Add needed functionalities to generate utt2spk and utt2char files

quillotm
1 parent 9a08c7beb3
Showing 1 changed file with 36 additions and 2 deletions Side-by-side Diff
volia/skyrim.py
@@ -29,7 +29,29 @@
             full_path = os.path.join(root, language, "wave-audio", wav)
  
             f.write(f"{id_} {full_path}\n")
-        pass
+
+
+def utt2spk(lst: str, outfile: str):
+    lst_ = read_lst(lst)
+
+    with open(outfile, "w") as f:
+        for id_ in lst:
+            splited = id_.split(",")
+            spk_id = splited[0] + "-" + splited[1]
+            f.write(id_ + " " + spk_id + "\n")
+
+
+def utt2char(lst: str, outfile: str):
+    lst_ = read_lst(lst)
+
+    with open(outfile, "w") as f:
+        for id_ in lst:
+            splited = id_.split(",")
+            char_id = splited[1]
+            f.write(id_ + " " + char_id + "\n")
+
+
+def utt2char():
     pass
  
  
  
@@ -44,12 +66,24 @@
     parser_lst2wav.add_argument("--outfile", type=str, help="output wav.scp-like file")
     parser_lst2wav.set_defaults(which="lst2wav")
  
+    parser_utt2spk = subparsers.add_parser("utt2spk", help="generate utt2spk file")
+    parser_utt2spk.add_argument("--lst", required=True, type=str, help="list file .lst")
+    parser_utt2spk.add_argument("--outfile", required=True, type=str, help="utt2spk output file")
+    parser_utt2spk.set_defaults(which="utt2spk")
+
+    parser_utt2char = subparsers.add_parser("utt2char", help="..")
+    parser_utt2char.add_argument("--lst", required=True, type=str, help="list file .lst")
+    parser_utt2char.add_argument("--outfile", required=True, type=str, help="utt2char output file")
+    parser_utt2char.set_defaults(which="utt2char")
+
     # Parse
     args = parser.parse_args()
  
     # Run commands
     runner = SubCommandRunner({
-        "lst2wav" : lst2wav
+        "lst2wav" : lst2wav,
+        "utt2spk": utt2spk,
+        "utt2char": utt2char
     })
  
     runner.run(args.which, args.__dict__, remove="which")
...	...	@@ -29,7 +29,29 @@
29	29	full_path = os.path.join(root, language, "wave-audio", wav)
30	30
31	31	f.write(f"{id_} {full_path}\n")
32		- pass
	32	+
	33	+
	34	+def utt2spk(lst: str, outfile: str):
	35	+ lst_ = read_lst(lst)
	36	+
	37	+ with open(outfile, "w") as f:
	38	+ for id_ in lst:
	39	+ splited = id_.split(",")
	40	+ spk_id = splited[0] + "-" + splited[1]
	41	+ f.write(id_ + " " + spk_id + "\n")
	42	+
	43	+
	44	+def utt2char(lst: str, outfile: str):
	45	+ lst_ = read_lst(lst)
	46	+
	47	+ with open(outfile, "w") as f:
	48	+ for id_ in lst:
	49	+ splited = id_.split(",")
	50	+ char_id = splited[1]
	51	+ f.write(id_ + " " + char_id + "\n")
	52	+
	53	+
	54	+def utt2char():
33	55	pass
34	56
35	57
36	58
...	...	@@ -44,12 +66,24 @@
44	66	parser_lst2wav.add_argument("--outfile", type=str, help="output wav.scp-like file")
45	67	parser_lst2wav.set_defaults(which="lst2wav")
46	68
	69	+ parser_utt2spk = subparsers.add_parser("utt2spk", help="generate utt2spk file")
	70	+ parser_utt2spk.add_argument("--lst", required=True, type=str, help="list file .lst")
	71	+ parser_utt2spk.add_argument("--outfile", required=True, type=str, help="utt2spk output file")
	72	+ parser_utt2spk.set_defaults(which="utt2spk")
	73	+
	74	+ parser_utt2char = subparsers.add_parser("utt2char", help="..")
	75	+ parser_utt2char.add_argument("--lst", required=True, type=str, help="list file .lst")
	76	+ parser_utt2char.add_argument("--outfile", required=True, type=str, help="utt2char output file")
	77	+ parser_utt2char.set_defaults(which="utt2char")
	78	+
47	79	# Parse
48	80	args = parser.parse_args()
49	81
50	82	# Run commands
51	83	runner = SubCommandRunner({
52		- "lst2wav" : lst2wav
	84	+ "lst2wav" : lst2wav,
	85	+ "utt2spk": utt2spk,
	86	+ "utt2char": utt2char
53	87	})
54	88
55	89	runner.run(args.which, args.__dict__, remove="which")