Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,183 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+arabic_recording.wav filter=lfs diff=lfs merge=lfs -text
+nid.wav filter=lfs diff=lfs merge=lfs -text
+pn.wav filter=lfs diff=lfs merge=lfs -text
+recordings/case_name_20250909_144019.wav filter=lfs diff=lfs merge=lfs -text
+recordings/case_name_20250909_144453.wav filter=lfs diff=lfs merge=lfs -text
+recordings/case_name_20250909_155450.wav filter=lfs diff=lfs merge=lfs -text
+recordings/case_name_20250909_155812.wav filter=lfs diff=lfs merge=lfs -text
+recordings/case_name_20250909_155921.wav filter=lfs diff=lfs merge=lfs -text
+recordings/case_name_20250909_170323.wav filter=lfs diff=lfs merge=lfs -text
+recordings/case_name_20250909_170702.wav filter=lfs diff=lfs merge=lfs -text
+recordings/case_name_20250924_200610.wav filter=lfs diff=lfs merge=lfs -text
+recordings/name_20250909_143929.wav filter=lfs diff=lfs merge=lfs -text
+recordings/name_20250909_143955.wav filter=lfs diff=lfs merge=lfs -text
+recordings/name_20250909_144405.wav filter=lfs diff=lfs merge=lfs -text
+recordings/name_20250909_144432.wav filter=lfs diff=lfs merge=lfs -text
+recordings/name_20250909_155412.wav filter=lfs diff=lfs merge=lfs -text
+recordings/name_20250909_155430.wav filter=lfs diff=lfs merge=lfs -text
+recordings/name_20250909_155755.wav filter=lfs diff=lfs merge=lfs -text
+recordings/name_20250909_155905.wav filter=lfs diff=lfs merge=lfs -text
+recordings/name_20250909_170307.wav filter=lfs diff=lfs merge=lfs -text
+recordings/name_20250909_170647.wav filter=lfs diff=lfs merge=lfs -text
+recordings/name_20250924_200542.wav filter=lfs diff=lfs merge=lfs -text
+recordings/name_20250924_202837.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_142059.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_143513.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_143739.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_143835.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_143901.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_144232.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_144254.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_144321.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_144751.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_144821.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_144845.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_155231.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_155548.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_155608.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_155738.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_155849.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_165246.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_165306.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_165751.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_165824.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_170138.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_170225.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_170249.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_170454.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_170510.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_170525.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_170628.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_171108.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_171128.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_171143.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_171403.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_172000.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_172022.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_172037.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250909_172215.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_190026.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_190048.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_190113.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_190139.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_193554.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_193645.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_193831.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_195919.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_200017.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_200225.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_200327.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_200456.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_200525.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_200946.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_201105.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_201847.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_201924.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_202232.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_202334.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_202518.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_202637.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_202658.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_202725.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_202856.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_202917.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_202939.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_202958.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_203018.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_203243.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_203318.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_203901.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_213039.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_213332.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_213415.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_215139.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_215742.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_220323.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_220610.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_220942.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_221305.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_221623.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_221735.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_221954.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_222031.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_222055.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250924_222126.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250925_103705.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250925_103842.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250925_104013.wav filter=lfs diff=lfs merge=lfs -text
+recordings/nid_20250925_104056.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_140723.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_141138.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_141212.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_141256.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_141359.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_142839.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_143206.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_143242.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_143308.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_143425.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_143449.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_144046.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_144111.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_144132.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_144211.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_144519.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_144552.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_144616.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_144644.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_144706.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_144728.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_155318.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_155349.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_155514.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_155531.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_155657.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_155833.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_170206.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_170436.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250909_170608.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_172013.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_173904.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_173928.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_174543.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_184304.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_184519.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_184618.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_185008.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_185556.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_185707.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_185935.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_190723.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_190828.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_190905.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_192105.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_192132.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_192201.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_193531.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_200629.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_200649.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_200714.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_200732.wav filter=lfs diff=lfs merge=lfs -text
+recordings/phone_20250924_200905.wav filter=lfs diff=lfs merge=lfs -text
+recordings/recording_20250909_140117.wav filter=lfs diff=lfs merge=lfs -text
+recordings/upload_nid_20250909_142240.wav filter=lfs diff=lfs merge=lfs -text
+recordings/upload_nid_20250909_145232.wav filter=lfs diff=lfs merge=lfs -text
+recordings/upload_nid_20250909_155235.wav filter=lfs diff=lfs merge=lfs -text
+recordings/upload_nid_20250909_155331.wav filter=lfs diff=lfs merge=lfs -text
+recordings/upload_nid_20250909_155351.wav filter=lfs diff=lfs merge=lfs -text
+recordings/upload_nid_20250909_155359.wav filter=lfs diff=lfs merge=lfs -text
+recordings/upload_nid_20250909_155418.wav filter=lfs diff=lfs merge=lfs -text
+recordings/upload_nid_20250909_155437.wav filter=lfs diff=lfs merge=lfs -text
+recordings/upload_nid_20250909_155517.wav filter=lfs diff=lfs merge=lfs -text
+recordings/upload_nid_20250909_155534.wav filter=lfs diff=lfs merge=lfs -text
+recordings/upload_nid_20250909_155550.wav filter=lfs diff=lfs merge=lfs -text
+recordings/upload_nid_20250909_155611.wav filter=lfs diff=lfs merge=lfs -text
+recordings/upload_phone_20250909_141748.wav filter=lfs diff=lfs merge=lfs -text
+recordings/upload_phone_20250909_142024.wav filter=lfs diff=lfs merge=lfs -text
+recordings/upload_phone_20250909_142041.wav filter=lfs diff=lfs merge=lfs -text
+recordings/upload_phone_20250909_142238.wav filter=lfs diff=lfs merge=lfs -text
+recordings/upload_phone_20250909_142949.wav filter=lfs diff=lfs merge=lfs -text
+stt_ar_fastconformer_large_pc_finetuned.nemo filter=lfs diff=lfs merge=lfs -text

Downloading model.txt

@@ -0,0 +1 @@
+https://huggingface.co/nvidia/stt_ar_fastconformer_hybrid_large_pcd_v1.0

app.py

@@ -0,0 +1,414 @@
+# import sounddevice as sd
+# import streamlit as st
+# import os
+# import re
+# import numpy as np
+# import wave
+# import requests
+# from datetime import datetime
+
+# API_URL = "http://localhost:8000/transcribe"
+
+# # Record audio
+# def record_audio(duration=15, fs=16000):
+#     st.write("🔴 Recording... Speak Arabic now!")
