Pull in new changes from original gentle repo

.travis.yml

@@ -13,5 +13,5 @@ script:
 after_success:
   - if [ "$TRAVIS_BRANCH" == "master" ]; then
     docker login -u="$DOCKER_USERNAME" -p="$DOCKER_PASSWORD";
-    docker push lowerquality/gentle:latest
+    docker push lowerquality/gentle:latest;
     fi

Pipfile

@@ -0,0 +1,12 @@
+[[source]]
+url = "https://pypi.org/simple"
+verify_ssl = true
+name = "pypi"
+
+[packages]
+twisted = "*"
+
+[dev-packages]
+
+[requires]
+python_version = "3.10"

README.md

@@ -24,6 +24,9 @@ curl -F "audio=@audio.mp3" -F "transcript=@words.txt" "http://localhost:8765/tra
 If you've downloaded the source code you can also run the aligner as a command line program:
 
 ```bash
+git clone https://github.com/lowerquality/gentle.git
+cd gentle
+./install.sh
 python3 align.py audio.mp3 words.txt
 ```
 

align.py

@@ -43,7 +43,7 @@ def on_progress(p):
         logging.debug("%s: %s" % (k, v))
 
 
-with open(args.txtfile) as fh:
+with open(args.txtfile, encoding="utf-8") as fh:
     transcript = fh.read()
 
 resources = gentle.Resources()
@@ -54,7 +54,7 @@ def on_progress(p):
     aligner = gentle.ForcedAligner(resources, transcript, nthreads=args.nthreads, disfluency=args.disfluency, conservative=args.conservative, disfluencies=disfluencies)
     result = aligner.transcribe(wavfile, progress_cb=on_progress, logging=logging)
 
-fh = open(args.output, 'w') if args.output else sys.stdout
+fh = open(args.output, 'w', encoding="utf-8") if args.output else sys.stdout
 fh.write(result.to_json(indent=2))
 if args.output:
     logging.info("output written to %s" % (args.output))

ext/install_kaldi.sh

@@ -4,9 +4,8 @@
 cd kaldi/tools
 make clean
 make
-./extras/install_openblas.sh
 cd ../src
-make clean
-./configure --static --static-math=yes --static-fst=yes --use-cuda=no --openblas-root=../tools/OpenBLAS/install
+# make clean (sometimes helpful after upgrading upstream?)
+./configure --static --static-math=yes --static-fst=yes --use-cuda=no
 make depend
 cd ../../

ext/k3.cc

@@ -7,6 +7,7 @@
 #include "online2/online-endpoint.h"
 #include "fstext/fstext-lib.h"
 #include "lat/lattice-functions.h"
+#include "lat/lattice-functions-transition-model.h"
 #include "lat/word-align-lattice.h"
 #include "nnet3/decodable-simple-looped.h"
 
@@ -15,7 +16,7 @@
 #endif
 
 const int arate = 8000;
-    
+
 void ConfigFeatureInfo(kaldi::OnlineNnet2FeaturePipelineInfo& info,
                        std::string ivector_model_dir) {
     // Configure inline to avoid absolute paths in ".conf" files
@@ -32,7 +33,7 @@ void ConfigFeatureInfo(kaldi::OnlineNnet2FeaturePipelineInfo& info,
                     &info.ivector_extractor_info.diag_ubm);
     ReadKaldiObject(ivector_model_dir + "/final.ie",
                     &info.ivector_extractor_info.extractor);
-    
+
     info.ivector_extractor_info.num_gselect = 5;
     info.ivector_extractor_info.min_post = 0.025;
     info.ivector_extractor_info.posterior_scale = 0.1;
@@ -77,12 +78,12 @@ int main(int argc, char *argv[]) {
     using namespace kaldi;
     using namespace fst;
 
-    setbuf(stdout, NULL);  
+    setbuf(stdout, NULL);
 
