whisper.rn 0.1.3 → 0.1.5

package/cpp/whisper.h

@@ -66,6 +66,7 @@ extern "C" {
     //
 
     struct whisper_context;
+    struct whisper_state;
 
     typedef int whisper_token;
 
@@ -101,11 +102,20 @@ extern "C" {
     WHISPER_API struct whisper_context * whisper_init_from_buffer(void * buffer, size_t buffer_size);
     WHISPER_API struct whisper_context * whisper_init(struct whisper_model_loader * loader);
 
-    // Frees all memory allocated by the model.
-    WHISPER_API void whisper_free(struct whisper_context * ctx);
+    // These are the same as the above, but the internal state of the context is not allocated automatically
+    // It is the responsibility of the caller to allocate the state using whisper_init_state() (#523)
+    WHISPER_API struct whisper_context * whisper_init_from_file_no_state(const char * path_model);
+    WHISPER_API struct whisper_context * whisper_init_from_buffer_no_state(void * buffer, size_t buffer_size);
+    WHISPER_API struct whisper_context * whisper_init_no_state(struct whisper_model_loader * loader);
+
+    WHISPER_API struct whisper_state * whisper_init_state(struct whisper_context * ctx);
+
+    // Frees all allocated memory
+    WHISPER_API void whisper_free      (struct whisper_context * ctx);
+    WHISPER_API void whisper_free_state(struct whisper_state * state);
 
     // Convert RAW PCM audio to log mel spectrogram.
-    // The resulting spectrogram is stored inside the provided whisper context.
+    // The resulting spectrogram is stored inside the default state of the provided whisper context.
     // Returns 0 on success
     WHISPER_API int whisper_pcm_to_mel(
             struct whisper_context * ctx,
@@ -113,17 +123,30 @@ extern "C" {
                                int   n_samples,
                                int   n_threads);
 
-    // Convert RAW PCM audio to log mel spectrogram but applies a Phase Vocoder to speed up the audio x2. 
-    // The resulting spectrogram is stored inside the provided whisper context.
+    WHISPER_API int whisper_pcm_to_mel_with_state(
+            struct whisper_context * ctx,
+              struct whisper_state * state,
+                       const float * samples,
+                               int   n_samples,
+                               int   n_threads);
+
+    // Convert RAW PCM audio to log mel spectrogram but applies a Phase Vocoder to speed up the audio x2.
+    // The resulting spectrogram is stored inside the default state of the provided whisper context.
     // Returns 0 on success
     WHISPER_API int whisper_pcm_to_mel_phase_vocoder(
-        struct whisper_context* ctx,
-        const float* samples,
-        int   n_samples,
-        int   n_threads);
-
-
-    // This can be used to set a custom log mel spectrogram inside the provided whisper context.
+        struct whisper_context * ctx,
+                   const float * samples,
+                           int   n_samples,
+                           int   n_threads);
+
+    WHISPER_API int whisper_pcm_to_mel_phase_vocoder_with_state(
+        struct whisper_context * ctx,
+          struct whisper_state * state,
+                   const float * samples,
+                           int   n_samples,
+                           int   n_threads);
+
+    // This can be used to set a custom log mel spectrogram inside the default state of the provided whisper context.
     // Use this instead of whisper_pcm_to_mel() if you want to provide your own log mel spectrogram.
     // n_mel must be 80
     // Returns 0 on success
@@ -133,7 +156,14 @@ extern "C" {
                                int   n_len,
                                int   n_mel);
 
-    // Run the Whisper encoder on the log mel spectrogram stored inside the provided whisper context.
+    WHISPER_API int whisper_set_mel_with_state(
+            struct whisper_context * ctx,
+              struct whisper_state * state,
+                       const float * data,
+                               int   n_len,
+                               int   n_mel);
+
+    // Run the Whisper encoder on the log mel spectrogram stored inside the default state in the provided whisper context.
     // Make sure to call whisper_pcm_to_mel() or whisper_set_mel() first.
     // offset can be used to specify the offset of the first frame in the spectrogram.
     // Returns 0 on success
@@ -142,6 +172,12 @@ extern "C" {
                                int   offset,
                                int   n_threads);
 
+    WHISPER_API int whisper_encode_with_state(
+            struct whisper_context * ctx,
+              struct whisper_state * state,
+                               int   offset,
+                               int   n_threads);
+
     // Run the Whisper decoder to obtain the logits and probabilities for the next token.
     // Make sure to call whisper_encode() first.
     // tokens + n_tokens is the provided context for the decoder.
@@ -155,6 +191,14 @@ extern "C" {
                                int   n_past,
                                int   n_threads);
 
+    WHISPER_API int whisper_decode_with_state(
+            struct whisper_context * ctx,
+              struct whisper_state * state,
+               const whisper_token * tokens,
+                               int   n_tokens,
+                               int   n_past,
+                               int   n_threads);
+
     // Convert the provided text into tokens.
     // The tokens pointer must be large enough to hold the resulting tokens.
     // Returns the number of tokens on success, no more than n_max_tokens
@@ -190,17 +234,26 @@ extern "C" {
                                int   n_threads,
                              float * lang_probs);
 
-    WHISPER_API int whisper_n_len          (struct whisper_context * ctx); // mel length
-    WHISPER_API int whisper_n_vocab        (struct whisper_context * ctx);
-    WHISPER_API int whisper_n_text_ctx     (struct whisper_context * ctx);
-    WHISPER_API int whisper_n_audio_ctx    (struct whisper_context * ctx);
-    WHISPER_API int whisper_is_multilingual(struct whisper_context * ctx);
+    WHISPER_API int whisper_lang_auto_detect_with_state(
+            struct whisper_context * ctx,
+              struct whisper_state * state,
+                               int   offset_ms,
+                               int   n_threads,
+                             float * lang_probs);
+
+    WHISPER_API int whisper_n_len           (struct whisper_context * ctx); // mel length
+    WHISPER_API int whisper_n_len_from_state(struct whisper_state * state); // mel length
+    WHISPER_API int whisper_n_vocab         (struct whisper_context * ctx);
+    WHISPER_API int whisper_n_text_ctx      (struct whisper_context * ctx);
+    WHISPER_API int whisper_n_audio_ctx     (struct whisper_context * ctx);
+    WHISPER_API int whisper_is_multilingual (struct whisper_context * ctx);
 
     // Token logits obtained from the last call to whisper_decode()
     // The logits for the last token are stored in the last row
     // Rows: n_tokens
     // Cols: n_vocab
-    WHISPER_API float * whisper_get_logits(struct whisper_context * ctx);
+    WHISPER_API float * whisper_get_logits           (struct whisper_context * ctx);
+    WHISPER_API float * whisper_get_logits_from_state(struct whisper_state * state);
 
     // Token Id -> String. Uses the vocabulary in the provided context
     WHISPER_API const char * whisper_token_to_str(struct whisper_context * ctx, whisper_token token);
@@ -218,7 +271,7 @@ extern "C" {
     WHISPER_API whisper_token whisper_token_translate (void);
     WHISPER_API whisper_token whisper_token_transcribe(void);
 
-    // Performance information
+    // Performance information from the default state.
     WHISPER_API void whisper_print_timings(struct whisper_context * ctx);
     WHISPER_API void whisper_reset_timings(struct whisper_context * ctx);
 
@@ -236,18 +289,19 @@ extern "C" {
     // Text segment callback
     // Called on every newly generated text segment
     // Use the whisper_full_...() functions to obtain the text segments
-    typedef void (*whisper_new_segment_callback)(struct whisper_context * ctx, int n_new, void * user_data);
+    typedef void (*whisper_new_segment_callback)(struct whisper_context * ctx, struct whisper_state * state, int n_new, void * user_data);
 
     // Encoder begin callback
     // If not NULL, called before the encoder starts
     // If it returns false, the computation is aborted
-    typedef bool (*whisper_encoder_begin_callback)(struct whisper_context * ctx, void * user_data);
+    typedef bool (*whisper_encoder_begin_callback)(struct whisper_context * ctx, struct whisper_state * state, void * user_data);
 
     // Logits filter callback
     // Can be used to modify the logits before sampling
     // If not NULL, called after applying temperature to logits
     typedef void (*whisper_logits_filter_callback)(
             struct whisper_context * ctx,
+              struct whisper_state * state,
           const whisper_token_data * tokens,
                                int   n_tokens,
                              float * logits,
@@ -334,6 +388,7 @@ extern "C" {
     WHISPER_API struct whisper_full_params whisper_full_default_params(enum whisper_sampling_strategy strategy);
 
     // Run the entire model: PCM -> log mel spectrogram -> encoder -> decoder -> text
+    // Not thread safe for same context
     // Uses the specified decoding strategy to obtain the text.
     WHISPER_API int whisper_full(
                 struct whisper_context * ctx,
@@ -341,7 +396,16 @@ extern "C" {
                            const float * samples,
                                    int   n_samples);
 
-    // Split the input audio in chunks and process each chunk separately using whisper_full()
+    WHISPER_API int whisper_full_with_state(
+                struct whisper_context * ctx,
+                  struct whisper_state * state,
+            struct whisper_full_params   params,
+                           const float * samples,
+                                   int   n_samples);
+
+    // Split the input audio in chunks and process each chunk separately using whisper_full_with_state()
+    // Result is stored in the default state of the context
+    // Not thread safe if executed in parallel on the same context.
     // It seems this approach can offer some speedup in some cases.
     // However, the transcription accuracy can be worse at the beginning and end of each chunk.
     WHISPER_API int whisper_full_parallel(
@@ -351,40 +415,56 @@ extern "C" {
                                    int   n_samples,
                                    int   n_processors);
 
-    // Number of generated text segments.
+    // Number of generated text segments
     // A segment can be a few words, a sentence, or even a paragraph.
-    WHISPER_API int whisper_full_n_segments(struct whisper_context * ctx);
+    WHISPER_API int whisper_full_n_segments           (struct whisper_context * ctx);
+    WHISPER_API int whisper_full_n_segments_from_state(struct whisper_state * state);
 