+#     recording = sd.rec(int(duration * fs), samplerate=fs, channels=1, dtype='int16')
+#     sd.wait()
+#     st.write("✅ Recording finished")
+#     return recording, fs
+
+# # Save recording
+# def save_wav(recording, fs, out_dir="recordings"):
+#     os.makedirs(out_dir, exist_ok=True)
+#     timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+#     file_path = os.path.join(out_dir, f"recording_{timestamp}.wav")
+
+#     # Normalize
+#     max_val = np.max(np.abs(recording))
+#     if max_val > 0:
+#         recording = (recording / max_val * 32767).astype(np.int16)
+
+#     with wave.open(file_path, "wb") as wf:
+#         wf.setnchannels(1)
+#         wf.setsampwidth(2)
+#         wf.setframerate(fs)
+#         wf.writeframes(recording.tobytes())
+
+#     return file_path
+
+# # Call API
+# def transcribe_audio(file_path):
+#     with open(file_path, "rb") as f:
+#         files = {"file": (file_path, f, "audio/wav")}
+#         response = requests.post(API_URL, files=files)
+
+#     if response.status_code == 200:
+#         return response.json().get("transcription", "")
+#     else:
+#         st.error(f"❌ API Error {response.status_code}: {response.text}")
+#         return ""
+
+# # Extract phone number
+# def extract_phone_number(text):
+#     match = re.search(r"01[0-9]{9}", text)
+#     return match.group(0) if match else None
+
+# # Extract national ID
+# def extract_national_id(text):
+#     digits = re.findall(r"\d+", text)
+#     candidate = "".join(digits)
+#     if len(candidate) == 14:
+#         return candidate
+#     elif len(candidate) > 14:
+#         return candidate[:14]
+#     elif 7 <= len(candidate) < 14:
+#         return f"⚠️ Incomplete ID: {candidate} ({len(candidate)} digits)"
+#     else:
+#         return None
+
+# # ---------------- UI ----------------
+# st.title("📞 Phone & National ID Capture (with Name + Case Name)")
+
+# # Session state
+# if "phone_number" not in st.session_state:
+#     st.session_state.phone_number = None
+# if "national_id" not in st.session_state:
+#     st.session_state.national_id = None
+# if "name" not in st.session_state:
+#     st.session_state.name = ""
+# if "case_name" not in st.session_state:
+#     st.session_state.case_name = ""
+
+# # Step 1: Phone number
+# st.subheader("Step 1: Provide your phone number")
+# col1, col2 = st.columns(2)
+
+# with col1:
+#     if st.button("🎙️ Record Phone Number"):
+#         rec, fs = record_audio()
+#         wav_path = save_wav(rec, fs)
+#         st.audio(wav_path)
+#         text = transcribe_audio(wav_path)
+#         st.write("📝 Transcription:", text)
+#         phone = extract_phone_number(text)
+#         if phone:
+#             st.session_state.phone_number = phone
+#             st.success(f"📱 Detected Phone Number: {phone}")
+#         else:
+#             st.error("❌ No valid phone number detected")
+
+# with col2:
+#     phone_upload = st.file_uploader("Or upload phone number audio", type=["wav", "mp3", "m4a"])
+#     if phone_upload is not None:
+#         temp_path = os.path.join("recordings", f"upload_phone_{datetime.now().strftime('%Y%m%d_%H%M%S')}.wav")
+#         os.makedirs("recordings", exist_ok=True)
+#         with open(temp_path, "wb") as f:
+#             f.write(phone_upload.read())
+#         st.audio(temp_path)
+#         text = transcribe_audio(temp_path)
+#         st.write("📝 Transcription:", text)
+#         phone = extract_phone_number(text)
+#         if phone:
+#             st.session_state.phone_number = phone
+#             st.success(f"📱 Detected Phone Number: {phone}")
+#         else:
+#             st.error("❌ No valid phone number detected")
+
+# # Step 2: National ID
+# st.subheader("Step 2: Provide your national ID")
+# col3, col4 = st.columns(2)
+
+# with col3:
+#     if st.button("🎙️ Record National ID"):
+#         rec, fs = record_audio()
+#         wav_path = save_wav(rec, fs)
+#         st.audio(wav_path)
+#         text = transcribe_audio(wav_path)
+#         st.write("📝 Transcription:", text)
+#         nid = extract_national_id(text)
+#         if nid:
+#             st.session_state.national_id = nid
+#             st.success(f"🪪 Detected National ID: {nid}")
+#         else:
+#             st.error("❌ No valid national ID detected")
+
+# with col4:
+#     nid_upload = st.file_uploader("Or upload national ID audio", type=["wav", "mp3", "m4a"])
+#     if nid_upload is not None:
+#         temp_path = os.path.join("recordings", f"upload_nid_{datetime.now().strftime('%Y%m%d_%H%M%S')}.wav")
+#         os.makedirs("recordings", exist_ok=True)
+#         with open(temp_path, "wb") as f:
+#             f.write(nid_upload.read())
+#         st.audio(temp_path)
+#         text = transcribe_audio(temp_path)
+#         st.write("📝 Transcription:", text)
+#         nid = extract_national_id(text)
+#         if nid:
+#             st.session_state.national_id = nid
+#             st.success(f"🪪 Detected National ID: {nid}")
+#         else:
+#             st.error("❌ No valid national ID detected")
+
+# # Step 3: Manual fields
+# st.subheader("Step 3: Provide additional info")
+# st.session_state.name = st.text_input("👤 Enter your Name", st.session_state.name)
+# st.session_state.case_name = st.text_input("📂 Enter Case Name", st.session_state.case_name)
+
+# # Final summary
+# st.subheader("📋 Summary")
+# if st.session_state.phone_number:
+#     st.info(f"📱 Phone Number: {st.session_state.phone_number}")
+# if st.session_state.national_id:
+#     st.info(f"🪪 National ID: {st.session_state.national_id}")
+# if st.session_state.name:
+#     st.info(f"👤 Name: {st.session_state.name}")
+# if st.session_state.case_name:
+#     st.info(f"📂 Case Name: {st.session_state.case_name}")
+
+# if st.session_state.phone_number and st.session_state.national_id and st.session_state.name and st.session_state.case_name:
+#     st.success("✅ All details captured successfully!")
+
+
+import sounddevice as sd
+import streamlit as st
+import os
+import re
+import numpy as np
+import wave
+import requests
+from datetime import datetime
+
+API_URL = "http://localhost:8070/transcribe"
+
+# Record audio
+def record_audio(duration=15, fs=16000):
+    st.write("🔴 Recording... Speak Arabic now!")
