auto-save 2026-04-01 09:03 (+8, ~2)

android-ocr-service/app/build.gradle.kts

@@ -0,0 +1,43 @@
+plugins {
+    id("com.android.application")
+    id("org.jetbrains.kotlin.android")
+}
+
+android {
+    namespace = "com.guiagent.ocr"
+    compileSdk = 31
+
+    defaultConfig {
+        applicationId = "com.guiagent.ocr"
+        minSdk = 26
+        targetSdk = 31
+        versionCode = 1
+        versionName = "1.0"
+    }
+
+    buildTypes {
+        release {
+            isMinifyEnabled = false
+        }
+    }
+
+    compileOptions {
+        sourceCompatibility = JavaVersion.VERSION_1_8
+        targetCompatibility = JavaVersion.VERSION_1_8
+    }
+
+    kotlinOptions {
+        jvmTarget = "1.8"
+    }
+}
+
+dependencies {
+    // ML Kit Text Recognition - bundled model (no GMS needed!)
+    implementation("com.google.mlkit:text-recognition-chinese:16.0.0")
+
+    // HTTP server
+    implementation("org.nanohttpd:nanohttpd:2.3.1")
+
+    // JSON
+    implementation("com.google.code.gson:gson:2.10.1")
+}

android-ocr-service/app/src/main/AndroidManifest.xml

@@ -0,0 +1,28 @@
+<?xml version="1.0" encoding="utf-8"?>
+<manifest xmlns:android="http://schemas.android.com/apk/res/android">
+
+    <uses-permission android:name="android.permission.READ_EXTERNAL_STORAGE"/>
+    <uses-permission android:name="android.permission.INTERNET"/>
+    <uses-permission android:name="android.permission.FOREGROUND_SERVICE"/>
+
+    <application
+        android:allowBackup="false"
+        android:label="OCR Service"
+        android:supportsRtl="true">
+
+        <activity
+            android:name=".MainActivity"
+            android:exported="true">
+            <intent-filter>
+                <action android:name="android.intent.action.MAIN"/>
+                <category android:name="android.intent.category.LAUNCHER"/>
+            </intent-filter>
+        </activity>
+
+        <service
+            android:name=".OcrService"
+            android:exported="true"
+            android:foregroundServiceType="dataSync"/>
+
+    </application>
+</manifest>

android-ocr-service/app/src/main/java/com/guiagent/ocr/MainActivity.kt

@@ -0,0 +1,23 @@
+package com.guiagent.ocr
+
+import android.app.Activity
+import android.content.Intent
+import android.os.Bundle
+import android.widget.TextView
+
+class MainActivity : Activity() {
+    override fun onCreate(savedInstanceState: Bundle?) {
+        super.onCreate(savedInstanceState)
+        val tv = TextView(this).apply {
+            text = "OCR Service\nPort: 18900\nStarting..."
+            textSize = 20f
+            setPadding(40, 40, 40, 40)
+        }
+        setContentView(tv)
+
+        // Start the service
+        val intent = Intent(this, OcrService::class.java)
+        startForegroundService(intent)
+        tv.text = "OCR Service\nPort: 18900\nRunning!"
+    }
+}

android-ocr-service/app/src/main/java/com/guiagent/ocr/OcrEngine.kt

@@ -0,0 +1,79 @@
+package com.guiagent.ocr
+
+import android.graphics.Bitmap
+import android.graphics.BitmapFactory
+import com.google.mlkit.vision.common.InputImage
+import com.google.mlkit.vision.text.TextRecognition
+import com.google.mlkit.vision.text.chinese.ChineseTextRecognizerOptions
+import java.io.File
+import java.util.concurrent.CountDownLatch
+import java.util.concurrent.TimeUnit
+
+data class TextBox(
+    val text: String,
+    val x: Int,
+    val y: Int,
+    val w: Int,
+    val h: Int,
+    val confidence: Float
+) {
+    val cx get() = x + w / 2
+    val cy get() = y + h / 2
+}
+
+object OcrEngine {
+
+    private val recognizer by lazy {
+        TextRecognition.getClient(ChineseTextRecognizerOptions.Builder().build())
+    }
+
+    fun recognize(imagePath: String): List<TextBox> {
+        val file = File(imagePath)
+        if (!file.exists()) return emptyList()
+        val bitmap = BitmapFactory.decodeFile(imagePath) ?: return emptyList()
+        return recognizeBitmap(bitmap)
+    }
+
+    /** 直接截屏并识别，不落盘 */
+    fun screencapAndRecognize(): List<TextBox> {
+        val process = Runtime.getRuntime().exec("screencap -p")
+        val bytes = process.inputStream.readBytes()
+        process.waitFor()
+        if (bytes.isEmpty()) return emptyList()
+        val bitmap = BitmapFactory.decodeByteArray(bytes, 0, bytes.size) ?: return emptyList()
+        return recognizeBitmap(bitmap)
+    }
+
+    fun recognizeBitmap(bitmap: Bitmap): List<TextBox> {
+        val image = InputImage.fromBitmap(bitmap, 0)
+        val results = mutableListOf<TextBox>()
+        val latch = CountDownLatch(1)
+
+        recognizer.process(image)
+            .addOnSuccessListener { visionText ->
+                for (block in visionText.textBlocks) {
+                    for (line in block.lines) {
+                        val box = line.boundingBox ?: continue
+                        results.add(
+                            TextBox(
+                                text = line.text,
+                                x = box.left,
+                                y = box.top,
+                                w = box.width(),
+                                h = box.height(),
+                                confidence = line.confidence ?: 0.8f
+                            )
+                        )
+                    }
+                }
+                latch.countDown()
+            }
+            .addOnFailureListener {
+                latch.countDown()
+            }
+
+        latch.await(10, TimeUnit.SECONDS)
+        bitmap.recycle()
+        return results
+    }
+}