-    // Language id associated with the current context
+    // Language id associated with the context's default state
     WHISPER_API int whisper_full_lang_id(struct whisper_context * ctx);
 
-    // Get the start and end time of the specified segment.
-    WHISPER_API int64_t whisper_full_get_segment_t0(struct whisper_context * ctx, int i_segment);
-    WHISPER_API int64_t whisper_full_get_segment_t1(struct whisper_context * ctx, int i_segment);
+    // Language id associated with the provided state
+    WHISPER_API int whisper_full_lang_id_from_state(struct whisper_state * state);
+
+    // Get the start and end time of the specified segment
+    WHISPER_API int64_t whisper_full_get_segment_t0           (struct whisper_context * ctx, int i_segment);
+    WHISPER_API int64_t whisper_full_get_segment_t0_from_state(struct whisper_state * state, int i_segment);
+
+    WHISPER_API int64_t whisper_full_get_segment_t1           (struct whisper_context * ctx, int i_segment);
+    WHISPER_API int64_t whisper_full_get_segment_t1_from_state(struct whisper_state * state, int i_segment);
+
+    // Get the text of the specified segment
+    WHISPER_API const char * whisper_full_get_segment_text           (struct whisper_context * ctx, int i_segment);
+    WHISPER_API const char * whisper_full_get_segment_text_from_state(struct whisper_state * state, int i_segment);
 
-    // Get the text of the specified segment.
-    WHISPER_API const char * whisper_full_get_segment_text(struct whisper_context * ctx, int i_segment);
+    // Get number of tokens in the specified segment
+    WHISPER_API int whisper_full_n_tokens           (struct whisper_context * ctx, int i_segment);
+    WHISPER_API int whisper_full_n_tokens_from_state(struct whisper_state * state, int i_segment);
 
-    // Get number of tokens in the specified segment.
-    WHISPER_API int whisper_full_n_tokens(struct whisper_context * ctx, int i_segment);
+    // Get the token text of the specified token in the specified segment
+    WHISPER_API const char * whisper_full_get_token_text           (struct whisper_context * ctx, int i_segment, int i_token);
+    WHISPER_API const char * whisper_full_get_token_text_from_state(struct whisper_context * ctx, struct whisper_state * state, int i_segment, int i_token);
 
-    // Get the token text of the specified token in the specified segment.
-    WHISPER_API const char * whisper_full_get_token_text(struct whisper_context * ctx, int i_segment, int i_token);
-    WHISPER_API whisper_token whisper_full_get_token_id (struct whisper_context * ctx, int i_segment, int i_token);
+    WHISPER_API whisper_token whisper_full_get_token_id           (struct whisper_context * ctx, int i_segment, int i_token);
+    WHISPER_API whisper_token whisper_full_get_token_id_from_state(struct whisper_state * state, int i_segment, int i_token);
 
-    // Get token data for the specified token in the specified segment.
+    // Get token data for the specified token in the specified segment
     // This contains probabilities, timestamps, etc.
-    WHISPER_API whisper_token_data whisper_full_get_token_data(struct whisper_context * ctx, int i_segment, int i_token);
+    WHISPER_API whisper_token_data whisper_full_get_token_data           (struct whisper_context * ctx, int i_segment, int i_token);
+    WHISPER_API whisper_token_data whisper_full_get_token_data_from_state(struct whisper_state * state, int i_segment, int i_token);
 
-    // Get the probability of the specified token in the specified segment.
-    WHISPER_API float whisper_full_get_token_p(struct whisper_context * ctx, int i_segment, int i_token);
+    // Get the probability of the specified token in the specified segment
+    WHISPER_API float whisper_full_get_token_p           (struct whisper_context * ctx, int i_segment, int i_token);
+    WHISPER_API float whisper_full_get_token_p_from_state(struct whisper_state * state, int i_segment, int i_token);
 
     ////////////////////////////////////////////////////////////////////////////
 
     // Temporary helpers needed for exposing ggml interface
 
     WHISPER_API int whisper_bench_memcpy(int n_threads);
+    WHISPER_API const char * whisper_bench_memcpy_str(int n_threads);
     WHISPER_API int whisper_bench_ggml_mul_mat(int n_threads);
+    WHISPER_API const char * whisper_bench_ggml_mul_mat_str(int n_threads);
 
 #ifdef __cplusplus
 }

package/ios/RNWhisper.h

@@ -1,7 +1,9 @@
 #ifdef __cplusplus
 #import "whisper.h"
+#import "rn-whisper.h"
 #endif
 
+
 #import <React/RCTBridgeModule.h>
 
 @interface RNWhisper : NSObject <RCTBridgeModule>

package/ios/RNWhisper.mm

@@ -1,6 +1,7 @@
 
 #import "RNWhisper.h"
 #include <stdlib.h>
+#include <string>
 
 @interface WhisperContext : NSObject {
 }
@@ -74,7 +75,7 @@ RCT_REMAP_METHOD(transcribe,
 
     const int max_threads = options[@"maxThreads"] != nil ?
       [options[@"maxThreads"] intValue] :
-      MIN(8, (int)[[NSProcessInfo processInfo] processorCount]) - 2;
+      MIN(4, (int)[[NSProcessInfo processInfo] processorCount]);
 
     if (options[@"beamSize"] != nil) {
         params.strategy = WHISPER_SAMPLING_BEAM_SEARCH;
@@ -93,15 +94,18 @@ RCT_REMAP_METHOD(transcribe,
     params.no_context       = true;
     params.single_segment   = false;
 
