correction importation

README.md

@@ -115,7 +115,7 @@ python3 decouper_wav_mots_uniquement.py annotation file "mot1" "mot2" "mot3" ...
 ```
 
 ## 2. Matching with similarity
-### • This script (inspired by [the github of the paper : Spoken-Term Discovery using Discrete Speech Units](https://github.com/bshall/dusted)) assumes you have a dusted folder with the match_2.py script in the folder
+### • This script (inspired by [the github of the paper : Spoken-Term Discovery using Discrete Speech Units](https://github.com/bshall/dusted)). The match_2.py script is adapted from the match.py script of the github of DUSTED
 **matchQueryHS.py :** Output : csv files for every query with score of matching 
 
 ```bash

matchQueryHS.py

@@ -12,8 +12,8 @@ import numba
 
 
 def print_dtw_terminal(x_tokens, y_tokens, W1, W2, threshold):
-    from dusted.match_2 import match_rescore
-    from dusted.match_2 import score
+    from match_2 import match_rescore
+    from match_2 import score
     x_tokens = np.array(x_tokens, dtype=np.int64)
     y_tokens = np.array(y_tokens, dtype=np.int64)
     print(x_tokens)
@@ -94,7 +94,7 @@ def process_pair(target_path, match, sim, query_path, W1, W2, threshold=None, mi
 
 def do_match(args, query_path, W1_value, W2_value, out_path, visualize=False):
     segment_paths = sorted(args.segments_dir.rglob("*.npz"))
-    from dusted.match_2 import match_rescore as match_fn
+    from match_2 import match_rescore as match_fn
     print(segment_paths)
     all_tokens = []
     for seg_path in segment_paths + [query_path]: