Embed URL to coordinates, kml or geojson

Question

I've got a complex directions URL and embed URL that I would like to get polylines for. Once I can get them into polylines or something similar I can convert to final format: GeoJSON.

Direction Link

-or-

Embed Link

I have looked at the API's and I can't find anything that accepts or would decode the PB (what is this? it's not a protocol buffer). So far this is as far as I've got:

//php
$pb_array = explode('!', $pb);
foreach($pb_array as $key => $value){
    echo "$key - $value<br/>";
}

===

1 - 1m73 2 - 1m12 3 - 1m3 4 - 1d1472548.9575794793 5 - 2d-72.8191002664707 6 - 3d43.87505426780168 7 - 2m3 8 - 1f0 9 - 2f0 10 - 3f0 11 - 3m2 12 - 1i1024 13 - 2i768 14 - 4f13.1 15 - 4m58 16 - 3e0 17 - 4m5 18 - 1s0x0%3A0xa58b3d6041ba69f8 19 - 2sGuilford+Welcome+Center 20 - 3m2 21 - 1d42.8120069 22 - 2d-72.56614689999999 23 - 4m3 24 - 3m2 25 - 1d43.3893165 26 - 2d-72.40772249999999 27 - 4m5 28 - 1s0x4cb52e78df455c83%3A0xb6946ec850907db8 29 - 2s130+Lower+Michigan+Road%2C+Pittsfield%2C+VT+05762 30 - 3m2 31 - 1d43.76898 32 - 2d-72.815214 33 - 4m4 34 - 1s0x0%3A0xea2de48bba82cc86 35 - 3m2 36 - 1d44.042544 37 - 2d-72.6046997 38 - 4m5 39 - 1s0x0%3A0x6bb602ed58bf4413 40 - 2sJay+Peak+Resort 41 - 3m2 42 - 1d44.9379515 43 - 2d-72.5045433 44 - 4m5 45 - 1s0x4cb392aaa4333a07%3A0x160aef1559868340 46 - 2sDolly+Copp+Campground+Rd%2C+Gorham%2C+NH+03581 47 - 3m2 48 - 1d44.335842199999995 49 - 2d-71.21837339999999 50 - 4m5 51 - 1s0x4cb392684201a94d%3A0xfa4a6f490a05429d 52 - 2sMt+Washington+Auto+Road%2C+1+Mount+Washington+Auto+Road%2C+Gorham%2C+NH+03581 53 - 3m2 54 - 1d44.288384099999995 55 - 2d-71.22459599999999 56 - 4m5 57 - 1s0x4cb38e798f42c3d9%3A0xc3b88e4dac01db12 58 - 2sMt+Washington 59 - 3m2 60 - 1d44.270585399999995 61 - 2d-71.3032723 62 - 4m5 63 - 1s0x89e2a7fa444124d5%3A0xe3ed24b6f864eba0 64 - 2sWells%2C+ME 65 - 3m2 66 - 1d43.322232899999996 67 - 2d-70.5805209 68 - 4m5 69 - 1s0x89e2ba813e828c71%3A0x8cdf74380f6a933d 70 - 2sLibby's+Oceanside+Camp%2C+York+Street%2C+York%2C+ME 71 - 3m2 72 - 1d43.147162 73 - 2d-70.626173 74 - 5e1 75 - 3m2 76 - 1sen 77 - 2sus 78 - 4v1472497940601

The closest hints I could find are from this thread. I will keep looking but I'm stuck.

I'm trying to create an API based solution that has an input of one of these URL's and returns a GeoJSON.

Answer 1

would decode the PB (what is this? it's not a protocol buffer)

For the record, because this overflow question keeps popping up on google results: it is a protocol buffer. PB litteraly stands for protocol buffer.

It's just a different ASCII encoding (a compact URL encoding reminiscent of the binary encoding, not the usual JSON-like text encoding. When you squint at it it's not that much different than torrent's structure encoding), and Google doesn't provide us the .proto file.

