Add git precommit hook and github actions (#67)

Author: chen-argmax

.githooks/pre-commit

@@ -0,0 +1,34 @@
+#!/bin/sh
+
+# Get list of files that are staged for commit
+KT_FILES=$(git diff --cached --name-only --diff-filter=ACMR | grep -E '^android/.*\.kt$' || true)
+GRADLE_FILES=$(git diff --cached --name-only --diff-filter=ACMR | grep -E '.*\.gradle\.kts$' || true)
+CPP_FILES=$(git diff --cached --name-only --diff-filter=ACMR | grep -E '^(jni/.*\.(cpp|h|c|hpp|cc)|cli/.*\.(cpp|h|c|hpp|cc)|cpp/.*\.(cpp|h|c|hpp|cc))$' || true)
+
+# Run spotless checks if needed
+if [ -n "$KT_FILES" ]; then
+    echo "Running spotlessKotlinCheck..."
+    ./gradlew spotlessKotlinCheck
+    if [ $? -ne 0 ]; then
+        echo "❌ Kotlin files need formatting. Please run './gradlew spotlessKotlinApply' or 'make format' and commit again."
+        exit 1
+    fi
+fi
+
+if [ -n "$GRADLE_FILES" ]; then
+    echo "Running spotlessKotlinGradleCheck..."
+    ./gradlew spotlessKotlinGradleCheck
+    if [ $? -ne 0 ]; then
+        echo "❌ Gradle files need formatting. Please run './gradlew spotlessKotlinGradleApply' or 'make format' and commit again."
+        exit 1
+    fi
+fi
+
+if [ -n "$CPP_FILES" ]; then
+    echo "Running spotlessCppCheck..."
+    ./gradlew spotlessCppCheck
+    if [ $? -ne 0 ]; then
+        echo "❌ C++ files need formatting. Please run './gradlew spotlessCppApply' or 'make format' and commit again."
+        exit 1
+    fi
+fi 

.github/workflows/pr-checks.yml

@@ -0,0 +1,81 @@
+name: PR Checks
+
+on:
+  pull_request:
+    types: [opened, synchronize, reopened]
+  push:
+    branches:
+      - main
+
+jobs:
+  check-format:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Set up JDK 17
+        uses: actions/setup-java@v3
+        with:
+          java-version: '17'
+          distribution: 'temurin'
+          cache: gradle
+
+      - name: Install clang-format 14 (native)
+        run: |
+          sudo apt update
+          sudo apt install -y clang-format-14
+          sudo ln -sf /usr/bin/clang-format-14 /usr/local/bin/clang-format
+          clang-format --version
+
+      - name: Verify clang-format installation
+        run: |
+          if ! command -v clang-format >/dev/null; then
+            echo "❌ clang-format not found"
+            exit 1
+          fi
+          clang-format --version
+
+      - name: Grant execute permission for gradlew
+        run: chmod +x gradlew
+
+      - name: Run spotlessCheck
+        run: |
+          echo "Running spotlessCheck..."
+          ./gradlew spotlessCheck
+          if [ $? -ne 0 ]; then
+            echo "❌ spotlessCheck failed. Please run './gradlew spotlessApply' locally to fix formatting issues."
+            exit 1
+          fi
+
+  build-test-kotlin:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Set up JDK 17
+        uses: actions/setup-java@v3
+        with:
+          java-version: '17'
+          distribution: 'temurin'
+          cache: gradle
+
+      - name: Grant execute permission for gradlew
+        run: chmod +x gradlew
+
+      - name: Run detekt
+        run: |
+          echo "Running detekt..."
+          ./gradlew detekt
+          if [ $? -ne 0 ]; then
+            echo "❌ detekt found code style issues. Please fix them locally."
+            exit 1
+          fi
+
+      - name: Run unit tests
+        run: |
+          echo "Running unit tests..."
+          ./gradlew testDebugUnitTest
+          if [ $? -ne 0 ]; then
+            echo "❌ Unit tests failed. Please fix the failing tests locally."
+            exit 1
+          fi 

README.md

@@ -230,6 +230,64 @@ With `all` option, it will conduct deep clean including open source components.
 
 WhisperKit Android is currently in the beta stage. We are actively developing the project and welcome contributions from the community.
 