+    # recording = sd.rec(int(duration * fs), samplerate=fs, channels=1, dtype='int16')
+    recording = sd.rec(int(16000 * 10), samplerate=16000, channels=1, dtype='int16')
+    sd.wait()
+    st.write("✅ Recording finished")
+    return recording, fs
+
+# Save recording
+# def save_wav(recording, fs, out_dir="recordings", prefix="recording"):
+#     os.makedirs(out_dir, exist_ok=True)
+#     timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+#     file_path = os.path.join(out_dir, f"{prefix}_{timestamp}.wav")
+
+#     # Normalize
+#     max_val = np.max(np.abs(recording))
+#     if max_val > 0:
+#         recording = (recording / max_val * 32767).astype(np.int16)
+
+#     with wave.open(file_path, "wb") as wf:
+#         wf.setnchannels(1)
+#         wf.setsampwidth(2)
+#         wf.setframerate(fs)
+#         wf.writeframes(recording.tobytes())
+
+#     return file_path
+import scipy.io.wavfile as wav
+
+# def save_wav(recording, fs, out_dir="recordings", prefix="recording"):
+#     os.makedirs(out_dir, exist_ok=True)
+#     timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+#     file_path = os.path.join(out_dir, f"{prefix}_{timestamp}.wav")
+
+#     # Save directly using scipy.io.wavfile
+#     wav.write(file_path, fs, recording)
+
+#     return file_path
+# Call API
+def transcribe_audio(file_path):
+    with open(file_path, "rb") as f:
+        files = {"file": (file_path, f, "audio/wav")}
+        response = requests.post(API_URL, files=files)
+
+    if response.status_code == 200:
+        return response.json().get("transcription", "")
+    else:
+        st.error(f"❌ API Error {response.status_code}: {response.text}")
+        return ""
+
+
+def save_wav(recording, fs, out_dir="recordings", prefix="recording"):
+    os.makedirs(out_dir, exist_ok=True)
+    timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+    file_path = os.path.join(out_dir, f"{prefix}_{timestamp}.wav")
+
+    # Flatten to 1D if stereo-like shape
+    if recording.ndim > 1:
+        recording = recording[:, 0]
+
+    wav.write(file_path, fs, recording.astype(np.int16))
+
+    return file_path
+
+# Extract phone number
+def extract_phone_number(text):
+    digits = re.findall(r"\d+", text)
+    candidate = "".join(digits)
+    return candidate if candidate.startswith("01") and len(candidate) == 11 else None
+
+# Extract national ID
+def extract_national_id(text):
+    digits = re.findall(r"\d+", text)
+    candidate = "".join(digits)
+    if len(candidate) == 14:
+        return candidate
+    elif len(candidate) > 14:
+        return candidate[:14]
+    elif 7 <= len(candidate) < 14:
+        return f"⚠️ Incomplete ID: {candidate} ({len(candidate)} digits)"
+    else:
+        return None
+
+# ---------------- UI ----------------
+st.title("📞 Phone, National ID, Name & Case Name Capture")
+
+# Session state
+for key in ["phone_number", "national_id", "name", "case_name"]:
+    if key not in st.session_state:
+        st.session_state[key] = None
+
+# Step 1: Phone number
+st.subheader("Step 1: Provide your phone number")
+col1, col2 = st.columns(2)
+
+with col1:
+    if st.button("🎙️ Record Phone Number"):
+        rec, fs = record_audio()
+        wav_path = save_wav(rec, fs, prefix="phone")
+        st.audio(wav_path)
+        text = transcribe_audio(wav_path)
+        st.write("📝 Transcription:", text)
+        phone = extract_phone_number(text)
+        if phone:
+            st.session_state.phone_number = phone
+            st.success(f"📱 Detected Phone Number: {phone}")
+        else:
+            st.error("❌ No valid phone number detected")
+
+with col2:
+    phone_upload = st.file_uploader("Or upload phone number audio", type=["wav", "mp3", "m4a"])
+    if phone_upload is not None:
+        temp_path = os.path.join("recordings", f"upload_phone_{datetime.now().strftime('%Y%m%d_%H%M%S')}.wav")
+        os.makedirs("recordings", exist_ok=True)
+        with open(temp_path, "wb") as f:
+            f.write(phone_upload.read())
+        st.audio(temp_path)
+        text = transcribe_audio(temp_path)
+        st.write("📝 Transcription:", text)
+        phone = extract_phone_number(text)
+        if phone:
+            st.session_state.phone_number = phone
+            st.success(f"📱 Detected Phone Number: {phone}")
+        else:
+            st.error("❌ No valid phone number detected")
+
+# Step 2: National ID
+st.subheader("Step 2: Provide your national ID")
+col3, col4 = st.columns(2)
+
+with col3:
+    if st.button("🎙️ Record National ID"):
+        rec, fs = record_audio()
+        wav_path = save_wav(rec, fs, prefix="nid")
+        st.audio(wav_path)
+        text = transcribe_audio(wav_path)
+        st.write("📝 Transcription:", text)
+        nid = extract_national_id(text)
+        if nid:
+            st.session_state.national_id = nid
+            st.success(f"🪪 Detected National ID: {nid}")
+        else:
+            st.error("❌ No valid national ID detected")
+
+with col4:
+    nid_upload = st.file_uploader("Or upload national ID audio", type=["wav", "mp3", "m4a"])
+    if nid_upload is not None:
+        temp_path = os.path.join("recordings", f"upload_nid_{datetime.now().strftime('%Y%m%d_%H%M%S')}.wav")
+        os.makedirs("recordings", exist_ok=True)
+        with open(temp_path, "wb") as f:
+            f.write(nid_upload.read())
+        st.audio(temp_path)
+        text = transcribe_audio(temp_path)
+        st.write("📝 Transcription:", text)
+        nid = extract_national_id(text)
