Currently, building the javadoc for libjava throws the following warnings:

.../mapbox-java/libjava/lib/src/main/java/com/mapbox/services/directions/v4/models/FeatureProperties.java:16: warning - Tag @link: reference not found: DirectionFeature
.../mapbox-java/libjava/lib/src/main/java/com/mapbox/services/directions/v4/models/FeatureProperties.java:16: warning - Tag @link: reference not found: DirectionFeature
.../mapbox-java/libjava/lib/src/main/java/com/mapbox/services/directions/v5/MapboxDirections.java:184: warning - @return tag has no arguments.
.../mapbox-java/libjava/lib/src/main/java/com/mapbox/services/directions/v5/MapboxDirections.java:197: warning - @return tag has no arguments.
.../mapbox-java/libjava/lib/src/main/java/com/mapbox/services/directions/v5/MapboxDirections.java:219: warning - @return tag has no arguments.
.../mapbox-java/libjava/lib/src/main/java/com/mapbox/services/directions/v4/models/FeatureProperties.java:16: warning - Tag @link: reference not found: DirectionFeature
6 warnings

Could you look into them, @cammace ?

Review https://github.com/mapbox/geometry.hpp and study approaches to make API and functionality consistent across both projects.

This will allow for POIs to be filtered by a given region.

https://github.com/Turfjs/turf/tree/master/packages/turf-within

Geocoder now includes wikidata information. MAS needs to include getter and setter methods. API-documentation

cc: @zugaldia

To run the test app accompanying libandroid you need to add a dev.xml file with the mapbox key. Maybe it's nice to add a simple page to the wiki with the steps involved. Right now, the app starts and crashes.

The Mapbox SDK is introducing a MapboxAccountManager as a centralized place for developers to configure their apps to use Mapbox technology on Android. We want to enable this as well in MAS to handle token management, for ease of use and consistency.

/cc: @bleege

New version v5 of Mapbox Directions is still unreleased. This ticket tracks its implementation when the time comes.

Would make navigation in github on long run a bit easier:

Example labels:

• static-api
• geocoder
• ..

I'm using GeocoderAutoCompleteView to get nearest POI with respect to proximity but its strange that I'm not getting results of nearby places. I'm from Nepal, currently living in Kathmandu.

autocomplete.setType(GeocodingCriteria.TYPE_POI);
autocomplete.setProximity(Position.fromCoordinates(
                27.6796678, 85.3490278));

Google's geocoding works fine in the same case. Why not mapbox's?

Because the rules to build this branch are different from master-derived branches. We should either have a specific rule, or remove it from the build process.

Likely on the README.md file.

Hello, the API is working fine for me on emulators. But on the physical device, it crashes and return.
java.lang.NullPointerException: Attempt to invoke virtual method 'java.util.List com.mapbox.services.directions.v4.models.DirectionsResponse.getRoutes()' on a null object reference at net.fjos.distancefinder.MainActivity.finnDistanse(MainActivity.java:158) at net.fjos.distancefinder.MainActivity$1.onClick(MainActivity.java:101) at android.view.View.performClick(View.java:5204) at android.view.View$PerformClick.run(View.java:21155) at android.os.Handler.handleCallback(Handler.java:739) at android.os.Handler.dispatchMessage(Handler.java:95) at android.os.Looper.loop(Looper.java:148) at android.app.ActivityThread.main(ActivityThread.java:5422) at java.lang.reflect.Method.invoke(Native Method) at com.android.internal.os.ZygoteInit$MethodAndArgsCaller.run(ZygoteInit.java:726) at com.android.internal.os.ZygoteInit.main(ZygoteInit.java:616)

Background

This repository (mapbox-java) contains two new artifacts, mapbox-java-services and mapbox-android-services which are the result of combining the original mapbox-geocoder-android and mapbox-directions-android projects as discussed in https://github.com/mapbox/gl-internal/issues/387#issuecomment-200399195.

New features

Besides combining our previous libraries, the codebase now has a number of additional features:

• Separates the pure Java code (e.g. for server use) from the Android-specific code.
• Includes support for v5 of the upcoming directions API.
• Supports the Mapbox Static API.
• Supports RxJava for all geocoding and directions requests in addition to the classic Call mechanism.
• Includes a lightweight implementation for Java GeoJSON objects.
• Includes additional tests.

