bandsaunter/tests/test_kml.py

"""The map of who was heard.

KML is a text format, and every one of these reads back what was written
with an XML parser rather than by matching strings, so a file that would not
open in Google Earth fails here first.

Nothing touches the network: the licence data is built by hand.
"""
import time
import xml.etree.ElementTree as ET

import pytest

from bandsaunter.callsign import Callsign, CallsignBook, grid_to_latlon
from bandsaunter.kml import Contact, Hearing, KmlLog

KML_NS = "{http://www.opengis.net/kml/2.2}"


@pytest.fixture(autouse=True)
def no_network(monkeypatch):
    def refuse(self, url, call):
        raise AssertionError(f"a test tried to look up {call} for real")

    monkeypatch.setattr(CallsignBook, "_request", refuse)


def _when(text: str) -> float:
    return time.mktime(time.strptime(text, "%Y-%m-%d %H:%M:%S"))


def _entry(call="KU0W", **kw) -> Callsign:
    body = dict(name="Rod R Gowdy", location="Tucson, AZ",
                country="United States", grid="DM42lj", oper_class="EXTRA",
                latitude=32.3345, longitude=-111.0421, status="found")
    body.update(kw)
    return Callsign(call=call, **body)


def _placemarks(path):
    root = ET.parse(path).getroot()
    return root.iter(KML_NS + "Placemark")


def _fields(placemark) -> dict:
    out = {}
    for data in placemark.iter(KML_NS + "Data"):
        value = data.find(KML_NS + "value")
        out[data.get("name")] = value.text if value is not None else ""
    return out


# ---------------------------------------------------------------------------
# One station, one pin
# ---------------------------------------------------------------------------

def test_a_callsign_becomes_a_placemark_where_it_is_licensed(tmp_path):
    log = KmlLog(tmp_path / "map.kml")
    log.add(_entry(), 146.88e6, _when("2026-08-22 13:01:44"),
            recording="0146.880000MHz-nfm.wav")
    written = log.save()
    assert written is not None and written.exists()

    marks = list(_placemarks(written))
    assert len(marks) == 1
    fields = _fields(marks[0])
    assert fields["callsign"] == "KU0W"
    assert fields["licensee"] == "Rod R Gowdy"
    assert fields["location"] == "Tucson, AZ"
    assert fields["grid"] == "DM42lj"
    point = marks[0].find(f"{KML_NS}Point/{KML_NS}coordinates")
    lon, lat, _ = point.text.split(",")
    # KML puts longitude first, which is the easy thing to get backwards.
    assert float(lon) == pytest.approx(-111.0421, abs=1e-4)
    assert float(lat) == pytest.approx(32.3345, abs=1e-4)


def test_the_name_on_the_pin_is_the_callsign_and_the_licensee(tmp_path):
    log = KmlLog(tmp_path / "map.kml")
    log.add(_entry(), 146.88e6, _when("2026-08-22 13:01:44"))
    mark = next(_placemarks(log.save()))
    assert mark.find(KML_NS + "name").text == "KU0W — Rod R Gowdy"


def test_the_balloon_names_the_frequency_and_the_time(tmp_path):
    log = KmlLog(tmp_path / "map.kml")
    log.add(_entry(), 146.88e6, _when("2026-08-22 13:01:44"))
    body = next(_placemarks(log.save())).find(KML_NS + "description").text
    assert "146.88 MHz" in body
    assert "2026-08-22 13:01:44" in body
    assert "2 m Amateur" in body or "2 m" in body
    assert "Rod R Gowdy" in body


def test_hearing_a_station_again_adds_a_line_not_a_second_pin(tmp_path):
    log = KmlLog(tmp_path / "map.kml")
    for offset in (0, 300, 900):
        log.add(_entry(), 146.88e6, _when("2026-08-22 13:01:44") + offset)
    written = log.save()
    assert len(list(_placemarks(written))) == 1
    heard = _fields(next(_placemarks(written)))["heard"].splitlines()
    assert len(heard) == 3


def test_the_very_same_transmission_is_not_recorded_twice(tmp_path):
    """A re-run over the same recordings must not inflate the count."""
    log = KmlLog(tmp_path / "map.kml")
    at = _when("2026-08-22 13:01:44")
    assert log.add(_entry(), 146.88e6, at) is True
    assert log.add(_entry(), 146.88e6, at) is False
    heard = _fields(next(_placemarks(log.save())))["heard"].splitlines()
    assert len(heard) == 1