+    if (options[@"maxLen"] != nil) {
+        params.max_len = [options[@"maxLen"] intValue];
+    }
+    params.token_timestamps = options[@"tokenTimestamps"] != nil ? [options[@"tokenTimestamps"] boolValue] : false;
+
     if (options[@"bestOf"] != nil) {
         params.greedy.best_of = [options[@"bestOf"] intValue];
     }
     if (options[@"maxContext"] != nil) {
         params.n_max_text_ctx = [options[@"maxContext"] intValue];
     }
-    if (options[@"maxLen"] != nil) {
-        params.max_len = [options[@"maxLen"] intValue];
-    }
+    
     if (options[@"offset"] != nil) {
         params.offset_ms = [options[@"offset"] intValue];
     }
@@ -117,6 +121,15 @@ RCT_REMAP_METHOD(transcribe,
     if (options[@"temperatureInc"] != nil) {
         params.temperature_inc = [options[@"temperature_inc"] floatValue];
     }
+    
+    if (options[@"prompt"] != nil) {
+        std::string *prompt = new std::string([options[@"prompt"] UTF8String]);
+        rn_whisper_convert_prompt(
+            context.ctx,
+            params,
+            prompt
+        );
+    }
 
     whisper_reset_timings(context.ctx);
     int code = whisper_full(context.ctx, params, waveFile, count);
@@ -132,11 +145,25 @@ RCT_REMAP_METHOD(transcribe,
 
     NSString *result = @"";
     int n_segments = whisper_full_n_segments(context.ctx);
+
+    NSMutableArray *segments = [[NSMutableArray alloc] init];
     for (int i = 0; i < n_segments; i++) {
         const char * text_cur = whisper_full_get_segment_text(context.ctx, i);
         result = [result stringByAppendingString:[NSString stringWithUTF8String:text_cur]];
-    }
-    resolve(result);
+
+        const int64_t t0 = whisper_full_get_segment_t0(context.ctx, i);
+        const int64_t t1 = whisper_full_get_segment_t1(context.ctx, i);
+        NSDictionary *segment = @{
+            @"text": [NSString stringWithUTF8String:text_cur],
+            @"t0": [NSNumber numberWithLongLong:t0],
+            @"t1": [NSNumber numberWithLongLong:t1]
+        };
+        [segments addObject:segment];
+    }
+    resolve(@{
+        @"result": result,
+        @"segments": segments
+    });
 }
 
 RCT_REMAP_METHOD(releaseContext,

package/jest/mock.js

@@ -3,7 +3,10 @@ const { NativeModules } = require('react-native')
 if (!NativeModules.RNWhisper) {
   NativeModules.RNWhisper = {
     initContext: jest.fn(() => Promise.resolve(1)),
-    transcribe: jest.fn(() => Promise.resolve('TEST')),
+    transcribe: jest.fn(() => Promise.resolve({
+      result: ' Test',
+      segments: [{ text: ' Test', t0: 0, t1: 33 }],
+    })),
     releaseContext: jest.fn(() => Promise.resolve()),
     releaseAllContexts: jest.fn(() => Promise.resolve()),
   }

package/lib/commonjs/index.js

@@ -21,9 +21,7 @@ class WhisperContext {
   }
   async transcribe(path) {
     let options = arguments.length > 1 && arguments[1] !== undefined ? arguments[1] : {};
-    return RNWhisper.transcribe(this.id, path, options).then(result => ({
-      result
-    }));
+    return RNWhisper.transcribe(this.id, path, options);
   }
   async release() {
     return RNWhisper.releaseContext(this.id);

