Android Native Development Kit

In order to use SensiML Knowledge Packs with Android, the application must be configured to use the Android NDK.

You can Create or import a native project to do this. For this guide, we will show you how to integrate the library version of a Knowledge Pack or the Source Code version of a Knowledge Pack.

Using A Library Download

When you download a model, you will get a zip file with a structure that looks similar to this:

knowledgepack-libs
   ├── model.json
   ├── include
       ├── kb.h
       ├── kb_defines.h
       ├── kb_typedefs.h
       └── kb_debug.h
   ├── src
        ├── arm64-v8a
            └── libsensiml.so
        ├── armeabi-v7a
            └── libsensiml.so
        ├── x86
            └── libsensiml.so
        ├── x86_64
            └── libsensiml.so
knowledgepack_project
        └── recognition_run
            ├── sml_recognition_run.cpp
            ├── CMakeLists.txt
            └── README.md

Copy the Library Files

You can add the knowledgepack-libs folder to your application folder, for instance app/knowledgepack-libs

Copy and rename the wrapper functions

Copy the knowledgepack_project/recognition_run/sml_recognition_run.cpp file to the app/src/main/cpp directory. This file may need changes, as it is set up in such a way that it will be exported to an activity named KnowledgePackActivity. If your application does not have a KnowledgePackActivity, rename the JNIExport calls:

extern "C"
JNIEXPORT jboolean JNICALL
Java_com_sensiml_knowledgepackrunner_KnowledgePackActivity_usingTestData(JNIEnv *env,
                                                                    jobject call_object){
#ifdef USE_TEST_RAW_SAMPLES
    return true;
#else
    return false;
#endif
}

// RENAMED VERSION HERE:

extern "C"
JNIEXPORT jboolean JNICALL
Java_com_sensiml_knowledgepackrunner_YOURRECOGNITIONACTIVITY_usingTestData(JNIEnv *env,
                                                                    jobject call_object){
#ifdef USE_TEST_RAW_SAMPLES
    return true;
#else
    return false;
#endif
}

This will need to be done for all of the calls.

Update CMakeLists.txt

The Android NDK uses CMake to compile and link native code to java. When you generate a SensiML Knowledge Pack for Android NDK we will automatically generate a CMakeLists.txt file in your knowledgepack_projectrecognition_run directory that you can use to compile in your application.

If you already have a CMakeLists.txt file that you are using in your project then you will need to add the following relevant content to your file:

set(kp_lib_DIR ${CMAKE_SOURCE_DIR}/../../../knowledgepack-libs)
set(kp_inc_DIR ${kp_lib_DIR}/include)
# Creates and names a library, sets it as either STATIC
# or SHARED, and provides the relative paths to its source code.
# You can define multiple libraries, and CMake builds them for you.
# Gradle automatically packages shared libraries with your APK.

add_library(
        sensiml
        SHARED
        IMPORTED
)
set_target_properties(sensiml
        PROPERTIES IMPORTED_LOCATION
        ${kp_lib_DIR}/src/${CMAKE_ANDROID_ARCH_ABI}/libsensiml.so)
include_directories(${kp_inc_DIR})

# Searches for a specified prebuilt library and stores the path as a
# variable. Because CMake includes system libraries in the search path by
# default, you only need to specify the name of the public NDK library
# you want to add. CMake verifies that the library exists before
# completing its build.

add_library( # Sets the name of the library.
        knowledgepack-wrapper

        # Sets the library as a shared library.
        SHARED

        # Provides a relative path to your source file(s).
        sensiml_recognition_run.cpp)

target_include_directories(knowledgepack-wrapper
        PRIVATE
        ${kp_inc_DIR})

target_link_libraries( # Specifies the target library.

        # YOUR OTHER LIBRARIES GO HERE
        knowledgepack-wrapper
        android
        sensiml

        # Links the target library to the log library
        # included in the NDK.
        log)

Using Source Code Download

When using the source code download option, you will also get the precompiled libraries, as well as the model source. The above steps for sml_recognition_run.cpp still apply. You will now see a directory structure like this:

knowledgepack-libs
   ├── model.json
   ├── include
       ├── kb.h
       ├── kb_defines.h
       ├── kb_typedefs.h
       └── kb_debug.h
   ├── src
        ├── arm64-v8a
            └── libsensiml.so
        ├── armeabi-v7a
            └── libsensiml.so
        ├── x86
            └── libsensiml.so
        ├── x86_64
            └── libsensiml.so
knowledgepack_project
        └── recognition_run
            ├── sml_recognition_run.cpp
            ├── CMakeLists.txt
            └── README.md
libsensiml
    └── cpp
        ├── model source files
        └── model header files

Copy the libsensiml/cpp folder to the directory of your choice. Here, we will use MAIN_DIR/libsensiml/cpp

Update CMakeLists.txt

The Android NDK uses CMake to compile and link native code to java. When you generate a SensiML Knowledge Pack for Android NDK we will automatically generate a CMakeLists.txt file in your knowledgepack_projectrecognition_run directory that you can use to compile in your application.

If you already have a CMakeLists.txt file that you are using in your project then you will need to add the following relevant content to your file:

The CMakeLists.txt file changes slightly from that of the library version. Instead of pulling in pre-compiled .so files, the source code will directly be built first.

set(KNOWLEDGEPACK_MODEL_SOURCE_DIR ${CMAKE_SOURCE_DIR}/../../../libsensiml/cpp)
file(GLOB KNOWLEDGEPACK_SRCS ${KNOWLEDGEPACK_MODEL_SOURCE_DIR}/ *.c)
# Creates and names a library, sets it as either STATIC
# or SHARED, and provides the relative paths to its source code.
# You can define multiple libraries, and CMake builds them for you.
# Gradle automatically packages shared libraries with your APK.

add_library(
        sensiml
        SHARED
        ${KNOWLEDGEPACK_SRCS}
)
find_library( # Sets the name of the path variable.
          log-lib

          # Specifies the name of the NDK library that
          # you want CMake to locate.
          log )
include_directories(${KNOWLEDGEPACK_MODEL_SOURCE_DIR})

# Searches for a specified prebuilt library and stores the path as a
# variable. Because CMake includes system libraries in the search path by
# default, you only need to specify the name of the public NDK library
# you want to add. CMake verifies that the library exists before
# completing its build.

add_library( # Sets the name of the library.
        knowledgepack-wrapper

        # Sets the library as a shared library.
        SHARED

        # Provides a relative path to your source file(s).
        sensiml_recognition_run.cpp)

target_include_directories(knowledgepack-wrapper
        PRIVATE
        ${kp_inc_DIR})

target_link_libraries( # Specifies the target library.

        # YOUR OTHER LIBRARIES GO HERE
        knowledgepack-wrapper
        android
        sensiml

        # Links the target library to the log library
        # included in the NDK.
        log)

Integrating the Knowledge Pack in your Application

README.md Is provided in the `knowledgepack_project/recognition_run` folder of a download. This file contains integration instructions for Kotlin and Java langagues, ensuring that the NDK library components and expected calls to the SensiML models will be set up properly.

Initializing the Models

The models will need to be initialized just once. The initSensiMLKnowledgePackModels() call will do that. This call should be placed in the Activity creation where data will be processed.