+        if nid:
+            st.session_state.national_id = nid
+            st.success(f"🪪 Detected National ID: {nid}")
+        else:
+            st.error("❌ No valid national ID detected")
+
+# Step 3: Name (audio input)
+st.subheader("Step 3: Provide your Name")
+col5, col6 = st.columns(2)
+
+with col5:
+    if st.button("🎙️ Record Name"):
+        rec, fs = record_audio()
+        wav_path = save_wav(rec, fs, prefix="name")
+        st.audio(wav_path)
+        text = transcribe_audio(wav_path)
+        st.write("📝 Transcription:", text)
+        if text.strip():
+            st.session_state.name = text.strip()
+            st.success(f"👤 Name: {text.strip()}")
+
+with col6:
+    name_upload = st.file_uploader("Or upload name audio", type=["wav", "mp3", "m4a"])
+    if name_upload is not None:
+        temp_path = os.path.join("recordings", f"upload_name_{datetime.now().strftime('%Y%m%d_%H%M%S')}.wav")
+        os.makedirs("recordings", exist_ok=True)
+        with open(temp_path, "wb") as f:
+            f.write(name_upload.read())
+        st.audio(temp_path)
+        text = transcribe_audio(temp_path)
+        st.write("📝 Transcription:", text)
+        if text.strip():
+            st.session_state.name = text.strip()
+            st.success(f"👤 Name: {text.strip()}")
+
+# Step 4: Case Name (audio input)
+st.subheader("Step 4: Provide Case Name")
+col7, col8 = st.columns(2)
+
+with col7:
+    if st.button("🎙️ Record Case Name"):
+        rec, fs = record_audio()
+        wav_path = save_wav(rec, fs, prefix="case_name")
+        st.audio(wav_path)
+        text = transcribe_audio(wav_path)
+        st.write("📝 Transcription:", text)
+        if text.strip():
+            st.session_state.case_name = text.strip()
+            st.success(f"📂 Case Name: {text.strip()}")
+
+with col8:
+    case_upload = st.file_uploader("Or upload case name audio", type=["wav", "mp3", "m4a"])
+    if case_upload is not None:
+        temp_path = os.path.join("recordings", f"upload_case_{datetime.now().strftime('%Y%m%d_%H%M%S')}.wav")
+        os.makedirs("recordings", exist_ok=True)
+        with open(temp_path, "wb") as f:
+            f.write(case_upload.read())
+        st.audio(temp_path)
+        text = transcribe_audio(temp_path)
+        st.write("📝 Transcription:", text)
+        if text.strip():
+            st.session_state.case_name = text.strip()
+            st.success(f"📂 Case Name: {text.strip()}")
+
+# Final summary
+st.subheader("📋 Summary")
+if st.session_state.phone_number:
+    st.info(f"📱 Phone Number: {st.session_state.phone_number}")
+if st.session_state.national_id:
+    st.info(f"🪪 National ID: {st.session_state.national_id}")
+if st.session_state.name:
+    st.info(f"👤 Name: {st.session_state.name}")
+if st.session_state.case_name:
+    st.info(f"📂 Case Name: {st.session_state.case_name}")
+
+if st.session_state.phone_number and st.session_state.national_id and st.session_state.name and st.session_state.case_name:
+    st.success("✅ All details captured successfully!")

app_api.py

@@ -0,0 +1,306 @@
+from fastapi import FastAPI, File, UploadFile, Form
+from fastapi.responses import JSONResponse
+import uvicorn
+import tempfile
+import nemo.collections.asr as nemo_asr
+import re
+import os
+from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+import torch
+from word2number import w2n
+from deep_translator import GoogleTranslator
+
+# ===== Arabic number mapping (expanded) =====
+arabic_numbers = {
+    # Basic digits
+    "صفر": "0", "زيرو": "0", "٠": "0","زيو": "0","زير": "0","زر": "0","زروا": "0","زرا": "0","زيره ": "0","زرو ": "0",
+    "واحد": "1", "واحدة": "1", "١": "1",
+    "اتنين": "2", "اثنين": "2", "إثنين": "2", "اثنان": "2", "إثنان": "2", "٢": "2",
+    "تلاتة": "3", "ثلاثة": "3", "٣": "3",
+    "اربعة": "4", "أربعة": "4", "٤": "4",
+    "خمسة": "5", "٥": "5",
+    "ستة": "6", "٦": "6",
+    "سبعة": "7", "٧": "7",
+    "تمانية": "8", "ثمانية": "8", "٨": "8",
+    "تسعة": "9", "٩": "9",
+
+    # Teens
+    "عشرة": "10", "١٠": "10",
+    # 11
+    "احد عشر": "11", "واحد عشر": "11", "حداشر": "11",
+    "١ عشر": "11", "1 عشر": "11", "١عشر": "11", "1عشر": "11",
+    "١١": "11", "11": "11",
+
+    # 12
+    "اثنا عشر": "12", "اثني عشر": "12", "اتناشر": "12",
+    "٢ عشر": "12", "2 عشر": "12", "٢عشر": "12", "2عشر": "12",
+    "١٢": "12", "12": "12",
+
+    # 13
+    "ثلاثة عشر": "13", "تلاتة عشر": "13", "تلتاشر": "13",
+    "٣ عشر": "13", "3 عشر": "13", "٣عشر": "13", "3عشر": "13",
+    "١٣": "13", "13": "13",
+
+    # 14
+    "أربعة عشر": "14", "اربعة عشر": "14", "اربعتاشر": "14",
+    "٤ عشر": "14", "4 عشر": "14", "٤عشر": "14", "4عشر": "14",
+    "١٤": "14", "14": "14",
+
+    # 15
+    "خمسة عشر": "15", "خمسه عشر": "15", "خمستاشر": "15",
+    "٥ عشر": "15", "5 عشر": "15", "٥عشر": "15", "5عشر": "15",
+    "١٥": "15", "15": "15",
+
+    # 16
+    "ستة عشر": "16", "سته عشر": "16", "ستاشر": "16",
+    "٦ عشر": "16", "6 عشر": "16", "٦عشر": "16", "6عشر": "16",
+    "١٦": "16", "16": "16",
+
+    # 17
+    "سبعة عشر": "17", "سبعه عشر": "17", "سبعتاشر": "17",
+    "٧ عشر": "17", "7 عشر": "17", "٧عشر": "17", "7عشر": "17",
+    "١٧": "17", "17": "17",
+
+    # 18
+    "ثمانية عشر": "18", "تمانية عشر": "18", "طمنتاشر": "18",
+    "٨ عشر": "18", "8 عشر": "18", "٨عشر": "18", "8عشر": "18",
+    "١٨": "18", "18": "18",
+
+    # 19
+    "تسعة عشر": "19", "تسعه عشر": "19", "تسعتاشر": "19",
+    "٩ عشر": "19", "9 عشر": "19", "٩عشر": "19", "9عشر": "19",
+    "١٩": "19", "19": "19",
+
+    # Tens
