Update app.py

app.py

@@ -1,3 +1,22 @@
+You've found another mapping inconsistency. My apologies for that. This error is happening for the same reason as the Punjabi issue, just with a different language.
+
+The Language ID (LID) model is correctly identifying the language as Hindi and outputting the simple code **`hin`**. However, our mapping dictionary was only looking for the more complex code `hin_Deva`.
+
+-----
+
+### \#\# The Correction
+
+The solution is to make our `LID_TO_ASR_LANG_MAP` more robust by including mappings for both the simple three-letter codes (like `hin`) and the complex ones (`hin_Deva`). This will ensure that no matter which code the LID model outputs, it will be correctly mapped to the two-letter code expected by the ASR model.
+
+I have updated the dictionary below to be comprehensive.
+
+-----
+
+### **Final Corrected Code**
+
+Here is the complete script with the updated and more robust mapping dictionary.
+
+```python
 from __future__ import annotations
 import torch
 import torchaudio
@@ -24,17 +43,23 @@ print("✅ Language ID Model loaded.")
 
 
 # --- Language Mappings ---
+# Updated to be more robust, handling both simple and complex LID codes.
 LID_TO_ASR_LANG_MAP = {
+    # MMS-style codes (e.g., hin_Deva)
     "asm_Beng": "as", "ben_Beng": "bn", "brx_Deva": "br", "doi_Deva": "doi",
     "guj_Gujr": "gu", "hin_Deva": "hi", "kan_Knda": "kn", "kas_Arab": "ks",
     "kas_Deva": "ks", "gom_Deva": "kok", "mai_Deva": "mai", "mal_Mlym": "ml",
     "mni_Beng": "mni", "mar_Deva": "mr", "nep_Deva": "ne", "ory_Orya": "or",
     "pan_Guru": "pa", "san_Deva": "sa", "sat_Olck": "sat", "snd_Arab": "sd",
     "tam_Taml": "ta", "tel_Telu": "te", "urd_Arab": "ur",
-    "pan": "pa"  # <-- ADDED THIS FIX FOR PUNJABI
+    # Simple 3-letter codes (e.g., hin) as a fallback
+    "as": "as", "ben": "bn", "brx": "br", "doi": "doi", "guj": "gu", "hin": "hi",
+    "kan": "kn", "kas": "ks", "kok": "kok", "mai": "mai", "mal": "ml", "mni": "mni",
+    "mar": "mr", "nep": "ne", "ory": "or", "pan": "pa", "san": "sa", "sat": "sat",
+    "snd": "sd", "tam": "ta", "tel": "te", "urd": "ur", "eng": "en"
 }
 
-ASR_CODE_TO_NAME = { "as": "Assamese", "bn": "Bengali", "br": "Bodo", "doi": "Dogri", "gu": "Gujarati", "hi": "Hindi", "kn": "Kannada", "ks": "Kashmiri", "kok": "Konkani", "mai": "Maithili", "ml": "Malayalam", "mni": "Manipuri", "mr": "Marathi", "ne": "Nepali", "or": "Odia", "pa": "Punjabi", "sa": "Sanskrit", "sat": "Santali", "sd": "Sindhi", "ta": "Tamil", "te": "Telugu", "ur": "Urdu"}
+ASR_CODE_TO_NAME = { "as": "Assamese", "bn": "Bengali", "br": "Bodo", "doi": "Dogri", "gu": "Gujarati", "hi": "Hindi", "kn": "Kannada", "ks": "Kashmiri", "kok": "Konkani", "mai": "Maithili", "ml": "Malayalam", "mni": "Manipuri", "mr": "Marathi", "ne": "Nepali", "or": "Odia", "pa": "Punjabi", "sa": "Sanskrit", "sat": "Santali", "sd": "Sindhi", "ta": "Tamil", "te": "Telugu", "ur": "Urdu", "en": "English"}
 
 
 @spaces.GPU
@@ -60,12 +85,8 @@ def transcribe_audio_with_lid(audio_path):
         asr_lang_code = LID_TO_ASR_LANG_MAP.get(detected_lid_code)
         
         if not asr_lang_code:
-            # Fallback for simple codes like 'pan' from other LID models
-            if detected_lid_code in LID_TO_ASR_LANG_MAP:
-                 asr_lang_code = LID_TO_ASR_LANG_MAP[detected_lid_code]
-            else:
-                detected_lang_str = f"Detected '{detected_lid_code}', which is not supported by the ASR model."
-                return detected_lang_str, "N/A", "N/A"
+            detected_lang_str = f"Detected '{detected_lid_code}', which is not supported by the ASR model."
+            return detected_lang_str, "N/A", "N/A"
 
         detected_lang_str = f"Detected Language: {ASR_CODE_TO_NAME.get(asr_lang_code, 'Unknown')}"
 
@@ -104,4 +125,5 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
     )
 
 if __name__ == "__main__":
-    demo.queue().launch()
+    demo.queue().launch()
+```