For each field:

• first character is the id (identifies the field according to the corresponding .proto file)
• second character is the type of the field
  • m is for message
  • s is for string
  • i, j, u, v are for various type of ints
  • f, d are for floating points
  • e is for enum
• the rest is the payload

So to unpack the fields you're seeing (even if we don't have the .proto file):

• 1m73 message of type 1, containing 73 elements (the whole message set)
  • 1m12 submessage of type 1, contains 12 elements (probably information about the view box in the map box)
    • 1m3 sub-sub-message type 1, contains 3 elements (probably map coordinates)
      • 1d1472548.9575794793 first double field (probably zoom level)
      • 2d-72.8191002664707 second double field (probably longitude)
      • 3d43.87505426780168 third double field (probably latitude)
    • 2m3 second sub-sub-sub message (no idea given that it's not filled. Maybe a starting point if you code a route instead of a single point ?)
      • 1f0, 2f0, 3f0 the three members, currently just zero
    • 3m2 third block (looks like a screen resolution)
      • 1i1024, 2i768 : 1024x768 ? (and probably the omited field 3 would have been the color depth if present ??)
      • 4f13.1 no idea, but it's a float
  • 4m58 next message with 58 elements (to me it looks like a bunch of POI that we need to display in the box)

    • 3e0 an enum, set to zero (this one would be completely impossible to interpret without a proto or without experimenting, as you need the list of enums)

    • 4m5 five more elements probably a map poi - 1s0x0%3A0xa58b3d6041ba69f8 string '0x0:0xa58b3d6041ba69f8', note the use of Url_encoded character. In turn it looks like a pair of hex numbers, maybe a GUID ? - 2sGuilford+Welcome+Center string, the name with plus instead of blank (like most URLs) - 3m2 two elements to come - 1d42.8120069 and 2d-72.56614689999999 doubles probably map coordinates

    • 4m3 again a message of type 4 in this level, so probably another poi - 3m2, 1d43.3893165, 2d-72.40772249999999 but this one only specifies coordinates, and nothing else

    • 4m5 another poi - 1s0x4cb52e78df455c83%3A0xb6946ec850907db8 different pair of hex, GUID

      ...you got the idea...

  • 5e1 another bunch of information probably general settings
    • 3m2 this setting is a message (and looks like a locale)
      • 1sen, 2sus locale is en_US
    • 4v1472497940601 some other large number...

Note: the original proto that Google doesn't show us, is probably a single multi level structure. Thus, the sub-sub-message ID don't have always the same meaning: they aren't global ID, but ID within the parent message.

• inside a sub message ID 1 (view box ?), the sub-sub message 3 seems to be resolution.
• inside a sub message ID 4 (POIs ?), the sub-ID 3 isn't even a message but some enum
• inside a sub message ID 5 (parameters), the sub-sub message 3 is a locale

Answer 2

Well, here's my sloppy but working solution. My needs required GeoJSON, but others can use the google maps service to get the desired output once you have the lat/lng array.