+    "عشرين": "20", "٢٠": "20",
+    "تلاتين": "30", "ثلاثين": "30", "٣٠": "30",
+    "اربعين": "40", "أربعين": "40", "٤٠": "40",
+    "خمسين": "50", "٥٠": "50",
+    "ستين": "60", "٦٠": "60",
+    "سبعين": "70", "٧٠": "70",
+    "تمانين": "80", "ثمانين": "80", "٨٠": "80","تمانون": "80","ثمانون": "80",
+    "تسعين": "90", "٩٠": "90",
+
+    # Hundreds
+    "مية": "100", "مائة": "100", "مئة": "100", "١٠٠": "100",
+    "ميتين": "200", "مائتين": "200",
+    "تلاتمية": "300", "ثلاثمائة": "300",
+    "اربعمية": "400", "أربعمائة": "400",
+    "خمسمية": "500", "خمسمائة": "500",
+    "ستمية": "600", "ستمائة": "600",
+    "سبعمية": "700", "سبعمائة": "700",
+    "تمانمية": "800", "ثمانمائة": "800",
+    "تسعمية": "900", "تسعمائة": "900",
+
+    # Thousands
+    "ألف": "1000", "الف": "1000", "١٠٠٠": "1000",
+    "ألفين": "2000", "الفين": "2000",
+    "تلات تلاف": "3000", "ثلاثة آلاف": "3000",
+    "اربعة آلاف": "4000", "أربعة آلاف": "4000",
+    "خمسة آلاف": "5000",
+    "ستة آلاف": "6000",
+    "سبعة آلاف": "7000",
+    "تمانية آلاف": "8000", "ثمانية آلاف": "8000",
+    "تسعة آلاف": "9000",
+
+    # Large numbers
+    "عشرة آلاف": "10000",
+    "مية ألف": "100000", "مائة ألف": "100000",
+    "مليون": "1000000", "١٠٠٠٠٠٠": "1000000",
+    "ملايين": "1000000",
+    "مليار": "1000000000", "١٠٠٠٠٠٠٠٠٠": "1000000000",
+    # ===== Compound tens (Arabic + digit forms) =====
+    "واحد وعشرون": "21", "1 وعشرون": "21",
+    "اثنان وعشرون": "22", "٢ وعشرون": "22",
+    "ثلاثة وعشرون": "23", "٣ وعشرون": "23",
+    "اربعة وعشرون": "24", "٤ وعشرون": "24",
+    "خمسة وعشرون": "25", "٥ وعشرون": "25",
+    "ستة وعشرون": "26", "٦ وعشرون": "26",
+    "سبعة وعشرون": "27", "٧ وعشرون": "27",
+    "تمانية وعشرون": "28", "ثمانية وعشرون": "28", "٨ وعشرون": "28",
+    "تسعة وعشرون": "29", "٩ وعشرون": "29",
+
+    "ثمانية وثمانون": "88", "8 وثمانون": "88",
+    "اثنان وثمانون": "82", "٢ وثمانون": "82",
+    "خمسة وستون": "65", "5 وستون": "65",
+    "ستة عشر": "16", "٦ عشر": "16",
+    "اثنا عشر": "12", "١٢": "12",
+    "ثلاثة وثلاثون": "33", "٣٣": "33", "33": "33",
+    "أربعة وأربعون": "44", "٤٤": "44", "44": "44",
+    "خمسة وخمسون": "55", "٥٥": "55", "55": "55",
+    "ستة وستون": "66", "٦٦": "66", "66": "66",
+    "سبعة وسبعون": "77", "٧٧": "77", "77": "77",
+    "ثمانية وثمانون": "88", "٨٨": "88", "88": "88",
+    "تسعة وتسعون": "99", "٩٩": "99", "99": "99",
+}
+
+def replace_arabic_numbers(text: str) -> str:
+    for word, digit in arabic_numbers.items():
+        text = re.sub(rf"\b{word}\b", digit, text)
+    return text
+
+
+# ===== FastAPI app =====
+app = FastAPI(title="Arabic ASR API", description="ASR API with NeMo and Arabic digit conversion")
+
+# Load model once on startup
+@app.on_event("startup")
+def load_model():
+    global asr_model
+    global model
+    global tokenizer
+    global device
+    #model_path = os.getenv("NEMO_MODEL_PATH", "C:/Users/thegh/Python_Projects/Expertflow/UnderProgress/Peter_Projects/nvidia_asr_eg_conformer_better_than_whisper/stt_ar_fastconformer_hybrid_large_pcd_v1.0.nemo") 
+    model_path = "C:/Users/thegh/Python_Projects/Expertflow/UnderProgress/Peter_Projects/NP_Detection_Nvidia_conformer/stt_ar_fastconformer_hybrid_large_pc_v1.0.nemo"
+    asr_model = nemo_asr.models.EncDecCTCModel.restore_from(model_path)
+    # Load once globally
+    # model_name = "alaasayed_ai/Egyptian_Arabic_to_English"
+    model_translator_name = "ukaAi/Egyptian_dialect_to_arabic"
+    tokenizer = AutoTokenizer.from_pretrained(model_translator_name)
+    model = AutoModelForSeq2SeqLM.from_pretrained(model_translator_name)
+
+    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+    model = model.to(device)
+def translate_egyptian_to_english(text: str) -> str:
+    """
+    Translates Egyptian Arabic text to English using the fine-tuned NLLB model.
+    
+    Parameters:
+    - text (str): The input Egyptian Arabic text
+    
+    Returns:
+    - str: The translated English text
+    """
+    tokenizer.src_lang = "arz_Arab"
+    forced_bos_token_id = tokenizer.convert_tokens_to_ids("eng_Latn")
+
+    inputs = tokenizer(text, return_tensors="pt", padding=True, truncation=True, max_length=512)
+    inputs = {k: v.to(device) for k, v in inputs.items()}
+
+    translated = model.generate(
+        **inputs,
+        forced_bos_token_id=forced_bos_token_id,
+        max_length=512,
+        num_beams=4,
+        early_stopping=True
+    )
+
+    return tokenizer.decode(translated[0], skip_special_tokens=True)
+# Cardinal and tens
+WORD_TO_NUM = {
+    "zero": 0, "one": 1, "two": 2, "three": 3, "four": 4, "five": 5,
+    "six": 6, "seven": 7, "eight": 8, "nine": 9, "ten": 10,
+    "eleven": 11, "twelve": 12, "thirteen": 13, "fourteen": 14,
+    "fifteen": 15, "sixteen": 16, "seventeen": 17, "eighteen": 18,
+    "nineteen": 19, "twenty": 20, "thirty": 30, "forty": 40,
+    "fifty": 50, "sixty": 60, "seventy": 70, "eighty": 80, "ninety": 90
+}
+
+# Ordinals
+ORDINAL_TO_NUM = {
+    "first": 1, "second": 2, "third": 3, "fourth": 4, "fifth": 5,
+    "sixth": 6, "seventh": 7, "eighth": 8, "ninth": 9, "tenth": 10,
+    "eleventh": 11, "twelfth": 12, "thirteenth": 13, "fourteenth": 14,
+    "fifteenth": 15, "sixteenth": 16, "seventeenth": 17, "eighteenth": 18,
+    "nineteenth": 19, "twentieth": 20, "thirtieth": 30, "fortieth": 40,
+    "fiftieth": 50, "sixtieth": 60, "seventieth": 70, "eightieth": 80, "ninetieth": 90
+}
+
+def normalize_token(token: str):
+    """Convert a single token or hyphenated token into a number if possible."""
