wip : initial WASM port

Works but it is very slow because no SIMD is used.
For example, jfk.wav is processed in ~23 seconds using "tiny.en" model

bindings/javascript/.gitignore

@@ -0,0 +1 @@
+publish.log

bindings/javascript/CMakeLists.txt

@@ -0,0 +1,34 @@
+set(TARGET libwhisper)
+
+add_executable(${TARGET}
+    emscripten.cpp
+    )
+
+target_link_libraries(${TARGET} PRIVATE
+    whisper
+    )
+
+unset(EXTRA_FLAGS)
+if (WHISPER_WASM_SINGLE_FILE)
+    set(EXTRA_FLAGS "-s SINGLE_FILE=1")
+    message(STATUS "Embedding WASM inside whisper.js")
+
+    add_custom_command(
+        TARGET libwhisper POST_BUILD
+        COMMAND ${CMAKE_COMMAND} -E copy
+        ${CMAKE_BINARY_DIR}/bin/libwhisper.js
+        ${CMAKE_CURRENT_SOURCE_DIR}/whisper.js
+        )
+endif()
+
+set_target_properties(${TARGET} PROPERTIES LINK_FLAGS " \
+    --bind \
+    -s MODULARIZE=1 \
+    -s ASSERTIONS=1 \
+    -s USE_PTHREADS=1 \
+    -s PTHREAD_POOL_SIZE=8 \
+    -s TOTAL_MEMORY=536870912 \
+    -s FORCE_FILESYSTEM=1 \
+    -s EXPORT_NAME=\"'whisper_factory'\" \
+    ${EXTRA_FLAGS} \
+    ")

bindings/javascript/emscripten.cpp

@@ -0,0 +1,70 @@
+#include "whisper.h"
+
+#include <emscripten.h>
+#include <emscripten/bind.h>
+
+#include <vector>
+
+std::vector<struct whisper_context *> g_contexts(4, nullptr);
+
+EMSCRIPTEN_BINDINGS(whisper) {
+    emscripten::function("init", emscripten::optional_override([](const std::string & path_model) {
+        for (size_t i = 0; i < g_contexts.size(); ++i) {
+            if (g_contexts[i] == nullptr) {
+                g_contexts[i] = whisper_init(path_model.c_str());
+                return i + 1;
+            }
+        }
+
+        return (size_t) 0;
+    }));
+
+    emscripten::function("free", emscripten::optional_override([](size_t index) {
+        --index;
+
+        if (index < g_contexts.size()) {
+            whisper_free(g_contexts[index]);
+            g_contexts[index] = nullptr;
+        }
+    }));
+
+    emscripten::function("full_default", emscripten::optional_override([](size_t index, const emscripten::val & audio) {
+        --index;
+
+        if (index >= g_contexts.size()) {
+            return -1;
+        }
+
+        if (g_contexts[index] == nullptr) {
+            return -2;
+        }
+
+        struct whisper_full_params params = whisper_full_default_params(whisper_sampling_strategy::WHISPER_SAMPLING_GREEDY);
+
+        params.print_realtime       = true;
+        params.print_progress       = false;
+        params.print_timestamps     = true;
+        params.print_special_tokens = false;
+        params.translate            = false;
+        params.language             = "en";
+        params.n_threads            = 4;
+        params.offset_ms            = 0;
+
+        std::vector<float> pcmf32;
+        const int n = audio["length"].as<int>();
+
+        emscripten::val heap = emscripten::val::module_property("HEAPU8");
+        emscripten::val memory = heap["buffer"];
+
+        pcmf32.resize(n);
+
+        emscripten::val memoryView = audio["constructor"].new_(memory, reinterpret_cast<uintptr_t>(pcmf32.data()), n);
+        memoryView.call<void>("set", audio);
+
+        int ret = whisper_full(g_contexts[index], params, pcmf32.data(), pcmf32.size());
+
+        whisper_print_timings(g_contexts[index]);
+
+        return ret;
+    }));
+}