def test_a_station_heard_on_two_bands_keeps_both(tmp_path):
    log = KmlLog(tmp_path / "map.kml")
    at = _when("2026-08-22 13:01:44")
    log.add(_entry(), 146.88e6, at)
    log.add(_entry(), 7.242e6, at + 60)
    heard = _fields(next(_placemarks(log.save())))["heard"]
    assert "146.88 MHz" in heard and "7.242 MHz" in heard


def test_the_time_span_covers_first_to_last(tmp_path):
    log = KmlLog(tmp_path / "map.kml")
    at = _when("2026-08-22 13:01:44")
    log.add(_entry(), 146.88e6, at)
    log.add(_entry(), 146.88e6, at + 3600)
    mark = next(_placemarks(log.save()))
    span = mark.find(KML_NS + "TimeSpan")
    assert span is not None
    begin = span.find(KML_NS + "begin").text
    end = span.find(KML_NS + "end").text
    assert begin < end and begin.endswith("Z")


# ---------------------------------------------------------------------------
# Continuing an existing map
# ---------------------------------------------------------------------------

def test_a_later_run_adds_to_the_map_rather_than_replacing_it(tmp_path):
    path = tmp_path / "map.kml"
    first = KmlLog(path)
    first.add(_entry(), 146.88e6, _when("2026-08-22 13:01:44"))
    first.save()

    second = KmlLog(path)
    assert "KU0W" in second.contacts, "the existing map was not read back"
    second.add(_entry("W1AW", name="Arrl Hq Operators Club",
                      location="Newington, CT", grid="FN31pr",
                      latitude=41.7147, longitude=-72.7270),
               3.965e6, _when("2026-08-23 20:00:00"))
    written = second.save()
    calls = {_fields(m)["callsign"] for m in _placemarks(written)}
    assert calls == {"KU0W", "W1AW"}


def test_what_was_heard_before_survives_the_round_trip(tmp_path):
    path = tmp_path / "map.kml"
    first = KmlLog(path)
    at = _when("2026-08-22 13:01:44")
    first.add(_entry(), 146.88e6, at, recording="one.wav")
    first.save()

    second = KmlLog(path)
    heard = second.contacts["KU0W"].hearings
    assert len(heard) == 1
    assert heard[0].frequency == pytest.approx(146.88e6)
    assert heard[0].at == pytest.approx(at)
    assert heard[0].recording == "one.wav"
    # and the same transmission read back is still not added twice
    assert second.add(_entry(), 146.88e6, at) is False


def test_a_position_learned_later_updates_the_pin(tmp_path):
    path = tmp_path / "map.kml"
    first = KmlLog(path)
    first.add(Callsign(call="KU0W", status="offline"), 146.88e6, 1.0)
    first.save()

    second = KmlLog(path)
    second.add(_entry(), 146.88e6, 2.0)
    mark = next(_placemarks(second.save()))
    assert mark.find(f"{KML_NS}Point/{KML_NS}coordinates") is not None


def test_a_failed_lookup_does_not_blank_out_what_is_known(tmp_path):
    log = KmlLog(tmp_path / "map.kml")
    log.add(_entry(), 146.88e6, 1.0)
    log.add(Callsign(call="KU0W", status="offline"), 146.88e6, 2.0)
    fields = _fields(next(_placemarks(log.save())))
    assert fields["licensee"] == "Rod R Gowdy"


# ---------------------------------------------------------------------------
# Stations with nowhere to put them
# ---------------------------------------------------------------------------

def test_a_callsign_with_no_licence_is_kept_but_not_placed(tmp_path):
    log = KmlLog(tmp_path / "map.kml")
    log.add(Callsign(call="2E0XYZ", country="United Kingdom",
                     status="unlisted"), 145.5e6, 1.0)
    written = log.save()
    mark = next(_placemarks(written))
    assert _fields(mark)["callsign"] == "2E0XYZ"
    assert mark.find(f"{KML_NS}Point/{KML_NS}coordinates") is None
    root = ET.parse(written).getroot()
    folder = root.find(f".//{KML_NS}Folder/{KML_NS}name")
    assert folder is not None and "no location" in folder.text


def test_a_grid_square_places_a_station_the_licence_did_not(tmp_path):
    log = KmlLog(tmp_path / "map.kml")
    log.add(Callsign(call="W1AW", grid="FN31pr", status="found"),
            3.965e6, 1.0)
    mark = next(_placemarks(log.save()))
    point = mark.find(f"{KML_NS}Point/{KML_NS}coordinates")
    lon, lat, _ = point.text.split(",")
    assert float(lat) == pytest.approx(41.73, abs=0.05)
    assert float(lon) == pytest.approx(-72.71, abs=0.05)
    # and it says so, because a grid square is kilometres wide
    assert "grid square" in mark.find(KML_NS + "description").text


def test_a_record_with_no_coordinates_is_not_placed_off_africa(tmp_path):
    """Zeroed coordinates are missing data, not a position in the Atlantic."""
    entry = Callsign(call="K7XYZ", status="found",
                     latitude=0.0, longitude=0.0)
    log = KmlLog(tmp_path / "map.kml")
    log.add(entry, 146.0e6, 1.0)
    mark = next(_placemarks(log.save()))
    assert mark.find(f"{KML_NS}Point/{KML_NS}coordinates") is None