+    token = token.lower()
+
+    # Handle ordinals
+    if token in ORDINAL_TO_NUM:
+        return ORDINAL_TO_NUM[token]
+
+    # Handle hyphenated compounds like 'thirty-nine'
+    if "-" in token:
+        parts = token.split("-")
+        nums = [WORD_TO_NUM.get(p) for p in parts if p in WORD_TO_NUM]
+        if nums:
+            return sum(nums)
+
+    # Handle normal cardinals
+    return WORD_TO_NUM.get(token)
+
+
+def words_to_numbers(phrase: str):
+    tokens = phrase.lower().strip().split()
+    nums = [normalize_token(t) for t in tokens if normalize_token(t) is not None]
+
+    if not nums:
+        return []
+
+    # Case: three tokens like "two one ninety" → 91
+    if len(nums) == 3:
+        return [int(f"{nums[0]}{nums[1]}") + nums[2]]
+
+    # Case: two tokens like "five thirty" → 35
+    if len(nums) == 2:
+        if nums[1] >= 20:
+            return [nums[0] + nums[1]]
+        else:
+            return [int("".join(str(n) for n in nums))]
+
+    # Otherwise, return each token separately
+    return nums
+
+
+def parse_numbers(text: str):
+    chunks = re.split(r"[,\.;]", text)
+    result = []
+    for chunk in chunks:
+        result.extend(words_to_numbers(chunk))
+    return " ".join(str(n) for n in result)
+@app.post("/transcribe")
+async def transcribe_audio(file: UploadFile = File(...)):
+    # Save uploaded file to a temp path
+    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp:
+        tmp.write(await file.read())
+        tmp_path = tmp.name
+
+    try:
+        # Run transcription
+        result = asr_model.transcribe([tmp_path])
+        print(result)
+        raw_text = result[0].text
+        print(raw_text)
+       
+        result = translate_egyptian_to_english(raw_text)
+
+        print("\n=== English Translation ===\n")
+        print(result)
+        
+        print(parse_numbers(result))
+        # print (w2n.word_to_num(result))
+        # Convert Arabic numbers
+        # cleaned_text = replace_arabic_numbers(raw_text)
+        # print("\n\n")
+        # print(cleaned_text)
+        # print("\n\n")
+        return JSONResponse(content={"transcription": raw_text})
+
+    finally:
+        os.remove(tmp_path)
+
+
+@app.post("/transcribe-bytes")
+async def transcribe_audio_bytes(audio_bytes: bytes = File(...)):
+    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp:
+        tmp.write(audio_bytes)
+        tmp_path = tmp.name
+
+    try:
+        result = asr_model.transcribe([tmp_path])
+        raw_text = result[0].text
+        cleaned_text = replace_arabic_numbers(raw_text)
+
+        return JSONResponse(content={"transcription": cleaned_text})
+
+    finally:
+        os.remove(tmp_path)
+
+
+if __name__ == "__main__":
+    uvicorn.run(app, host="0.0.0.0", port=8000, reload=True)

app_api_2.py

@@ -0,0 +1,490 @@
+# from fastapi import FastAPI, File, UploadFile
+# from fastapi.responses import JSONResponse
+# import uvicorn
+# import tempfile
+# import nemo.collections.asr as nemo_asr
+# import re
+# import os
+# import librosa
+# import soundfile as sf
+
+# # ===== Arabic number mapping (expanded) =====
+# arabic_numbers = {
+#     "صفر": "0", "زيرو": "0", "٠": "0", "زيو": "0", "زير": "0",
+#     "واحد": "1", "واحدة": "1", "١": "1",
+#     "اتنين": "2", "اثنين": "2", "اثنان": "2", "٢": "2",
+#     "تلاتة": "3", "ثلاثة": "3", "٣": "3","ثلاث": "3","تلات": "3",
+#     "اربعة": "4", "أربعة": "4", "٤": "4",
+#     "خمسة": "5", "٥": "5","خمسه": "5",
+#     "ستة": "6", "٦": "6",
+#     "سبعة": "7", "٧": "7","سبعه": "7",
+#     "تمانية": "8", "ثمانية": "8", "٨": "8",
+#     "تسعة": "9", "٩": "9",
+#     "عشرة": "10", "١٠": "10","عشره": "10",
+#     "حداشر": "11", "احد عشر": "11", "احداشر": "11",
+#     "اتناشر": "12", "اثنا عشر": "12",
+#     "تلتاشر": "13", "ثلاثة عشر": "13",
+#     "اربعتاشر": "14", "أربعة عشر": "14",
+#     "خمستاشر": "15", "خمسة عشر": "15",
+#     "ستاشر": "16", "ستة عشر": "16",
+#     "سبعتاشر": "17", "سبعة عشر": "17",
+#     "طمنتاشر": "18", "ثمانية عشر": "18",
+#     "تسعتاشر": "19", "تسعة عشر": "19",
+#     "عشرين": "20", "٢٠": "20",
+#     "تلاتين": "30", "ثلاثين": "30", "٣٠": "30",
+#     "اربعين": "40", "أربعين": "40", "٤٠": "40",
+#     "خمسين": "50", "٥٠": "50",
+#     "ستين": "60", "٦٠": "60",
+#     "سبعين": "70", "٧٠": "70",
+#     "تمانين": "80", "ثمانين": "80", "٨٠": "80",
+#     "تسعين": "90", "٩٠": "90",
+#     "مية": "100", "مائة": "100", "مئة": "100", "١٠٠": "100",
+#     "ميتين": "200", "مائتين": "200",
+#     "تلاتمية": "300", "ثلاثمائة": "300",
+#     "اربعمية": "400", "أربعمائة": "400",
+#     "خمسمية": "500", "خمسمائة": "500",
+#     "ستمية": "600", "ستمائة": "600",
+#     "سبعمية": "700", "سبعمائة": "700",
+#     "تمانمية": "800", "ثمانمائة": "800",
