Text summarization guide for Android

The MediaPipe Text Summarizer task lets you identify the most important information in a text and generate a shorter version while maintaining the original context meaning. These instructions show you how to use the Text Summarizer in an Android app.

For more information about the capabilities, models, and configuration options of this task, see the Overview.

Code example

The Text Summarizer Android example app demonstrates the API on a physical Android device or emulator.

You can use the app as a starting point for your own Android app, or refer to it when modifying an existing app. The Text Summarizer example code is hosted on GitHub.

Setup

This section describes key steps for setting up your development environment and code projects specifically to use Text Summarizer. For general information on setting up your development environment for using MediaPipe tasks, including platform version requirements, see the Setup guide for Android.

Attention: This MediaPipe Solutions Preview is an early release. Learn more.

Dependencies

Text Summarizer uses the com.google.mediapipe:tasks-text libraries. Add this dependency to the build.gradle file of your Android app development project. You can import the required dependencies with the following code:

dependencies {
    implementation 'com.google.mediapipe:tasks-text:latest.release'
}

Model

The MediaPipe Text Summarizer task requires a trained model that is compatible with this task. For more information on available trained models for Text Summarizer, see the task overview Models section.

Select and download the model, and then store it within your project directory:

<dev-project-root>/src/main/assets

Note: The Android build system automatically checks the project directory for file resources.

Specify the model within the options object using either a path or a file descriptor. Use the setModelPath() function to specify the path used by the model, or use the setModelAssetFileDescriptor() function to specify the file descriptor of the model asset.

Create the task

You can use the createFromOptions() function to create the task. The createFromOptions() function accepts configuration options to set the task options. For more information on configuration options, see Configuration options.

The following code demonstrates how to build and configure this task.

Model Path

// Configure the options using model path
TextSummarizerOptions options = TextSummarizerOptions.builder()
    .setModelPath("path/to/your/model.litertlm")
    .setMode(TextSummarizerOptions.Mode.TLDR)
    .build();

File Descriptor

// Configure the options using file descriptor
// Note: 'parcelFileDescriptor' is the ParcelFileDescriptor of
// the model asset
TextSummarizerOptions options = TextSummarizerOptions.builder()
    .setModelAssetFileDescriptor(parcelFileDescriptor)
    .setMode(TextSummarizerOptions.Mode.TLDR)
    .build();

Create the engine from options:

// Create the engine from options
// Note: 'context' is the Android Context
TextSummarizer summarizer = TextSummarizer.createFromOptions(context, options);

Configuration options

This task has the following configuration options for Android apps:

Option Name	Description	Value Range	Default Value
setMode	The summarization mode of the text summarizer task. Can be a short summary paragraph or bulleted list of key points.	TextSummarizerOptions.Mode.TLDR, TextSummarizerOptions.Mode.KEYPOINTS	TextSummarizerOptions.Mode.KEYPOINTS
setMaxNumTokens	The maximum number of tokens for summarization tasks. If set, the summarization will be truncated if the input and output exceed this value. If not set, then the default limit is decided by the model capacity.	Integer	Model default (8k)
setModelPath	The model path to a model asset file in the Android app assets folder.	File path as a string	Not set
setModelAssetFileDescriptor	The file descriptor integer of a model asset file.	Integer specifying the file descriptor	Not set

Prepare data

Text Summarizer accepts text (String) data and doesn't require any preparations.

String inputText = "Some input text for the summarization task";

Run the task

To summarize text, you can use the synchronous approach which blocks the execution thread until the summary is returned, or the streaming approach which returns immediately and streams the result content through a callback interface.

Option A: Synchronous Summarization

Use the synchronous approach when you want to wait for the entire text summary to be generated and returned at once.

import android.util.Log;
import com.google.mediapipe.tasks.text.textsummarizer.TextSummarizerResult;

String textToSummarize = "Here is a long meeting transcript or block of text...";

// Run summarization (blocks current thread until done)
TextSummarizerResult result = summarizer.summarize(textToSummarize);
String summary = result.getSummary();
Log.i("TextSummarizer", "Summary: " + summary);

Option B: Streaming Summarization

The streaming approach is recommended for longer generation tasks or maintaining UI responsiveness. Here, a callback is provided to receive partial chunks of the summary text as they are computed.

import android.util.Log;
import com.google.mediapipe.tasks.text.textsummarizer.TextSummarizerStreamingResult;

String textToSummarize = "Here is a long meeting transcript or block of text...";
// Starts summarization and returns immediately. Iteratively invokes callback.
summarizer.summarizeStreaming(textToSummarize, new TextSummarizer.SummarizationResultCallback() {
     @Override
     public void onNext(TextSummarizerStreamingResult result) {
         // Appends the next stream chunk
         String chunk = result.getChunk();

         // e.g., runOnUiThread(() -> resultTextView.append(chunk));
     }

     @Override
     public void onError(Throwable throwable) {
         // Handle summarization error
         Log.e("TextSummarizer", "Summarization failed: " + throwable.getMessage());
     }

     @Override
     public void onDone() {
         // Fired when summarization completes
         Log.i("TextSummarizer", "Summarization finished!");
     }
});

Handle and display results

On task execution:

• For synchronous summarization, the task returns a TextSummarizerResult containing the final summary text.
• For streaming summarization, the task outputs TextSummarizerStreamingResult chunks repeatedly to the callback interface.

Clean up

To free up memory and C++ resources, you must explicitly close the task when done using it.

@Override
protected void onDestroy() {
    super.onDestroy();
    if (summarizer != null) {
        summarizer.close(); // Cleans up underlying handle
    }
}