$embed = '<iframe src="https://www.google.com/maps/embed?pb=!1m73!1m12!1m3!1d1472548.9575794793!2d-72.8191002664707!3d43.87505426780168!2m3!1f0!2f0!3f0!3m2!1i1024!2i768!4f13.1!4m58!3e0!4m5!1s0x0%3A0xa58b3d6041ba69f8!2sGuilford+Welcome+Center!3m2!1d42.8120069!2d-72.56614689999999!4m3!3m2!1d43.3893165!2d-72.40772249999999!4m5!1s0x4cb52e78df455c83%3A0xb6946ec850907db8!2s130+Lower+Michigan+Road%2C+Pittsfield%2C+VT+05762!3m2!1d43.76898!2d-72.815214!4m4!1s0x0%3A0xea2de48bba82cc86!3m2!1d44.042544!2d-72.6046997!4m5!1s0x0%3A0x6bb602ed58bf4413!2sJay+Peak+Resort!3m2!1d44.9379515!2d-72.5045433!4m5!1s0x4cb392aaa4333a07%3A0x160aef1559868340!2sDolly+Copp+Campground+Rd%2C+Gorham%2C+NH+03581!3m2!1d44.335842199999995!2d-71.21837339999999!4m5!1s0x4cb392684201a94d%3A0xfa4a6f490a05429d!2sMt+Washington+Auto+Road%2C+1+Mount+Washington+Auto+Road%2C+Gorham%2C+NH+03581!3m2!1d44.288384099999995!2d-71.22459599999999!4m5!1s0x4cb38e798f42c3d9%3A0xc3b88e4dac01db12!2sMt+Washington!3m2!1d44.270585399999995!2d-71.3032723!4m5!1s0x89e2a7fa444124d5%3A0xe3ed24b6f864eba0!2sWells%2C+ME!3m2!1d43.322232899999996!2d-70.5805209!4m5!1s0x89e2ba813e828c71%3A0x8cdf74380f6a933d!2sLibby&#39;s+Oceanside+Camp%2C+York+Street%2C+York%2C+ME!3m2!1d43.147162!2d-70.626173!5e1!3m2!1sen!2sus!4v1472497940601" width="600" height="450" frameborder="0" style="border:0" allowfullscreen></iframe>';

$array = array();
preg_match( '/src="([^"]*)"/i', $embed, $array ) ;
list($pre, $pb) = split("pb=", $array[1]);

if($pb == "" || strpos($pb, "!") === false)
    die(json_encode(array("success"=>false)));

//echo "PB Extracted:<br>";
//echo $pb;
///echo "<br><br>Decode:<br/>";
$pb_array = explode('!', $pb);

$coords = array();
$address;$addressHex;
$results = array();
foreach($pb_array as $key => $value){
    //uncomment to debug output
    //echo "$key - $value<br/>";

    if($value == "3m2" || $value == "2m2"){
        //3m2 seems to be the divider of these 'places'
        if(count($coords) != 3) //don't add the center of map data (3 coordinates [height, lng, lat])
            array_push($results, array("coords"=>$coords,"address"=>$address,"addressHex"=>$addressHex));
        $coords = array(); //reset array
    }else{
        $type = substr($value, 1, 1);
        $stype = substr($value, 0, 2);
        $value = substr($value, 2);
        //echo "$type - $value<br/>";
        if($type == "d"){
            //Found Lat,Lng
            array_push($coords, $value);
        }else if($stype == "2s"){
            //Address
            $address = $value;
        }else if($stype == "1s"){
            //Address Encoded in some way
            $addressHex = $value;
        }
    }
}

//echo "<br><br>Google Result<br/>";
//echo json_encode($results);


//echo "<br><br>Mapbox API:<br/>";
$waypoints = array();
for($i=0;$i<count($results);$i++){
    if(count($results[$i]["coords"])){
        $lat = $results[$i]["coords"][0];
        $lng = $results[$i]["coords"][1];
        array_push($waypoints, "$lng%2C$lat");
    }
}
$waypoints = implode("%3B", $waypoints); //convert to string

$mapbox_api_key = "pk.eyJ1I.....";
$url = "https://api.mapbox.com/directions/v5/mapbox/driving/$waypoints.json?steps=false&alternatives=false&overview=full&geometries=geojson&access_token=$mapbox_api_key";

//echo "<br><br>Mapbox Response:<br/>";
$response = file_get_contents($url); 

$json = json_decode($response,true);
//echo "<br><br>Mapbox Geometry:<br/>";
$coordinates = $json["routes"][0]["geometry"]["coordinates"];

$geojson = (array("type"=>"FeatureCollection","features"=>array(array("type"=>"Feature","geometry"=>array("type"=>"LineString","coordinates"=>$coordinates),"properties"=>array()))));

echo json_encode(array("success"=>true, "geojson"=>$geojson));