+## Development Setup
+
+### Installing clang-format
+
+The project uses clang-format 14 for C++ code formatting. You'll need to install it based on your operating system:
+
+#### macOS
+```bash
+# Install LLVM 14
+brew install llvm@14
+
+# Create symlink for clang-format
+sudo ln -sf /opt/homebrew/opt/llvm@14/bin/clang-format /opt/homebrew/bin/clang-format
+```
+
+#### Linux (Ubuntu/Debian)
+```bash
+sudo apt update
+sudo apt install -y clang-format-14
+sudo ln -sf /usr/bin/clang-format-14 /usr/local/bin/clang-format
+```
+
+Verify the installation:
+```bash
+clang-format --version
+```
+
+To check C++ code formatting:
+```bash
+./gradlew spotlessCppCheck
+```
+
+## Git Hooks
+
+This project uses Git hooks to maintain code quality. These hooks help ensure consistent code formatting and quality standards.
+
+### Setup
+
+To use the Git hooks, run the following command in your repository root:
+
+```bash
+git config core.hooksPath .githooks
+```
+
+### Available Hooks
+
+#### pre-commit
+- Runs `spotlessApply` to automatically fix code formatting issues
+- If formatting fixes are applied, they are automatically staged
+- The commit will be blocked if `spotlessApply` fails
+
+### Troubleshooting
+
+If you need to bypass the hooks temporarily (not recommended), you can use:
+```bash
+git commit --no-verify
+```
+
 # License
 
 - We release WhisperKit Android under [MIT License](LICENSE).

build.gradle.kts

@@ -37,6 +37,6 @@ spotless {
             "cpp/**/*.cpp", "cpp/**/*.h", "cpp/**/*.c", "cpp/**/*.hpp", "cpp/**/*.cc",
         )
         targetExclude("**/build/**", "**/external/**")
-        clangFormat("20.1.5")
+        clangFormat("14.0.6")
     }
 }

cpp/src/Models/TextDecoder.cpp

@@ -101,15 +101,15 @@ bool is_exact_match_for_separate_kv_cache_no_alignment_heads(const tflite::Model
     }
 
     /* helper functions to calculate the number of inputs and outputs for a given number of layers */
-    auto calculate_num_inputs_for_variant_with_layers = [=](const int num_layers) -> auto {
+    auto calculate_num_inputs_for_variant_with_layers = [=](const int num_layers) -> auto{
         return num_shared_inputs + num_layers * kv_factor;
     };
 
-    auto calculate_num_outputs_for_variant_with_layers = [=](const int num_layers) -> auto {
+    auto calculate_num_outputs_for_variant_with_layers = [=](const int num_layers) -> auto{
         return kv_factor * num_layers + 1;
     };
 
-    auto output_names_for_variant_with_layers = [=](const int num_layers) -> auto {
+    auto output_names_for_variant_with_layers = [=](const int num_layers) -> auto{
         std::unordered_set<std::string> output_names;
         output_names.insert(std::string("logits"));
         for (int i = 0; i < num_layers; ++i) {
@@ -119,7 +119,7 @@ bool is_exact_match_for_separate_kv_cache_no_alignment_heads(const tflite::Model
         return output_names;
     };
 
-    auto input_names_for_variant_with_layers = [](const int num_layers) -> auto {
+    auto input_names_for_variant_with_layers = [](const int num_layers) -> auto{
         std::unordered_set<std::string> input_names;
         input_names.insert(std::string("x"));
         input_names.insert(std::string("index"));

cpp/src/Text/post_proc.cpp

@@ -5,7 +5,7 @@
 #define LOGITS_TO_NEG_INF(start, end) \
     for (auto iter = (start); iter != (end); iter++) *iter = -1e9;
 
-#define DEC_2_ROUND(x) (round((x) * 100.0) / 100.0)
+#define DEC_2_ROUND(x) (round((x)*100.0) / 100.0)
 
 using namespace std;
 using json = nlohmann::json;

cpp/src/Text/post_proc.hpp

@@ -14,7 +14,7 @@ constexpr const uint32_t SAMPLE_BEGIN = 1;
 class PostProcModel : public MODEL_SUPER_CLASS {
    public:
     PostProcModel(Tokenizer* tokenizer, bool timestamp_text = false);
-    virtual ~PostProcModel() {};
+    virtual ~PostProcModel(){};
 
     bool initialize(bool debug = false);
     virtual void invoke(bool measure_time = false);

cpp/src/WhisperKitConfiguration.cpp

@@ -4,7 +4,7 @@
 #include "WhisperKitPipeline.hpp"
 #include "backend_class.hpp"
 
-whisperkit_configuration_t::whisperkit_configuration_t() {};
+whisperkit_configuration_t::whisperkit_configuration_t(){};
 
 void whisperkit_configuration_t::set_audio_encoder(const char* audio_encoder) noexcept {
     this->audio_encoder = audio_encoder;