     std::string nnet_dir = "exp/tdnn_7b_chain_online";
     std::string graph_dir = nnet_dir + "/graph_pp";
     std::string fst_rxfilename = graph_dir + "/HCLG.fst";
-    
+
     if(argc == 3) {
       nnet_dir = argv[1];
       graph_dir = nnet_dir + "/graph_pp";
@@ -92,7 +93,7 @@ int main(int argc, char *argv[]) {
       usage();
       return EXIT_FAILURE;
     }
-	
+
 #ifdef HAVE_CUDA
     fprintf(stdout, "Cuda enabled\n");
     CuDevice &cu_device = CuDevice::Instantiate();
@@ -102,7 +103,7 @@ int main(int argc, char *argv[]) {
 #endif
     const std::string ivector_model_dir = nnet_dir + "/ivector_extractor";
     const std::string nnet3_rxfilename = nnet_dir + "/final.mdl";
-    
+
     const std::string word_syms_rxfilename = graph_dir + "/words.txt";
     const string word_boundary_filename = graph_dir + "/phones/word_boundary.int";
     const string phone_syms_rxfilename = graph_dir + "/phones.txt";
@@ -116,7 +117,7 @@ int main(int argc, char *argv[]) {
     ConfigDecoding(nnet3_decoding_config);
     OnlineEndpointConfig endpoint_config;
     ConfigEndpoint(endpoint_config);
-    
+
 
     BaseFloat frame_shift = feature_info.FrameShiftInSeconds();
 
@@ -133,16 +134,16 @@ int main(int argc, char *argv[]) {
     nnet_simple_looped_opts.acoustic_scale = 1.0; // changed from 0.1?
 
     nnet3::DecodableNnetSimpleLoopedInfo de_nnet_simple_looped_info(nnet_simple_looped_opts, &am_nnet);
-    
+
     fst::Fst<fst::StdArc> *decode_fst = ReadFstKaldi(fst_rxfilename);
 
     fst::SymbolTable *word_syms =
       fst::SymbolTable::ReadText(word_syms_rxfilename);
 
     fst::SymbolTable* phone_syms =
       fst::SymbolTable::ReadText(phone_syms_rxfilename);
-    
-    
+
+
     OnlineIvectorExtractorAdaptationState adaptation_state(feature_info.ivector_extractor_info);
 
     OnlineNnet2FeaturePipeline feature_pipeline(feature_info);
@@ -151,7 +152,7 @@ int main(int argc, char *argv[]) {
     OnlineSilenceWeighting silence_weighting(
                                              trans_model,
                                              feature_info.silence_weighting_config);
-        
+
     SingleUtteranceNnet3Decoder decoder(nnet3_decoding_config,
                                         trans_model,
 					de_nnet_simple_looped_info,
@@ -172,7 +173,7 @@ int main(int argc, char *argv[]) {
     else if(strcmp(cmd,"reset\n") == 0) {
       feature_pipeline.~OnlineNnet2FeaturePipeline();
       new (&feature_pipeline) OnlineNnet2FeaturePipeline(feature_info);
-      
+
       decoder.~SingleUtteranceNnet3Decoder();
       new (&decoder) SingleUtteranceNnet3Decoder(nnet3_decoding_config,
                                                  trans_model,
@@ -188,11 +189,11 @@ int main(int argc, char *argv[]) {
       fgets(cmd, sizeof(cmd), stdin);
       sscanf(cmd, "%d\n", &chunk_len);
 
-      int16_t audio_chunk[chunk_len];  
+      int16_t audio_chunk[chunk_len];
       Vector<BaseFloat> wave_part = Vector<BaseFloat>(chunk_len);
-      
+
       fread(&audio_chunk, 2, chunk_len, stdin);
-      
+
       // We need to copy this into the `wave_part' Vector<BaseFloat> thing.
       // From `gst-audio-source.cc' in gst-kaldi-nnet2
       for (int i = 0; i < chunk_len ; ++i) {
@@ -214,14 +215,14 @@ int main(int argc, char *argv[]) {
       fprintf(stdout, "ok\n");
     }
     else if(strcmp(cmd, "get-final\n") == 0) {
-      feature_pipeline.InputFinished(); // XXX: this is new: what does it do?
-
+      feature_pipeline.InputFinished(); // Computes last few frames of input
+      decoder.AdvanceDecoding();        // Decodes remaining frames
       decoder.FinalizeDecoding();
 
       Lattice final_lat;
       decoder.GetBestPath(true, &final_lat);
       CompactLattice clat;
-      ConvertLattice(final_lat, &clat);      
+      ConvertLattice(final_lat, &clat);
 
       // Compute prons alignment (see: kaldi/latbin/nbest-to-prons.cc)
       CompactLattice aligned_clat;
@@ -254,12 +255,12 @@ int main(int argc, char *argv[]) {
       }
 
       fprintf(stdout, "done with words\n");
-      
+
     }
     else {
 
       fprintf(stderr, "unknown command %s\n", cmd);
-      
+
     }
   }
 }