package/lib/commonjs/index.js.map

@@ -1 +1 @@
-{"version":3,"names":["_reactNative","require","LINKING_ERROR","Platform","select","ios","default","RNWhisper","NativeModules","Proxy","get","Error","WhisperContext","constructor","id","transcribe","path","options","arguments","length","undefined","then","result","release","releaseContext","initWhisper","filePath","initContext","releaseAllWhisper","releaseAllContexts"],"sourceRoot":"../../src","sources":["index.tsx"],"mappings":";;;;;;;AAAA,IAAAA,YAAA,GAAAC,OAAA;AAEA,MAAMC,aAAa,GAChB,sEAAqEC,qBAAQ,CAACC,MAAM,CAAC;EAAEC,GAAG,EAAE,gCAAgC;EAAEC,OAAO,EAAE;AAAG,CAAC,CAC3I,oDAAmD;AAEtD,MAAMC,SAAS,GAAGC,0BAAa,CAACD,SAAS,GACrCC,0BAAa,CAACD,SAAS,GACvB,IAAIE,KAAK,CACT,CAAC,CAAC,EACF;EACEC,GAAGA,CAAA,EAAG;IACJ,MAAM,IAAIC,KAAK,CAACT,aAAa,CAAC;EAChC;AACF,CAAC,CACF;AAoBH,MAAMU,cAAc,CAAC;EAGnBC,WAAWA,CAACC,EAAU,EAAE;IACtB,IAAI,CAACA,EAAE,GAAGA,EAAE;EACd;EAEA,MAAMC,UAAUA,CAACC,IAAY,EAA8D;IAAA,IAA5DC,OAA0B,GAAAC,SAAA,CAAAC,MAAA,QAAAD,SAAA,QAAAE,SAAA,GAAAF,SAAA,MAAG,CAAC,CAAC;IAC5D,OAAOX,SAAS,CAACQ,UAAU,CAAC,IAAI,CAACD,EAAE,EAAEE,IAAI,EAAEC,OAAO,CAAC,CAACI,IAAI,CAAEC,MAAc,KAAM;MAC5EA;IACF,CAAC,CAAC,CAAC;EACL;EAEA,MAAMC,OAAOA,CAAA,EAAG;IACd,OAAOhB,SAAS,CAACiB,cAAc,CAAC,IAAI,CAACV,EAAE,CAAC;EAC1C;AACF;AAEO,eAAeW,WAAWA,CAAA,EAEN;EAAA,IADzB;IAAEC;EAAgC,CAAC,GAAAR,SAAA,CAAAC,MAAA,QAAAD,SAAA,QAAAE,SAAA,GAAAF,SAAA,MAAG,CAAC,CAAC;EAExC,MAAMJ,EAAE,GAAG,MAAMP,SAAS,CAACoB,WAAW,CAACD,QAAQ,CAAC;EAChD,OAAO,IAAId,cAAc,CAACE,EAAE,CAAC;AAC/B;AAEO,eAAec,iBAAiBA,CAAA,EAAkB;EACvD,OAAOrB,SAAS,CAACsB,kBAAkB,EAAE;AACvC"}
+{"version":3,"names":["_reactNative","require","LINKING_ERROR","Platform","select","ios","default","RNWhisper","NativeModules","Proxy","get","Error","WhisperContext","constructor","id","transcribe","path","options","arguments","length","undefined","release","releaseContext","initWhisper","filePath","initContext","releaseAllWhisper","releaseAllContexts"],"sourceRoot":"../../src","sources":["index.tsx"],"mappings":";;;;;;;AAAA,IAAAA,YAAA,GAAAC,OAAA;AAEA,MAAMC,aAAa,GAChB,sEAAqEC,qBAAQ,CAACC,MAAM,CAAC;EAAEC,GAAG,EAAE,gCAAgC;EAAEC,OAAO,EAAE;AAAG,CAAC,CAC3I,oDAAmD;AAEtD,MAAMC,SAAS,GAAGC,0BAAa,CAACD,SAAS,GACrCC,0BAAa,CAACD,SAAS,GACvB,IAAIE,KAAK,CACT,CAAC,CAAC,EACF;EACEC,GAAGA,CAAA,EAAG;IACJ,MAAM,IAAIC,KAAK,CAACT,aAAa,CAAC;EAChC;AACF,CAAC,CACF;AA6BH,MAAMU,cAAc,CAAC;EAGnBC,WAAWA,CAACC,EAAU,EAAE;IACtB,IAAI,CAACA,EAAE,GAAGA,EAAE;EACd;EAEA,MAAMC,UAAUA,CAACC,IAAY,EAA8D;IAAA,IAA5DC,OAA0B,GAAAC,SAAA,CAAAC,MAAA,QAAAD,SAAA,QAAAE,SAAA,GAAAF,SAAA,MAAG,CAAC,CAAC;IAC5D,OAAOX,SAAS,CAACQ,UAAU,CAAC,IAAI,CAACD,EAAE,EAAEE,IAAI,EAAEC,OAAO,CAAC;EACrD;EAEA,MAAMI,OAAOA,CAAA,EAAG;IACd,OAAOd,SAAS,CAACe,cAAc,CAAC,IAAI,CAACR,EAAE,CAAC;EAC1C;AACF;AAEO,eAAeS,WAAWA,CAAA,EAEN;EAAA,IADzB;IAAEC;EAAgC,CAAC,GAAAN,SAAA,CAAAC,MAAA,QAAAD,SAAA,QAAAE,SAAA,GAAAF,SAAA,MAAG,CAAC,CAAC;EAExC,MAAMJ,EAAE,GAAG,MAAMP,SAAS,CAACkB,WAAW,CAACD,QAAQ,CAAC;EAChD,OAAO,IAAIZ,cAAc,CAACE,EAAE,CAAC;AAC/B;AAEO,eAAeY,iBAAiBA,CAAA,EAAkB;EACvD,OAAOnB,SAAS,CAACoB,kBAAkB,EAAE;AACvC"}

