Delete normalize_bm_input.py

normalize_bm_input.py

@@ -1,80 +0,0 @@
-import re
-
-# Define the de-contraction dictionary.
-# Keys are the contracted forms (what you want to replace).
-# Values are the expanded forms (what you want to replace them with).
-DE_CONTRACTIONS = {
-    # Keys with apostrophes/special characters for multi-word expansion
-    "k'a": "ka a",
-    "a b'a": "a be a",
-    "n'be": "ne be",
-    "n'b'a":"ne be a",
-    "b'a": "be a",
-    "k'o": "ko o", # Corrected key-value based on original request
-    "b'i": "be i",
-    "k'i":"ka i",
-    "k'aw":"ka aw",
-    
-    # Single-word keys (no apostrophe) for multi-word expansion
-    "kɔkɔ": "kɔgɔ",
-    "bɛ": "be"
-}
-
-def normalize_bm_input(text: str) -> str:
-    """
-    De-contracts (expands) specific contracted forms in a string
-    based on the DE_CONTRACTIONS dictionary.
-    """
-    
-    # 1. Ensure the text is lowercase for consistent matching
-    text = text.lower()
-    
-    # --- Part 1: Handle Multi-Word Expansions ---
-    
-    # The condition for 'multi-word expansion' must check the VALUE (the expanded form)
-    # not the KEY (the contracted form).
-    multi_word_expansions = {k: v for k, v in DE_CONTRACTIONS.items() if ' ' in v}
-    
-    # Sort keys (contracted forms) by length descending. This is CRUCIAL 
-    # for regex to match longer contracted forms (e.g., "a b'a") before
-    # shorter ones that might be contained within them.
-    sorted_multi_word = sorted(multi_word_expansions.items(), key=lambda item: len(item[0]), reverse=True)
-    
-    # Apply replacement for contracted forms that expand to multi-word phrases
-    for contracted_form, expanded_phrase in sorted_multi_word:
-        
-        # Create a pattern to match the full contracted form, ensuring it's 
-        # surrounded by word boundaries. This ensures "b'a" is not matched 
-        # within "b'adi".
-        pattern = r'\b' + re.escape(contracted_form) + r'\b'
-        
-        # Replace the full matched pattern with the expanded phrase
-        text = re.sub(pattern, expanded_phrase, text)
-
-    # --- Part 2: Handle Single-Word Expansions (e.g., 'kɔkɔ' -> 'kɔgɔ') ---
-    
-    # Filter for contractions that expand to a single word (no spaces in the value)
-    single_word_expansions = {k: v for k, v in DE_CONTRACTIONS.items() if ' ' not in v}
-    
-    def replace_single_word(match):
-        """Looks up the matched word (key) and returns the single-word expansion (value)."""
-        word = match.group(0)
-        # Use .get() to replace only the words present in the dictionary.
-        return single_word_expansions.get(word, word)
-        
-    # Apply the replacement function to all whole words
-    # This also catches cases like kɔkɔ and bɛ.
-    text = re.sub(r'\b\S+\b', replace_single_word, text)
-    
-    # 2. Capitalize the first letter of the result for presentation
-    return text[:1].upper() + text[1:]
-
-# --- Example Usage ---
-
-#input_text_4 = "k'a di a b'i fɛ kɔkɔ n'b'a fɔ. Bɛ jɛ."
-
-#print(f"Original Text: {input_text_4}")
-#normalized_4 = normalize_bm_input(input_text_4)
-#print(f"Normalized Text: {normalized_4}\n")
-
-# Expected Output: Ka a di a be i fɛ kɔgɔ ne be a fɔ. Be jɛ.