+#     "تسعمية": "900", "تسعمائة": "900",
+#     "ألف": "1000", "الف": "1000", "١٠٠٠": "1000",
+#     "ألفين": "2000", "الفين": "2000",
+#     "تلات تلاف": "3000", "ثلاثة آلاف": "3000",
+#     "اربعة آلاف": "4000", "أربعة آلاف": "4000",
+#     "خمسة آلاف": "5000",
+#     "ستة آلاف": "6000",
+#     "سبعة آلاف": "7000",
+#     "تمانية آلاف": "8000", "ثمانية آلاف": "8000",
+#     "تسعة آلاف": "9000",
+#     "عشرة آلاف": "10000",
+#     "مية ألف": "100000", "مائة ألف": "100000",
+#     "مليون": "1000000", "ملايين": "1000000",
+#     "مليار": "1000000000"
+# }
+
+# # ===== Helpers =====
+# def normalize_arabic(text: str) -> str:
+#     diacritics = re.compile(r'[\u0617-\u061A\u064B-\u0652]')
+#     text = re.sub(diacritics, '', text)
+#     text = re.sub(r'[إأآا]', 'ا', text)
+#     text = re.sub(r'ى', 'ي', text)
+#     text = re.sub(r'ؤ', 'و', text)
+#     text = re.sub(r'ئ', 'ي', text)
+#     text = re.sub(r'ة', 'ه', text)
+#     return text
+
+# def replace_arabic_numbers(text: str) -> str:
+#     for word, digit in arabic_numbers.items():
+#         text = re.sub(fr"(?:^|\s){word}(?:$|\s)", f" {digit} ", text)
+#     return " ".join(text.split())
+
+# def join_digit_sequences(text: str) -> str:
+#     tokens = text.split()
+#     out, buffer = [], []
+#     for tok in tokens:
+#         if tok.isdigit() and len(tok) == 1:
+#             buffer.append(tok)
+#         else:
+#             if buffer:
+#                 out.append("".join(buffer))
+#                 buffer = []
+#             out.append(tok)
+#     if buffer:
+#         out.append("".join(buffer))
+#     return " ".join(out)
+
+# def ensure_16k_wav(input_path, output_path):
+#     y, sr = librosa.load(input_path, sr=16000, mono=True)
+#     sf.write(output_path, y, 16000)
+
+# # ===== FastAPI app =====
+# app = FastAPI(title="Arabic ASR API", description="ASR API with NeMo and Arabic digit conversion")
+
+# @app.on_event("startup")
+# def load_model():
+#     global asr_model
+#     model_path = "C:/Users/thegh/Python_Projects/Expertflow/UnderProgress/Peter_Projects/NP_Detection_Nvidia_conformer/asr-egyptian-nemo-v2.0.nemo"
+#     asr_model = nemo_asr.models.EncDecCTCModel.restore_from(model_path)
+
+# @app.post("/transcribe")
+# async def transcribe_audio(file: UploadFile = File(...)):
+#     with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp:
+#         tmp.write(await file.read())
+#         tmp_path = tmp.name
+
+#     # Resample to 16kHz
+#     resampled_path = tmp_path.replace(".wav", "_16k.wav")
+#     ensure_16k_wav(tmp_path, resampled_path)
+
+#     try:
+#         result = asr_model.transcribe([resampled_path])
+#         raw_text = result[0].text
+
+#         raw_text = normalize_arabic(raw_text)
+#         cleaned_text = replace_arabic_numbers(raw_text)
+#         cleaned_text = join_digit_sequences(cleaned_text)
+
+#         return JSONResponse(content={"transcription": cleaned_text})
+
+#     finally:
+#         os.remove(tmp_path)
+#         if os.path.exists(resampled_path):
+#             os.remove(resampled_path)
+
+# @app.post("/transcribe-bytes")
+# async def transcribe_audio_bytes(audio_bytes: bytes = File(...)):
+#     with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp:
+#         tmp.write(audio_bytes)
+#         tmp_path = tmp.name
+
+#     resampled_path = tmp_path.replace(".wav", "_16k.wav")
+#     ensure_16k_wav(tmp_path, resampled_path)
+
+#     try:
+#         result = asr_model.transcribe([resampled_path])
+#         raw_text = result[0].text
+
+#         raw_text = normalize_arabic(raw_text)
+#         cleaned_text = replace_arabic_numbers(raw_text)
+#         cleaned_text = join_digit_sequences(cleaned_text)
+
+#         return JSONResponse(content={"transcription": cleaned_text})
+
+#     finally:
+#         os.remove(tmp_path)
+#         if os.path.exists(resampled_path):
+#             os.remove(resampled_path)
+
+# if __name__ == "__main__":
+#     uvicorn.run(app, host="0.0.0.0", port=8000, reload=True)
+from fastapi import FastAPI, File, UploadFile
+from fastapi.responses import JSONResponse
+import uvicorn
+import tempfile
+import nemo.collections.asr as nemo_asr
+import re
+import os
+import librosa
+import soundfile as sf
+
+# ===== Arabic + English number mapping (expanded) =====
+arabic_numbers = {
+    # Arabic digits
+    "صفر": "0", "زيرو": "0", "٠": "0", "زيو": "0", "زير": "0","زينوا": "0",
+    "واحد": "1", "واحدة": "1", "١": "1",
+    "اتنين": "2", "اثنين": "2", "اثنان": "2", "٢": "2",
+    "تلاتة": "3", "ثلاثة": "3", "٣": "3", "ثلاث": "3", "تلات": "3",
+    "اربعة": "4", "أربعة": "4", "٤": "4",
+    "خمسة": "5", "٥": "5", "خمسه": "5",
+    "ستة": "6", "٦": "6",
+    "سبعة": "7", "٧": "7", "سبعه": "7",