Release checklist

• Documentation
  • Complete Javadoc for classes @cammace @danswick
  • Extract sample Android code from the TestApp for the website @cammace @danswick
  • Prepare CHANGELOG.md @zugaldia
  • Decide the final location for Javadoc and samples in https://www.mapbox.com/android-sdk/ @danswick @cammace #9
• Release a version of Sirius that uses mapbox-android-services without legacy libraries @bleege @zugaldia https://github.com/mapbox/sirius-android/issues/276
• General code review @tobrun @bleege
• Comms: Blog post et al. @Lndsay @zugaldia https://github.com/mapbox/blog/pull/1107
• Make this repo public.
• Publish artifacts to Maven Central @zugaldia
• Add deprecation notice in previous repos.

After-release checklist

• Include Java as a new language in Mapbox's new API Documentation site https://github.com/mapbox/api-documentation/issues/138 @tmcw @cammace @danswick

/cc: @1ec5 @bsudekum @twbell @lxbarth @gundersen

The service object is always null when MapboxGeocoding.getObservable() is called which results in a NullPointerException. This is due to the the internal service object being referenced directly instead of calling getService() which is what's responsible for the object's creation.

java.lang.NullPointerException: Attempt to invoke interface method 'rx.Observable com.mapbox.services.geocoding.v5.GeocodingService.getObservable(java.lang.String, java.lang.String, java.lang.String, java.lang.String, java.lang.String)' on a null object reference
    at com.mapbox.services.geocoding.v5.MapboxGeocoding.getObservable(MapboxGeocoding.java:96)

/cc @zugaldia

Releasing mapbox-java soon, we need to figure out the best place for the overview, examples, and javadoc. Pros and cons of the two options:

Incorporate it within mapbox.com/android-sdk
Pros:

• Better visibility to those who are already using Android SDK
• Easier to maintain

Cons:

• API tab immediately takes you to Android SDK Javadoc, would need option to pick which javadoc to view
• More clutter on each page
• Overview page would need separate page for each SDK

Make separate mapbox.com/mapbox-java page
Pros:

• API tab can take you directly to Mapbox-java javadoc
• Less cutter overall

Cons:

• Less visibility due to it being on separate page
• Harder to maintain

Thoughts? @danswick @lyzidiamond @tmcw @zugaldia @bleege @tobrun

We're currently trying different snapshots and it's difficult for consumers to know for sure the version they're targeting.

Currently the image dimensions are only validated on null but not the correct range. With some image loading libraries this results in the image just not appearing and no obvious error being logged. (Maybe because the api returns a status code 422, which is not that common).

Mapbox SDKs shouldn't contain Mapbox's brand colors from the Base Styleguide as they're for Mapbox projects and branding. While it's totally cool for apps and Web sites to make use of them on their own, the SDKs shouldn't provide them as a supported resource as "they could change at any minute and include some styles specific to our brand."

I'm gonna start the process to set up the javadoc builds for both projects. As discussed in #9 (comment) we'll host them on a gh-pages branch in this repo.

sample White House feature json:

"id": "address.7626097581139850", "type": "Feature", "text": "Pennsylvania Ave NW", "place_name": "1600 Pennsylvania Ave NW, Washington, District of Columbia 20006, United States", "relevance": 0.99, "properties": {}, "center": [ -77.036543, 38.897702 ], "geometry": { "type": "Point", "coordinates": [ -77.036543, 38.897702 ] }, "address": "1600", "context": [ { "id": "neighborhood.7130293780113160", "text": "Franklin Mcpherson Square" }, { "id": "place.62020", "text": "Washington" }, { "id": "postcode.8344092665816130", "text": "20006" }, { "id": "region.6884744206035790", "text": "District of Columbia" }, { "id": "country.12862386939497690", "text": "United States", "short_code": "us" } ]

This includes address parameter and relevance which have no getter/setters.