@pytest.mark.parametrize("grid,lat,lon", [
    ("DM42lj", 32.40, -111.04),
    ("FN31pr", 41.73, -72.71),
    ("EM79", 39.5, -85.0),
])
def test_grid_squares_convert_to_their_centre(grid, lat, lon):
    got = grid_to_latlon(grid)
    assert got == pytest.approx((lat, lon), abs=0.06)


@pytest.mark.parametrize("bad", ["", "nonsense", "ZZ99", "DM4", "DM42lj9"])
def test_what_is_not_a_grid_square_converts_to_nothing(bad):
    assert grid_to_latlon(bad) is None


# ---------------------------------------------------------------------------
# Not breaking the file
# ---------------------------------------------------------------------------

def test_a_name_with_xml_in_it_does_not_break_the_document(tmp_path):
    log = KmlLog(tmp_path / "map.kml")
    log.add(_entry(name="Smith & Sons <Radio> Club"), 146.88e6, 1.0)
    fields = _fields(next(_placemarks(log.save())))
    assert fields["licensee"] == "Smith & Sons <Radio> Club"


def test_an_unreadable_file_is_never_overwritten(tmp_path):
    path = tmp_path / "map.kml"
    path.write_text("this is not KML, and something else wrote it")
    log = KmlLog(path)
    log.add(_entry(), 146.88e6, 1.0)
    assert log.save() is None
    assert path.read_text().startswith("this is not KML")


def test_nothing_heard_writes_no_file(tmp_path):
    path = tmp_path / "map.kml"
    assert KmlLog(path).save() is None
    assert not path.exists()


def test_the_write_leaves_no_temporary_behind(tmp_path):
    log = KmlLog(tmp_path / "map.kml")
    log.add(_entry(), 146.88e6, 1.0)
    log.save()
    assert [p.name for p in tmp_path.iterdir()] == ["map.kml"]


def test_a_callsign_with_no_call_is_refused(tmp_path):
    log = KmlLog(tmp_path / "map.kml")
    assert log.add(Callsign(call=""), 146.0e6, 1.0) is False
    assert len(log) == 0


# ---------------------------------------------------------------------------
# The pieces
# ---------------------------------------------------------------------------

def test_a_hearing_reads_back_what_it_wrote():
    original = Hearing(146.88e6, _when("2026-08-22 13:01:44"),
                       "2 m Amateur", "capture.wav")
    again = Hearing.parse(original.describe())
    assert again.frequency == pytest.approx(original.frequency)
    assert again.at == pytest.approx(original.at)
    assert again.band == original.band
    assert again.recording == original.recording


def test_a_mangled_line_loses_only_itself():
    """Someone editing the map by hand must not destroy the rest of it."""
    assert Hearing.parse("") is None
    assert Hearing.parse("who knows what this is") is not None


def test_a_contact_orders_its_hearings_by_time():
    c = Contact(call="KU0W")
    c.add(Hearing(146.88e6, 300.0))
    c.add(Hearing(146.88e6, 100.0))
    assert [h.at for h in c.hearings] == [100.0, 300.0]
    assert c.first_heard == 100.0 and c.last_heard == 300.0


# ---------------------------------------------------------------------------
# A scan that hears somebody
# ---------------------------------------------------------------------------

CALLOOK = {
    "status": "VALID",
    "type": "PERSON",
    "current": {"callsign": "W1AW", "operClass": "CLUB"},
    "name": "ARRL HQ OPERATORS CLUB",
    "address": {"line1": "225 MAIN ST", "line2": "NEWINGTON, CT 06111"},
    "location": {"latitude": "41.714775", "longitude": "-72.727260",
                 "gridsquare": "FN31pr"},
    "otherInfo": {"expiryDate": "02/29/2032"},
}