package/lib/module/index.js

@@ -14,9 +14,7 @@ class WhisperContext {
   }
   async transcribe(path) {
     let options = arguments.length > 1 && arguments[1] !== undefined ? arguments[1] : {};
-    return RNWhisper.transcribe(this.id, path, options).then(result => ({
-      result
-    }));
+    return RNWhisper.transcribe(this.id, path, options);
   }
   async release() {
     return RNWhisper.releaseContext(this.id);

package/lib/module/index.js.map

@@ -1 +1 @@
-{"version":3,"names":["NativeModules","Platform","LINKING_ERROR","select","ios","default","RNWhisper","Proxy","get","Error","WhisperContext","constructor","id","transcribe","path","options","arguments","length","undefined","then","result","release","releaseContext","initWhisper","filePath","initContext","releaseAllWhisper","releaseAllContexts"],"sourceRoot":"../../src","sources":["index.tsx"],"mappings":"AAAA,SAASA,aAAa,EAAEC,QAAQ,QAAQ,cAAc;AAEtD,MAAMC,aAAa,GAChB,sEAAqED,QAAQ,CAACE,MAAM,CAAC;EAAEC,GAAG,EAAE,gCAAgC;EAAEC,OAAO,EAAE;AAAG,CAAC,CAC3I,oDAAmD;AAEtD,MAAMC,SAAS,GAAGN,aAAa,CAACM,SAAS,GACrCN,aAAa,CAACM,SAAS,GACvB,IAAIC,KAAK,CACT,CAAC,CAAC,EACF;EACEC,GAAGA,CAAA,EAAG;IACJ,MAAM,IAAIC,KAAK,CAACP,aAAa,CAAC;EAChC;AACF,CAAC,CACF;AAoBH,MAAMQ,cAAc,CAAC;EAGnBC,WAAWA,CAACC,EAAU,EAAE;IACtB,IAAI,CAACA,EAAE,GAAGA,EAAE;EACd;EAEA,MAAMC,UAAUA,CAACC,IAAY,EAA8D;IAAA,IAA5DC,OAA0B,GAAAC,SAAA,CAAAC,MAAA,QAAAD,SAAA,QAAAE,SAAA,GAAAF,SAAA,MAAG,CAAC,CAAC;IAC5D,OAAOV,SAAS,CAACO,UAAU,CAAC,IAAI,CAACD,EAAE,EAAEE,IAAI,EAAEC,OAAO,CAAC,CAACI,IAAI,CAAEC,MAAc,KAAM;MAC5EA;IACF,CAAC,CAAC,CAAC;EACL;EAEA,MAAMC,OAAOA,CAAA,EAAG;IACd,OAAOf,SAAS,CAACgB,cAAc,CAAC,IAAI,CAACV,EAAE,CAAC;EAC1C;AACF;AAEA,OAAO,eAAeW,WAAWA,CAAA,EAEN;EAAA,IADzB;IAAEC;EAAgC,CAAC,GAAAR,SAAA,CAAAC,MAAA,QAAAD,SAAA,QAAAE,SAAA,GAAAF,SAAA,MAAG,CAAC,CAAC;EAExC,MAAMJ,EAAE,GAAG,MAAMN,SAAS,CAACmB,WAAW,CAACD,QAAQ,CAAC;EAChD,OAAO,IAAId,cAAc,CAACE,EAAE,CAAC;AAC/B;AAEA,OAAO,eAAec,iBAAiBA,CAAA,EAAkB;EACvD,OAAOpB,SAAS,CAACqB,kBAAkB,EAAE;AACvC"}
+{"version":3,"names":["NativeModules","Platform","LINKING_ERROR","select","ios","default","RNWhisper","Proxy","get","Error","WhisperContext","constructor","id","transcribe","path","options","arguments","length","undefined","release","releaseContext","initWhisper","filePath","initContext","releaseAllWhisper","releaseAllContexts"],"sourceRoot":"../../src","sources":["index.tsx"],"mappings":"AAAA,SAASA,aAAa,EAAEC,QAAQ,QAAQ,cAAc;AAEtD,MAAMC,aAAa,GAChB,sEAAqED,QAAQ,CAACE,MAAM,CAAC;EAAEC,GAAG,EAAE,gCAAgC;EAAEC,OAAO,EAAE;AAAG,CAAC,CAC3I,oDAAmD;AAEtD,MAAMC,SAAS,GAAGN,aAAa,CAACM,SAAS,GACrCN,aAAa,CAACM,SAAS,GACvB,IAAIC,KAAK,CACT,CAAC,CAAC,EACF;EACEC,GAAGA,CAAA,EAAG;IACJ,MAAM,IAAIC,KAAK,CAACP,aAAa,CAAC;EAChC;AACF,CAAC,CACF;AA6BH,MAAMQ,cAAc,CAAC;EAGnBC,WAAWA,CAACC,EAAU,EAAE;IACtB,IAAI,CAACA,EAAE,GAAGA,EAAE;EACd;EAEA,MAAMC,UAAUA,CAACC,IAAY,EAA8D;IAAA,IAA5DC,OAA0B,GAAAC,SAAA,CAAAC,MAAA,QAAAD,SAAA,QAAAE,SAAA,GAAAF,SAAA,MAAG,CAAC,CAAC;IAC5D,OAAOV,SAAS,CAACO,UAAU,CAAC,IAAI,CAACD,EAAE,EAAEE,IAAI,EAAEC,OAAO,CAAC;EACrD;EAEA,MAAMI,OAAOA,CAAA,EAAG;IACd,OAAOb,SAAS,CAACc,cAAc,CAAC,IAAI,CAACR,EAAE,CAAC;EAC1C;AACF;AAEA,OAAO,eAAeS,WAAWA,CAAA,EAEN;EAAA,IADzB;IAAEC;EAAgC,CAAC,GAAAN,SAAA,CAAAC,MAAA,QAAAD,SAAA,QAAAE,SAAA,GAAAF,SAAA,MAAG,CAAC,CAAC;EAExC,MAAMJ,EAAE,GAAG,MAAMN,SAAS,CAACiB,WAAW,CAACD,QAAQ,CAAC;EAChD,OAAO,IAAIZ,cAAc,CAACE,EAAE,CAAC;AAC/B;AAEA,OAAO,eAAeY,iBAAiBA,CAAA,EAAkB;EACvD,OAAOlB,SAAS,CAACmB,kBAAkB,EAAE;AACvC"}

