From 34f649f9b05f9bd4da1d18d4324b56367c2b4f24 Mon Sep 17 00:00:00 2001
From: quillotm <mathias.quillot@univ-avignon.fr>
Date: Thu, 2 Sep 2021 15:39:26 +0200
Subject: [PATCH] Adding utt2sub and sub2utt commands

---
 volia/masseffect.py | 46 +++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 45 insertions(+), 1 deletion(-)

diff --git a/volia/masseffect.py b/volia/masseffect.py
index 8828dbd..fbe55b3 100644
--- a/volia/masseffect.py
+++ b/volia/masseffect.py
@@ -92,6 +92,34 @@ def converter(file: str, outtype: str, outfile: str):
                 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="...")
@@ -132,6 +160,20 @@ if __name__ == '__main__':
     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()
@@ -142,7 +184,9 @@ if __name__ == '__main__':
         "char2utt": char2utt,
         "wavscp": wavscp,
         "changelabels": changelabels,
-        "converter": converter
+        "converter": converter,
+        "utt2sub": utt2sub,
+        "sub2utt": sub2utt
     })
 
     runner.run(args.which, args.__dict__, remove="which")
-- 
1.8.2.3