From ec8bfe9df205608282e5297635363fc8fc8fe55b Mon Sep 17 00:00:00 2001
From: JoJoBarthold2
Date: Sat, 16 Sep 2023 15:48:16 +0200
Subject: created a method that prints the lexicon

---
 swr2_asr/utils/data.py | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)

(limited to 'swr2_asr/utils/data.py')

diff --git a/swr2_asr/utils/data.py b/swr2_asr/utils/data.py
index d551c98..f484bdd 100644
--- a/swr2_asr/utils/data.py
+++ b/swr2_asr/utils/data.py
@@ -343,3 +343,24 @@ class MLSDataset(Dataset):
             dataset_lookup_entry["chapterid"],
             idx,
         )  # type: ignore
+    
+    def create_lexicon(vocab_counts_path, lexicon_path):
+      
+        words_list = []  
+        with open(vocab_counts_path, 'r') as file:
+            for line in file:
+               
+                words = line.split()
+                if len(words) >= 1:
+                   
+                    word = words[0]
+                    words_list.append(word)
+        
+        with open(lexicon_path, 'w') as file:
+            for word in words_list:
+                file.write(f"{word} ")
+                for char in word:
+                    file.write(char + ' ')
+                file.write("|")
+                        
+    
\ No newline at end of file
-- 
cgit v1.2.3