package/lib/typescript/index.d.ts

@@ -1,7 +1,10 @@
 export type TranscribeOptions = {
+    language?: string;
+    translate?: boolean;
     maxThreads?: number;
     maxContext?: number;
     maxLen?: number;
+    tokenTimestamps?: boolean;
     offset?: number;
     duration?: number;
     wordThold?: number;
@@ -10,9 +13,15 @@ export type TranscribeOptions = {
     beamSize?: number;
     bestOf?: number;
     speedUp?: boolean;
+    prompt?: string;
 };
 export type TranscribeResult = {
     result: string;
+    segments: Array<{
+        text: string;
+        t0: number;
+        t1: number;
+    }>;
 };
 declare class WhisperContext {
     id: number;

package/lib/typescript/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/index.tsx"],"names":[],"mappings":"AAiBA,MAAM,MAAM,iBAAiB,GAAG;IAC9B,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,OAAO,CAAC,EAAE,OAAO,CAAC;CACnB,CAAA;AAED,MAAM,MAAM,gBAAgB,GAAG;IAC7B,MAAM,EAAE,MAAM,CAAC;CAChB,CAAA;AAED,cAAM,cAAc;IAClB,EAAE,EAAE,MAAM,CAAA;gBAEE,EAAE,EAAE,MAAM;IAIhB,UAAU,CAAC,IAAI,EAAE,MAAM,EAAE,OAAO,GAAE,iBAAsB,GAAG,OAAO,CAAC,gBAAgB,CAAC;IAMpF,OAAO;CAGd;AAED,wBAAsB,WAAW,CAC/B,EAAE,QAAQ,EAAE,GAAE;IAAE,QAAQ,CAAC,EAAE,MAAM,CAAA;CAAO,GACvC,OAAO,CAAC,cAAc,CAAC,CAGzB;AAED,wBAAsB,iBAAiB,IAAI,OAAO,CAAC,IAAI,CAAC,CAEvD"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/index.tsx"],"names":[],"mappings":"AAiBA,MAAM,MAAM,iBAAiB,GAAG;IAC9B,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,SAAS,CAAC,EAAE,OAAO,CAAC;IACpB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,eAAe,CAAC,EAAE,OAAO,CAAC;IAC1B,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,OAAO,CAAC,EAAE,OAAO,CAAC;IAClB,MAAM,CAAC,EAAE,MAAM,CAAC;CACjB,CAAA;AAED,MAAM,MAAM,gBAAgB,GAAG;IAC7B,MAAM,EAAE,MAAM,CAAC;IACf,QAAQ,EAAE,KAAK,CAAC;QACd,IAAI,EAAE,MAAM,CAAC;QACb,EAAE,EAAE,MAAM,CAAC;QACX,EAAE,EAAE,MAAM,CAAC;KACZ,CAAC,CAAC;CACJ,CAAA;AAED,cAAM,cAAc;IAClB,EAAE,EAAE,MAAM,CAAA;gBAEE,EAAE,EAAE,MAAM;IAIhB,UAAU,CAAC,IAAI,EAAE,MAAM,EAAE,OAAO,GAAE,iBAAsB,GAAG,OAAO,CAAC,gBAAgB,CAAC;IAIpF,OAAO;CAGd;AAED,wBAAsB,WAAW,CAC/B,EAAE,QAAQ,EAAE,GAAE;IAAE,QAAQ,CAAC,EAAE,MAAM,CAAA;CAAO,GACvC,OAAO,CAAC,cAAc,CAAC,CAGzB;AAED,wBAAsB,iBAAiB,IAAI,OAAO,CAAC,IAAI,CAAC,CAEvD"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "whisper.rn",
-  "version": "0.1.3",
+  "version": "0.1.5",
   "description": "React Native binding of whisper.cpp",
   "main": "lib/commonjs/index",
   "module": "lib/module/index",
@@ -47,7 +47,7 @@
     "speech recognition"
   ],
   "repository": "https://github.com/mybigday/whisper.rn",
-  "author": "Jhen <developer@jhen.me> (https://github.com/jhen0409)",
+  "author": "Jhen-Jie Hong <developer@jhen.me>",
   "license": "MIT",
   "bugs": {
     "url": "https://github.com/mybigday/whisper.rn/issues"
@@ -98,13 +98,17 @@
       "@commitlint/config-conventional"
     ]
   },
+  "publishConfig": {
+    "registry": "https://registry.npmjs.org/"
+  },
   "release-it": {
     "git": {
       "commitMessage": "chore: release ${version}",
       "tagName": "v${version}"
     },
     "npm": {
-      "publish": true
+      "publish": true,
+      "skipChecks": true
     },
     "github": {
       "release": true