MAS 1.0.0 was officially released on 19-April-2016 ( #3 ), however 1.0.0-SNAPSHOTS continue to be built. Let's bump this to a new version number (doesn't have to be final) like 1.1.0-SNAPSHOT so that new work is visibly seen to be different from the 1.0.0 final release.

/cc @zugaldia @tobrun @cammace @danswick

We need to update the styles found here to use v9 instead. Also deprecate emerald and add outdoor.

As the current validations that throw the Service Exception are mostly programming errors, it makes sense to make ServiceException an unchecked exception to avoid unneeded boilerplate in the consuming code.

One exception (no pun intended to this) is the the image size validation as these could potentially be more data driven. Although not checking for 0/negative integers on the caller side might as well be considered a programming error.

Now that #70 has landed, we need to add a sample activity to the TestApp. Something like http://mourner.github.io/simplify-js/.

/cc: @cammace

In the case where the geometry query parameter is set to false the response contains an empty object for the geometry geometry={} where I would expect null. For now I worked around it, but good to check if this is by design or not.

This library should set the User-Agent HTTP header on requests it sends, so that Mapbox Studio’s Stats page can identify this library under the “Browsers” column instead of saying “Other”.

/cc @zugaldia

On every build I'm receiving emails on build success/error. Could we disable these in same way as we are doing for gl-native? Or maybe only email the person who made the faulty commit?

trying getAddress() like so:

response.body().getFeatures().get(i).getAddress()

returns a null or empty string. Doesn't look like the API has this field anymore.

cc: @zugaldia

In the same way that we provide API compatibility for the stock Android geocoder, we could consider bringing compatibility for the https://developers.google.com/maps/documentation/directions/ API.

Use Paris as a target for the sample code as it should be an item with a complete response.

Related: #13

/cc: @cammace

Yes.

/cc: @bleege @1ec5

We should following the same technical approach that we decide in mapbox/maki#247.

Current icons:

• Preferred: https://www.mapbox.com/base/styling/icons/ aka https://github.com/mapbox/base/tree/mb-pages/latest/icons
• https://github.com/mapbox/mapbox-directions.js/tree/mb-pages/dist

/cc: @samanpwbb @1ec5 @bsudekum

Hi,

I just wanted to let you know, that the current version of your library is not working for many foreign devices, e.g. German, French, ..., on which a comma instead of a dot is used for decimal separation.
The line in question is probably MapboxDirections.java:262 resulting in
https://api.mapbox.com/v4/directions/mapbox.driving/-77.047237,38.892255;-77.019759,38.903568.json... for English / American devices and in https://api.mapbox.com/v4/directions/mapbox.driving/-77,047237,38,892255;-77,019759,38.903568.json... for e.g. German devices.
The latter request is failing, BTW.

See: mapbox/mapbox-gl-native#4802

cc: @1ec5

It's been almost a month since 1.0 #3 and during this time we've included a few new features (and bug fixes), it's time to publish an updated release.

New features

• Support for directions API v5 #31
• Support for Maki 3 and directions icons in Test App #1
• Removed extra dependencies to reduce the method count #44
• Improved documentation #20 #28
• Fixed a bug where locale was different from Locale.US #39 #53

Full list: https://github.com/mapbox/mapbox-java/issues?q=milestone%3Av1.1.0+is%3Aclosed

Release checklist

• Update Javadoc in http://mapbox.github.io/mapbox-java @zugaldia
• Update install instructions in README.md #29 @cammace
• Update install instructions in https://www.mapbox.com/android-sdk/#mapbox_android_services @cammace
• Update samples in https://www.mapbox.com/android-sdk/examples/ @cammace
• Update CHANGELOG.md @zugaldia
• Comms: Twitter @Lndsay @zugaldia

libjava

• Change version name to 1.1.0 in gradle.properties and build.gradle
• Publish 1.1.0 artifact to Maven Central
• Bump and change version name to 1.2.0-SNAPSHOT

libandroid

• Change version name to 1.1.0 in gradle.properties and build.gradle
• Update libjava dependency to 1.1.0 in build.gradle
• Publish 1.1.0 artifact to Maven Central
• Tag v1.1.0 in GitHub
• Close milestone v1.1.0
• Open new milestone v1.2.0
• Bump and change version name to 1.2.0-SNAPSHOT
• Update methodscount badge in README.md

/cc: @mapbox/mobile

When copy pasting example from blog

    MapboxStaticImage staticImage = new MapboxStaticImage.Builder()
                .setAccessToken(getString(R.string.mapbox_access_token))
                .setStyleId(SyncStateContract.Constants.MAPBOX_STYLE_SATELLITE)
                .setLon(lon).setLat(lat)
                .setZoom(15).setBearing(45).setPitch(60)
                .setWidth(500).setHeight(500)
                .setRetina(true)
                .build();

Android Studio is automatically importing an Androd SDK dependency instead of this repo.

