KML Output

[darksidelemm]

I suggest the following code to add KML output to the amateur telemetry endpoint (Could be used for radiosondes too?)
This would go here: https://github.com/projecthorus/sondehub-infra/blob/main/lambda/query_ham/__init__.py#L373

Apologies that this is not a pull request, I'm not sure how to test this properly.

    elif (
        "queryStringParameters" in event
        and "format" in event["queryStringParameters"]
        and event["queryStringParameters"]['format'] == "kml"
    ):
        content_type = "application/vnd.google-earth.kml+xml"

        # Extract some basic flight info for use in KML Metadata
        callsign = data[0]['payload_callsign']
        start_datetime = data[0]['datetime']

        # Only get unique date/time data.
        # This results in a much shorter dictionary.
        _temp_data = {}
        for _telem in data:
            _temp_data[_telem['datetime']] = f"{float(_telem['lon']):.6f},{float(_telem['lat']):.6f},{float(_telem['alt']):.1f}\n"

        # For a KML output, the data *must* be sorted, else the KML LineString becomes a complete mess.
        # Get the keys from the dictionary generated above, and sort them.
        data_keys = list(_temp_data.keys())
        data_keys.sort()

        # Now generate the LineString data (lon,lat,alt)
        # This could possibly done with a .join, but I suspect that wont be much more efficient.
        kml_coords = ""
        for _key in data_keys:
            kml_coords += _temp_data[_key]

        # Generate the output KML. 
        # This is probably the simplest way of handling this without bringing in
        # any other libraries.
        kml_out = f"""
        <kml:Document xmlns:kml="http://www.opengis.net/kml/2.2">
        <kml:visibility>1</kml:visibility>
        <kml:Folder id="{callsign}">
            <kml:name>{start_datetime} {callsign}</kml:name>
            <kml:description>Flight Path</kml:description>
            <kml:visibility>1</kml:visibility>
            <kml:Placemark id="Flight Path ID">
                <kml:name>{start_datetime} {callsign}</kml:name>
                <kml:visibility>1</kml:visibility>
                <kml:Style><kml:LineStyle>
                <kml:color>aaffffff</kml:color>
                <kml:width>2.0</kml:width>
                </kml:LineStyle>
                <kml:PolyStyle>
                <kml:color>20000000</kml:color>
                <kml:fill>1</kml:fill>
                <kml:outline>1</kml:outline>
                </kml:PolyStyle></kml:Style>
                <kml:LineString>
                    <kml:extrude>1</kml:extrude>
                    <kml:altitudeMode>absolute</kml:altitudeMode>
                    <kml:coordinates>
        {kml_coords}
                    </kml:coordinates>
                </kml:LineString>
        </kml:Placemark>
        </kml:Folder>
        </kml:Document>
        """

        # Finally replace the data with the kml data
        data = kml_out

Unsure how the indentation will work...

On the performance of this - I tested this on some downloaded json data for a HORUSBINARY flight, with 62529 individual telemetry packets.

Times are in seconds:

• De-Dupe and Generate lon,lat,alt: 0.06961917877197266
• Sort Keys: 0.0005059242248535156
• Generate LineString: 0.0006291866302490234
  So approx 80ms or so.

As a comparison, the json.dumps(data) line at https://github.com/projecthorus/sondehub-infra/blob/main/lambda/query_ham/__init__.py#L374 takes a whopping 292ms to run. So, this kml export, even with de-duping and sorting, will actually be faster than the JSON export!