package/src/index.tsx

@@ -16,9 +16,12 @@ const RNWhisper = NativeModules.RNWhisper
   )
 
 export type TranscribeOptions = {
+  language?: string,
+  translate?: boolean,
   maxThreads?: number,
   maxContext?: number,
   maxLen?: number,
+  tokenTimestamps?: boolean,
   offset?: number,
   duration?: number,
   wordThold?: number,
@@ -27,10 +30,16 @@ export type TranscribeOptions = {
   beamSize?: number,
   bestOf?: number,
   speedUp?: boolean,
+  prompt?: string,
 }
 
 export type TranscribeResult = {
   result: string,
+  segments: Array<{
+    text: string,
+    t0: number,
+    t1: number,
+  }>,
 }
 
 class WhisperContext {
@@ -41,9 +50,7 @@ class WhisperContext {
   }
 
   async transcribe(path: string, options: TranscribeOptions = {}): Promise<TranscribeResult> {
-    return RNWhisper.transcribe(this.id, path, options).then((result: string) => ({
-      result
-    }))
+    return RNWhisper.transcribe(this.id, path, options)
   }
 
   async release() {

package/whisper-rn.podspec

@@ -1,8 +1,9 @@
 require "json"
 
 package = JSON.parse(File.read(File.join(__dir__, "package.json")))
-base_compiler_flags = '-DGGML_USE_ACCELERATE -Wno-shorten-64-to-32'
-folly_compiler_flags = '-DFOLLY_NO_CONFIG -DFOLLY_MOBILE=1 -DFOLLY_USE_LIBCPP=1 -Wno-comma -Wno-shorten-64-to-32'
+base_compiler_flags = "-DGGML_USE_ACCELERATE -Wno-shorten-64-to-32"
+folly_compiler_flags = "-DFOLLY_NO_CONFIG -DFOLLY_MOBILE=1 -DFOLLY_USE_LIBCPP=1 -Wno-comma"
+base_optimizer_flags = "-O3 -DNDEBUG"
 
 Pod::Spec.new do |s|
   s.name         = "whisper-rn"
@@ -20,17 +21,23 @@ Pod::Spec.new do |s|
   s.dependency "React-Core"
 
   s.compiler_flags = base_compiler_flags
-  s.xcconfig = {
-    'OTHER_LDFLAGS' => '-framework Accelerate'
+  s.pod_target_xcconfig = {
+    "OTHER_LDFLAGS" => "-framework Accelerate",
+    "OTHER_CFLAGS[config=Release]" => base_optimizer_flags,
+    "OTHER_CPLUSPLUSFLAGS[config=Release]" => base_optimizer_flags
   }
 
   # Don't install the dependencies when we run `pod install` in the old architecture.
   if ENV['RCT_NEW_ARCH_ENABLED'] == '1' then
     s.compiler_flags = base_compiler_flags + " " + folly_compiler_flags + " -DRCT_NEW_ARCH_ENABLED=1"
-    s.pod_target_xcconfig    = {
-        "HEADER_SEARCH_PATHS" => "\"$(PODS_ROOT)/boost\"",
-        "OTHER_CPLUSPLUSFLAGS" => "-DFOLLY_NO_CONFIG -DFOLLY_MOBILE=1 -DFOLLY_USE_LIBCPP=1",
-        "CLANG_CXX_LANGUAGE_STANDARD" => "c++17"
+    new_arch_cpp_flags = "-DFOLLY_NO_CONFIG -DFOLLY_MOBILE=1 -DFOLLY_USE_LIBCPP=1"
+    s.pod_target_xcconfig = {
+      "CLANG_CXX_LANGUAGE_STANDARD" => "c++17",
+      "HEADER_SEARCH_PATHS" => "\"$(PODS_ROOT)/boost\"",
+      "OTHER_LDFLAGS" => "-framework Accelerate",
+      "OTHER_CFLAGS[config=Release]" => base_optimizer_flags,
+      "OTHER_CPLUSPLUSFLAGS[config=Debug]" => new_arch_cpp_flags,
+      "OTHER_CPLUSPLUSFLAGS[config=Release]" => new_arch_cpp_flags + " " + base_optimizer_flags
     }
     s.dependency "React-Codegen"
     s.dependency "RCT-Folly"