I'm suggesting renaming the class/package to something different?

We have been using the mapbox-gl-native for our main apps and our now integrating the static map api. It would be great if we can use the same LatLng location objects here to keep consistency. Especially as we built out our support methods based on the LatLng and like to re-use those.

This code is also used in the GL SDK. It would make sense making this a separate class in the commons package. If we at some point going to include this repo in the GL SDK we could remove some duplicate code.

Reference implementation: https://github.com/kaushikgopal/RxJava-Android-Samples/blob/master/app/src/main/java/com/morihacky/android/rxjava/fragments/DebounceSearchEmitterFragment.java

Based on https://github.com/JakeWharton/RxBinding/blob/master/rxbinding/src/main/java/com/jakewharton/rxbinding/widget/RxAutoCompleteTextView.java.

We've got a few failing tests that weren't being run:

:lib:test

com.mapbox.services.geojson.FeatureCollectionTest > fromJson FAILED
    com.google.gson.JsonParseException at FeatureCollectionTest.java:14
        Caused by: java.lang.ClassNotFoundException at FeatureCollectionTest.java:14

com.mapbox.services.geojson.FeatureCollectionTest > toJson FAILED
    com.google.gson.JsonParseException at FeatureCollectionTest.java:27
        Caused by: java.lang.ClassNotFoundException at FeatureCollectionTest.java:27

com.mapbox.services.geojson.FeatureTest > fromJson FAILED
    com.google.gson.JsonParseException at FeatureTest.java:14
        Caused by: java.lang.ClassNotFoundException at FeatureTest.java:14

com.mapbox.services.geojson.FeatureTest > toJson FAILED
    com.google.gson.JsonParseException at FeatureTest.java:22
        Caused by: java.lang.ClassNotFoundException at FeatureTest.java:22

com.mapbox.services.geojson.GeometryCollectionTest > fromJson FAILED
    com.google.gson.JsonParseException at GeometryCollectionTest.java:14
        Caused by: java.lang.ClassNotFoundException at GeometryCollectionTest.java:14

com.mapbox.services.geojson.GeometryCollectionTest > toJson FAILED
    com.google.gson.JsonParseException at GeometryCollectionTest.java:22
        Caused by: java.lang.ClassNotFoundException at GeometryCollectionTest.java:22

60 tests completed, 6 failed

Hello,

I realized an application like Waze (GPS Nagivation) with MapBox

But with some devices i have a Out Of Memory like this :

