whisper : add abort_callback on lang detection (#4077)

tests/CMakeLists.txt

@@ -113,6 +113,16 @@ target_compile_definitions(${ZERO_SAMPLES_TEST} PRIVATE
 add_test(NAME ${ZERO_SAMPLES_TEST} COMMAND ${ZERO_SAMPLES_TEST})
 set_tests_properties(${ZERO_SAMPLES_TEST} PROPERTIES LABELS "tiny;gh")
 
+# abort_callback must be honored during language auto-detection (#3888)
+set(LANG_DETECT_ABORT_TEST test-whisper-lang-detect-abort)
+add_executable(${LANG_DETECT_ABORT_TEST} ${LANG_DETECT_ABORT_TEST}.cpp)
+target_include_directories(${LANG_DETECT_ABORT_TEST} PRIVATE ../include ../ggml/include ../examples)
+target_link_libraries(${LANG_DETECT_ABORT_TEST} PRIVATE common)
+target_compile_definitions(${LANG_DETECT_ABORT_TEST} PRIVATE
+    WHISPER_MODEL_PATH="${PROJECT_SOURCE_DIR}/models/for-tests-ggml-tiny.bin")
+add_test(NAME ${LANG_DETECT_ABORT_TEST} COMMAND ${LANG_DETECT_ABORT_TEST})
+set_tests_properties(${LANG_DETECT_ABORT_TEST} PROPERTIES LABELS "tiny;gh")
+
 # VAD test tests VAD in isolation
 set(VAD_TEST test-vad)
 add_executable(${VAD_TEST} ${VAD_TEST}.cpp)

tests/test-whisper-lang-detect-abort.cpp

@@ -0,0 +1,50 @@
+#include "whisper.h"
+
+#include <cstdio>
+#include <vector>
+
+#ifdef NDEBUG
+#undef NDEBUG
+#endif
+#include <cassert>
+
+static int n_encoder_begin = 0;
+
+static bool encoder_begin_cb(struct whisper_context *, struct whisper_state *, void *) {
+    n_encoder_begin++;
+    return true; // don't block anything, just count
+}
+
+static bool abort_cb(void *) {
+    return true; // abort immediately
+}
+
+int main() {
+    ggml_backend_load_all();
+
+    struct whisper_context_params cparams = whisper_context_default_params();
+    cparams.use_gpu = false;
+
+    struct whisper_context * ctx = whisper_init_from_file_with_params(WHISPER_MODEL_PATH, cparams);
+    assert(ctx != nullptr);
+
+    std::vector<float> pcmf32(2*WHISPER_SAMPLE_RATE, 0.0f);   // 2 s of silence
+
+    struct whisper_full_params params = whisper_full_default_params(WHISPER_SAMPLING_GREEDY);
+    params.language       = "auto";
+    params.print_progress = false;
+    params.print_realtime = false;
+
+    params.encoder_begin_callback = encoder_begin_cb;
+    params.abort_callback         = abort_cb;
+
+    const int rc = whisper_full(ctx, params, pcmf32.data(), pcmf32.size());
+
+    assert(rc != 0);                // the call must fail because we aborted
+    assert(n_encoder_begin == 1);   // aborted during auto-detect: main-loop encoder never started
+
+    whisper_free(ctx);
+
+    printf("test-whisper-lang-detect-abort: OK\n");
+    return 0;
+}