android-ocr-service/app/src/main/java/com/guiagent/ocr/OcrHttpServer.kt

@@ -0,0 +1,88 @@
+package com.guiagent.ocr
+
+import android.graphics.BitmapFactory
+import com.google.gson.Gson
+import fi.iki.elonen.NanoHTTPD
+import java.io.ByteArrayOutputStream
+
+class OcrHttpServer(port: Int = 18900) : NanoHTTPD(port) {
+
+    private val gson = Gson()
+    private val defaultPath = "/sdcard/ocr_screen.png"
+
+    override fun serve(session: IHTTPSession): Response {
+        return when (session.uri) {
+            "/ocr" -> handleOcr(session)
+            "/snap" -> handleSnap(session)
+            "/health" -> jsonResponse(mapOf("status" to "ok", "engine" to "mlkit-chinese"))
+            else -> newFixedLengthResponse(Response.Status.NOT_FOUND, MIME_PLAINTEXT, "404")
+        }
+    }
+
+    /** 读文件方式 OCR */
+    private fun handleOcr(session: IHTTPSession): Response {
+        val params = session.parms ?: emptyMap()
+        val imagePath = params["path"] ?: defaultPath
+        return doOcr(params["text"]) { OcrEngine.recognize(imagePath) }
+    }
+
+    /** POST 图片数据直接 OCR，不存文件 */
+    private fun handleSnap(session: IHTTPSession): Response {
+        val params = session.parms ?: emptyMap()
+
+        if (session.method == Method.POST) {
+            // NanoHTTPD parseBody 将 binary data 存到临时文件
+            val bodyFiles = HashMap<String, String>()
+            session.parseBody(bodyFiles)
+
+            // postData 键对应临时文件路径
+            val tmpPath = bodyFiles["postData"]
+            if (tmpPath != null) {
+                val imageBytes = java.io.File(tmpPath).readBytes()
+                val bitmap = BitmapFactory.decodeByteArray(imageBytes, 0, imageBytes.size)
+                if (bitmap != null) {
+                    return doOcr(params["text"]) { OcrEngine.recognizeBitmap(bitmap) }
+                }
+                return jsonResponse(mapOf("error" to "decode failed", "size" to imageBytes.size, "count" to 0))
+            }
+            return jsonResponse(mapOf("error" to "no body received", "count" to 0))
+        }
+
+        // GET: 读文件方式 fallback
+        return handleOcr(session)
+    }
+
+    private fun doOcr(query: String?, recognize: () -> List<TextBox>): Response {
+        val startTime = System.currentTimeMillis()
+        var results = recognize()
+
+        if (!query.isNullOrBlank()) {
+            results = results.filter { it.text.contains(query) }
+        }
+
+        val elapsed = System.currentTimeMillis() - startTime
+
+        val response = mapOf(
+            "results" to results.map { box ->
+                mapOf(
+                    "text" to box.text,
+                    "x" to box.x,
+                    "y" to box.y,
+                    "w" to box.w,
+                    "h" to box.h,
+                    "cx" to box.cx,
+                    "cy" to box.cy,
+                    "confidence" to box.confidence
+                )
+            },
+            "count" to results.size,
+            "elapsed_ms" to elapsed
+        )
+        return jsonResponse(response)
+    }
+
+    private fun jsonResponse(data: Any): Response {
+        val json = gson.toJson(data)
+        return newFixedLengthResponse(Response.Status.OK, "application/json", json)
+    }
+}

android-ocr-service/app/src/main/java/com/guiagent/ocr/OcrService.kt

@@ -0,0 +1,49 @@
+package com.guiagent.ocr
+
+import android.app.*
+import android.content.Intent
+import android.os.Build
+import android.os.IBinder
+import android.util.Log
+
+class OcrService : Service() {
+
+    private var server: OcrHttpServer? = null
+    private val TAG = "OcrService"
+    private val PORT = 18900
+
+    override fun onStartCommand(intent: Intent?, flags: Int, startId: Int): Int {
+        startForegroundNotification()
+
+        if (server == null) {
+            server = OcrHttpServer(PORT).also {
+                it.start()
+                Log.i(TAG, "OCR HTTP server started on port $PORT")
+            }
+        }
+        return START_STICKY
+    }
+
+    override fun onDestroy() {
+        server?.stop()
+        server = null
+        Log.i(TAG, "OCR HTTP server stopped")
+        super.onDestroy()
+    }
+
+    override fun onBind(intent: Intent?): IBinder? = null
+
+    private fun startForegroundNotification() {
+        val channelId = "ocr_service"
+        if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.O) {
+            val channel = NotificationChannel(channelId, "OCR Service", NotificationManager.IMPORTANCE_LOW)
+            getSystemService(NotificationManager::class.java).createNotificationChannel(channel)
+        }
+        val notification = Notification.Builder(this, channelId)
+            .setContentTitle("OCR Service")
+            .setContentText("Running on port $PORT")
+            .setSmallIcon(android.R.drawable.ic_menu_camera)
+            .build()
+        startForeground(1, notification)
+    }
+}

android-ocr-service/app/src/main/res/values/strings.xml

@@ -0,0 +1,4 @@
+<?xml version="1.0" encoding="utf-8"?>
+<resources>
+    <string name="app_name">OCR Service</string>
+</resources>