05-10 15:40:52.744: W/Adreno-GSL(11228): <sharedmem_gpumem_alloc_id:1498>: sharedmem_gpumem_alloc: mmap failed errno 12 Out of memory
05-10 15:40:52.744: E/Adreno-GSL(11228): <gsl_memory_alloc_pure:2044>: GSL MEM ERROR: kgsl_sharedmem_alloc ioctl failed.
05-10 15:40:52.798: W/Adreno-GSL(11228): <sharedmem_gpumem_alloc_id:1498>: sharedmem_gpumem_alloc: mmap failed errno 12 Out of memory
05-10 15:40:52.798: E/Adreno-GSL(11228): <gsl_memory_alloc_pure:2044>: GSL MEM ERROR: kgsl_sharedmem_alloc ioctl failed.
05-10 15:40:52.799: W/Adreno-GSL(11228): <sharedmem_gpumem_alloc_id:1498>: sharedmem_gpumem_alloc: mmap failed errno 12 Out of memory
05-10 15:40:52.799: E/Adreno-GSL(11228): <gsl_memory_alloc_pure:2044>: GSL MEM ERROR: kgsl_sharedmem_alloc ioctl failed.
05-10 15:40:52.799: W/Adreno-GSL(11228): <sharedmem_gpumem_alloc_id:1498>: sharedmem_gpumem_alloc: mmap failed errno 12 Out of memory
05-10 15:40:52.799: E/Adreno-GSL(11228): <gsl_memory_alloc_pure:2044>: GSL MEM ERROR: kgsl_sharedmem_alloc ioctl failed.
05-10 15:40:52.800: W/Adreno-GSL(11228): <sharedmem_gpumem_alloc_id:1498>: sharedmem_gpumem_alloc: mmap failed errno 12 Out of memory
05-10 15:40:52.800: E/Adreno-GSL(11228): <gsl_memory_alloc_pure:2044>: GSL MEM ERROR: kgsl_sharedmem_alloc ioctl failed.
05-10 15:40:52.800: W/Adreno-GSL(11228): <sharedmem_gpumem_alloc_id:1498>: sharedmem_gpumem_alloc: mmap failed errno 12 Out of memory
05-10 15:40:52.800: E/Adreno-GSL(11228): <gsl_memory_alloc_pure:2044>: GSL MEM ERROR: kgsl_sharedmem_alloc ioctl failed.
05-10 15:40:52.811: W/Adreno-ES20(11228): <core_glBufferData:1175>: GL_OUT_OF_MEMORY
05-10 15:40:52.811: W/Adreno-GSL(11228): <sharedmem_gpumem_alloc_id:1498>: sharedmem_gpumem_alloc: mmap failed errno 12 Out of memory
05-10 15:40:52.811: E/Adreno-GSL(11228): <gsl_memory_alloc_pure:2044>: GSL MEM ERROR: kgsl_sharedmem_alloc ioctl failed.
05-10 15:40:52.811: W/Adreno-GSL(11228): <sharedmem_gpumem_alloc_id:1498>: sharedmem_gpumem_alloc: mmap failed errno 12 Out of memory
05-10 15:40:52.811: E/Adreno-GSL(11228): <gsl_memory_alloc_pure:2044>: GSL MEM ERROR: kgsl_sharedmem_alloc ioctl failed.
05-10 15:40:52.812: A/libc(11228): Fatal signal 11 (SIGSEGV), code 1, fault addr 0x54 in tid 12921 (Map Thread)
05-10 15:41:15.692: W/libc(11228): pthread_create failed: couldn't allocate 1040384-byte stack: Out of memory
05-10 15:41:15.692: E/libutils.threads(11228): androidCreateRawThreadEtc failed (entry=0xb6d24a95, res=11, errno=0)
05-10 15:41:15.692: E/libutils.threads(11228): (android threadPriority=0)

this occurs after a few minutes of using app.

For information, my application runs with MapBox 3.2.0,

I wonder if people have had this worries,

Where does this problem come from ?

It has no tests currently.

While converting Sirius to use the new Mapbox Android Services library I noticed that for Directions the previous V4 had a setOrigin(Waypoint origin) and setDestination(Waypoint destination) methods while the new V5 relies simply on setCoordinates(Position[] coordinates). @zugaldia explained that this is due to the underlying Mapbox Direction Data API, but that we could consider re-adding these convenience methods. I think that makes sense to do so as to me it seems like a fundamental part of a developer's mental model when they go to add Directions in their app. It'd also make upgrading from V4 to V5 much easier / quicker.

/cc @zugaldia @1ec5

In this ApplicationTest class from the libjava module there are imports from the android sdk.
I think we can just remove this class since instrumentation tests are not needed for the libjava.

We need to add them to all packages in both projects to help javadoc generate the right descriptions, as explained in http://docs.oracle.com/javase/specs/jls/se7/html/jls-7.html#jls-7.4.

/cc: @cammace

For consistency with the main SDK: mapbox/mapbox-gl-native#4112.

In order to make the returned images better cacheable on the client, it would be nice to set the precision in decimals on the Builder instead of manually rounding the latitude/longitude.

We need to add details on the required proguard settings.

Next steps:

• Add settings @zugaldia
• Document @cammace

/cc: @bleege