_1.parquet")