From 482306aa33024473a765db484c101119102b5140 Mon Sep 17 00:00:00 2001
From: Mathias Quillot <mathiasquillot@gmail.com>
Date: Thu, 2 Sep 2021 15:43:17 +0200
Subject: [PATCH] new utt2sub and sub2utt command

---
 volia/masseffect.py | 73 ++++++++++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 72 insertions(+), 1 deletion(-)

diff --git a/volia/masseffect.py b/volia/masseffect.py
index 3cbb813..c1be671 100644
--- a/volia/masseffect.py
+++ b/volia/masseffect.py
@@ -78,6 +78,48 @@ def changelabels(source: str, labels: str, outfile: str):
             core.data.write_line(",".join(splited), data_dict[key], out=f)
 
 
+def converter(file: str, outtype: str, outfile: str):
+    data = core.data.read_id_values(file)
+
+    with open(outfile, "w") as of:
+        for key in data:
+            splited = key.replace("\n", "").split(",")
+            masseffect_id = key.replace("\n", "")
+            kaldi_id = ",".join([splited[0], splited[1], splited[3]])
+            if outtype == "masseffect2kaldi":
+                of.write(f"{masseffect_id} {kaldi_id}\n")
+            elif outtype == "kaldi2masseffect":
+                of.write(f"{kaldi_id} {masseffect_id}\n")
+
+
+def utt2sub(self, file: str, subfile: str, outfile: str):
+    data = core.data.read_id_values(file)
+    keys = [key for key in data]
+
+    data_sub = core.data.read_id_values(subfile)
+    keys_sub = [key for key in data_sub]
+
+    with open(outfile) as of:
+        for key in keys:
+            subkeys = [subkey for subkey in keys_sub if subkey.startswith(key)]
+            subkeys_str = " ".join(subkeys)
+            of.write(f"{key} {subkeys_str}")
+
+
+def sub2utt(self, file: str, subfile: str, outfile: str):
+    data = core.data.read_id_values(file)
+    keys = [key for key in data]
+
+    data_sub = core.data.read_id_values(subfile)
+    keys_sub = [key for key in data_sub]
+
+    with open(outfile) as of:
+        for key in keys:
+            subkeys = [subkey for subkey in keys_sub if subkey.startswith(key)]
+            for subkey in subkeys:
+                of.write(f"{subkey} {key}")
+
+
 if __name__ == '__main__':
     # Main parser
     parser = argparse.ArgumentParser(description="...")
@@ -108,6 +150,31 @@ if __name__ == '__main__':
     parser_changelabels.add_argument("--outfile", required=True, type=str, help="Output file")
     parser_changelabels.set_defaults(which="changelabels")
 
+    # Create converter
+    parser_converter = subparsers.add_parser("converter", help="Create converter file")
+    parser_converter.add_argument("--file",
+                                  type=str,
+                                  required=True,
+                                  help="File with ids from which create converter.")
+    parser_converter.add_argument("--outtype", type=str, choices=["kaldi2masseffect", "masseffect2kaldi"])
+    parser_converter.add_argument("--outfile", type=str, required=True, help="")
+    parser_converter.set_defaults(which="converter")
+
+    # Create utt2sub
+    parser_utt2sub = subparsers.add_parser("utt2sub", help="generate utt2sub file")
+    parser_utt2sub.add_argument("--file", required=True, type=str, help="features, list or labels file with normal ids")
+    parser_utt2sub.add_argument("--subfile", required=True, type=str, help="features, list or labels file with sub ids")
+    parser_utt2sub.add_argument("--outfile", required=True, type=str, help="output file")
+    parser_utt2sub.set_defaults(which="utt2sub")
+
+    # Create sub2utt
+    parser_sub2utt = subparsers.add_parser("sub2utt", help="generate sub2utt file")
+    parser_sub2utt.add_argument("--file", required=True, type=str, help="features, list or labels file with normal ids")
+    parser_sub2utt.add_argument("--subfile", required=True, type=str, help="features, list or labels file sub ids")
+    parser_sub2utt.add_argument("--outfile", required=True, type=str, help="output file")
+    parser_sub2utt.set_defaults(which="sub2utt")
+
+
     # Parse
     args = parser.parse_args()
 
@@ -116,7 +183,11 @@ if __name__ == '__main__':
         "utt2char" : utt2char,
         "char2utt": char2utt,
         "wavscp": wavscp,
-        "changelabels": changelabels
+        "changelabels": changelabels,
+        "converter": converter,
+        "utt2sub": utt2sub,
+        "sub2utt": sub2utt
     })
 
     runner.run(args.which, args.__dict__, remove="which")
+
-- 
1.8.2.3