modify export_models.py

fun_text_processing/inverse_text_normalization/export_models.py

@@ -21,6 +21,12 @@ def parse_args():
     parser.add_argument(
         "--language", help="language", choices=['de', 'en', 'es', 'fr', 'id', 'ja', 'ko', 'pt', 'ru', 'vi', 'zh'], default="en", type=str
     )
+
+    parser.add_argument(
+        "--token_and_classify_and_verbalize", help="export the single token&classify and verbalize or combined", choices=['single', 'combine'],
+        default="single", type=str
+    )
+
     parser.add_argument(
         "--export_dir",
         help="path to export directory. Default to current directory.",
@@ -77,10 +83,17 @@ if __name__ == "__main__":
     os.makedirs(export_dir, exist_ok=True)
     tagger_far_file = os.path.join(export_dir, args.language + "_itn_tagger.far")
     verbalizer_far_file = os.path.join(export_dir, args.language + "_itn_verbalizer.far")
+    tager_and_verbalizer_far_file = os.path.join(export_dir, args.language, "model.far" )
 
     start_time = perf_counter()
     tagger_fst, verbalizer_fst = get_grammars(args.language)
-    generator_main(tagger_far_file, {"tokenize_and_classify": tagger_fst})
-    generator_main(verbalizer_far_file, {"verbalize": verbalizer_fst})
+
+    if args.token_and_classify_and_verbalize == 'single':
+        generator_main(tagger_far_file, {"tokenize_and_classify": tagger_fst})
+        generator_main(verbalizer_far_file, {"verbalize": verbalizer_fst})
+    elif args.token_and_classify_and_verbalize == 'combine':
+        if not os.path.exists(os.path.join(export_dir, args.language)):
+            os.makedirs(os.path.join(export_dir, args.language))
+        generator_main(tager_and_verbalizer_far_file, {"tokenize_and_classify": tagger_fst, "verbalize": verbalizer_fst})
     print(f'Time to generate graph: {round(perf_counter() - start_time, 2)} sec')
 

fun_text_processing/version.txt

@@ -1 +1 @@
-0.1.0
+0.1.1