@pytest.fixture
def fake_lookup(monkeypatch):
    """The licence database, answering from a fixture instead of the network."""
    asked = []

    def respond(self, url, call):
        asked.append(call)
        return dict(CALLOOK, current={"callsign": call, "operClass": "CLUB"})

    monkeypatch.setattr(CallsignBook, "_request", respond)
    return asked


@pytest.fixture
def heard_speech(monkeypatch):
    """A recogniser that hears one station identify itself."""
    from bandsaunter import transcribe as tr

    monkeypatch.setitem(
        tr._DISPATCH, "fake",
        lambda audio, rate, model, lang: tr.Transcript(
            text="Net control, this is W1AW, standing by.", engine="fake"))
    monkeypatch.setattr(tr, "ENGINES", ("fake",) + tr.ENGINES)
    monkeypatch.setattr(tr, "_is_present", lambda name: name == "fake")
    monkeypatch.setattr("bandsaunter.scanner.available_engine", lambda: "fake")


def _scan(tmp_path, **over):
    from bandsaunter.config import ScanConfig
    from bandsaunter.ranges import parse_range_list
    from bandsaunter.scanner import Scanner, ScannerCallbacks
    from bandsaunter.simulator import SimulatedDevice, VirtualTransmitter as V

    cfg = ScanConfig(ranges=parse_range_list("146.4M-146.6M"),
                     output_dir=str(tmp_path), record_seconds=4.0,
                     hang_seconds=1.0, threshold_db=12, dwell_seconds=0.05,
                     max_cycles=1, revisit_seconds=0.2, transcribe=True,
                     transcribe_engine="fake")
    on_status = over.pop("_on_status", None)
    for key, value in over.items():
        setattr(cfg, key, value)
    scanner = Scanner(cfg, device=SimulatedDevice(
        transmitters=[V(146_520_000, "nfm", 0.4, 12_500, "v")]).open(),
        callbacks=ScannerCallbacks(on_status=on_status) if on_status else None)
    scanner.prepare()
    scanner.run()
    return scanner


def test_a_scan_maps_the_stations_it_hears(tmp_path, heard_speech,
                                           fake_lookup):
    scanner = _scan(tmp_path)
    written = tmp_path / "callsigns.kml"
    assert written.exists(), "the scan heard a callsign but wrote no map"
    assert fake_lookup == ["W1AW"]

    mark = next(_placemarks(written))
    fields = _fields(mark)
    assert fields["callsign"] == "W1AW"
    assert fields["licensee"] == "ARRL HQ Operators Club"
    assert fields["location"] == "Newington, CT"
    # The frequency is the one the detector settled on, which is
    # near the transmitter rather than exactly on it.
    assert "146.5" in fields["heard"]
    assert "2 m FM Simplex" in fields["heard"]
    assert ".wav" in fields["heard"]
    assert mark.find(f"{KML_NS}Point/{KML_NS}coordinates") is not None
    assert scanner.heard.get("W1AW")


def test_a_scan_says_on_the_display_who_it_heard(tmp_path, heard_speech,
                                                 fake_lookup):
    said = []
    _scan(tmp_path, _on_status=said.append)
    spoken = [m for m in said if "W1AW" in m]
    assert spoken, f"the scan never announced the callsign: {said}"
    assert "ARRL HQ Operators Club" in spoken[0]


def test_the_same_station_is_only_announced_once(tmp_path, heard_speech,
                                                 fake_lookup):
    """A repeater net would otherwise fill the status line with one name."""
    said = []
    _scan(tmp_path, _on_status=said.append, max_cycles=3,
          revisit_seconds=0.0)
    assert len([m for m in said if m.startswith("heard W1AW")]) == 1


def test_the_map_is_off_when_no_file_is_named(tmp_path, heard_speech,
                                              fake_lookup):
    _scan(tmp_path, kml_file="")
    assert not list(tmp_path.glob("*.kml"))


def test_lookups_can_be_turned_off_entirely(tmp_path, heard_speech):
    """Nothing may reach the network with the lookup switched off."""
    def refuse(self, url, call):
        raise AssertionError("looked a callsign up with lookups off")

    import bandsaunter.callsign as cs
    original, cs.CallsignBook._request = cs.CallsignBook._request, refuse
    try:
        _scan(tmp_path, callsign_lookup=False)
    finally:
        cs.CallsignBook._request = original
    # The station is still on the map, named by its prefix rather than a
    # licence: that it was heard is worth recording either way.
    written = tmp_path / "callsigns.kml"
    assert written.exists()
    assert _fields(next(_placemarks(written)))["callsign"] == "W1AW"


def test_a_second_scan_adds_to_the_same_map(tmp_path, heard_speech,
                                            fake_lookup):
    _scan(tmp_path)
    first = len(_fields(next(_placemarks(tmp_path / "callsigns.kml")
                             ))["heard"].splitlines())
    _scan(tmp_path)
    marks = list(_placemarks(tmp_path / "callsigns.kml"))
    assert len(marks) == 1, "a second scan started a second pin"
    assert len(_fields(marks[0])["heard"].splitlines()) > first