+    "تمانية": "8", "ثمانية": "8", "٨": "8",
+    "تسعة": "9", "٩": "9",
+    "عشرة": "10", "١٠": "10", "عشره": "10",
+    "حداشر": "11", "احد عشر": "11", "احداشر": "11",
+    "اتناشر": "12", "اثنا عشر": "12",
+    "تلتاشر": "13", "ثلاثة عشر": "13",
+    "اربعتاشر": "14", "أربعة عشر": "14",
+    "خمستاشر": "15", "خمسة عشر": "15",
+    "ستاشر": "16", "ستة عشر": "16",
+    "سبعتاشر": "17", "سبعة عشر": "17",
+    "طمنتاشر": "18", "ثمانية عشر": "18",
+    "تسعتاشر": "19", "تسعة عشر": "19",
+    "عشرين": "20", "٢٠": "20",
+    "تلاتين": "30", "ثلاثين": "30", "٣٠": "30",
+    "اربعين": "40", "أربعين": "40", "٤٠": "40",
+    "خمسين": "50", "٥٠": "50",
+    "ستين": "60", "٦٠": "60",
+    "سبعين": "70", "٧٠": "70",
+    "تمانين": "80", "ثمانين": "80", "٨٠": "80",
+    "تسعين": "90", "٩٠": "90",
+    "مية": "100", "مائة": "100", "مئة": "100", "١٠٠": "100",
+    "ميتين": "200", "مائتين": "200",
+    "تلاتمية": "300", "ثلاثمائة": "300",
+    "اربعمية": "400", "أربعمائة": "400",
+    "خمسمية": "500", "خمسمائة": "500",
+    "ستمية": "600", "ستمائة": "600",
+    "سبعمية": "700", "سبعمائة": "700",
+    "تمانمية": "800", "ثمانمائة": "800",
+    "تسعمية": "900", "تسعمائة": "900",
+    "ألف": "1000", "الف": "1000", "١٠٠٠": "1000",
+    "ألفين": "2000", "الفين": "2000",
+    "تلات تلاف": "3000", "ثلاثة آلاف": "3000",
+    "اربعة آلاف": "4000", "أربعة آلاف": "4000",
+    "خمسة آلاف": "5000",
+    "ستة آلاف": "6000",
+    "سبعة آلاف": "7000",
+    "تمانية آلاف": "8000", "ثمانية آلاف": "8000",
+    "تسعة آلاف": "9000",
+    "عشرة آلاف": "10000",
+    "مية ألف": "100000", "مائة ألف": "100000",
+    "مليون": "1000000", "ملايين": "1000000",
+    "مليار": "1000000000",
+
+    # English digits
+    "zero": "0", "one": "1", "two": "2", "to": "2", "too": "2", "three": "3",
+    "four": "4", "for": "4", "five": "5", "six": "6", "seven": "7",
+    "eight": "8", "nine": "9", "ten": "10",
+    "eleven": "11", "twelve": "12", "thirteen": "13", "fourteen": "14",
+    "fifteen": "15", "sixteen": "16", "seventeen": "17",
+    "eighteen": "18", "nineteen": "19", "twenty": "20",
+    "thirty": "30", "forty": "40", "fifty": "50",
+    "sixty": "60", "seventy": "70", "eighty": "80", "ninety": "90",
+    "hundred": "100", "thousand": "1000", "million": "1000000",
+
+    # Arabic variants
+    "تلاته": "3", "اربعه": "4", "سته": "6",
+    "تمانيه": "8", "ثماني": "80", "تسعه": "9",
+    "واحده": "1", "عشره": "10",
+    "حداشر": "11", "اتناشر": "12",
+    "تلاته عشر": "13", "اربعه عشر": "14",
+    "خمسه عشر": "15", "سته عشر": "16",
+    "سبعه عشر": "17", "ثمانيه عشر": "18",
+    "تسعه عشر": "19",
+
+    # English tricky forms
+    "oh": "0",
+    "double zero": "00",
+    "double one": "11",
+    "double two": "22",
+    "double three": "33",
+    "double four": "44",
+    "double five": "55",
+    "double six": "66",
+    "double seven": "77",
+    "double eight": "88",
+    "double nine": "99",
+    "for": "4", "to": "2", "too": "2",
+    "nite": "9", "fiv": "5",
+
+    # 🔹 Repeated Digits
+    "واحد واحد": "11",
+    "اثنين اثنين": "22",
+    "اتنين اتنين": "22",
+    "ثلاثة ثلاثة": "33",
+    "تلاتة تلاتة": "33",
+    "أربعة أربعة": "44",
+    "اربعة اربعة": "44",
+    "خمسة خمسة": "55",
+    "ستة ستة": "66",
+    "سبعة سبعة": "77",
+    "ثمانية ثمانية": "88",
+    "تمانية تمانية": "88",
+    "تسعة تسعة": "99",
+
+    # 🔹 Hundreds
+    "مئة": "100",
+    "مية": "100",
+    "مئتين": "200",
+    "ميتين": "200",
+    "ثلاثمية": "300",
+    "تلتمية": "300",
+    "أربعمية": "400",
+    "اربعمية": "400",
+    "خمسمية": "500",
+    "ستمية": "600",
+    "سبعمية": "700",
+    "تمانمية": "800",
+    "تسعمية": "900",
+
+    # 🔹 Teens
+    "أحد عشر": "11",
+    "حداشر": "11",
+    "اثنا عشر": "12",
+    "اتناشر": "12",
+    "ثلاثة عشر": "13",
+    "تلاتاشر": "13",
+    "أربعة عشر": "14",
+    "اربعتاشر": "14",
+    "خمسة عشر": "15",
+    "خمسطاشر": "15",
+    "ستة عشر": "16",
+    "ستاشر": "16",
+    "سبعة عشر": "17",
+    "سبعتاشر": "17",
+    "ثمانية عشر": "18",
+    "طمنتاشر": "18",
+    "تسعة عشر": "19",
+    "تسعتاشر": "19",
+
+    # 🔹 Tens
+    "عشرين": "20",
+    "تلاتين": "30",
+    "ثلاثين": "30",
+    "أربعين": "40",
+    "اربعين": "40",
+    "خمسين": "50",
+    "ستين": "60",
+    "سبعين": "70",
+    "ثمانين": "80",
+    "تمانين": "80",
+    "ثامنين": "80",
+    "تسعين": "90",
+
+    # 🔹 Mixed Word + Digits
+    "خمسة صفر": "50",
+    "ثلاثة صفر صفر": "300",
+    "تلاتة صفر صفر": "300",
+    "واحد صفر": "10",
+    "واحد صفر واحد": "101",
+    "واحد اثنين": "12",
+    "واحد اتنين": "12",
+    "واحد ثلاثة": "13",
+    "واحد تلاتة": "13",
+    "واحد خمسة": "15",
+    "عشرة عشرة": "1010",
+    # Zero
+    "صفر": "0",
+    "زيرو": "0",
+    "زيو": "0",
+    "٠": "0",
+
+    # One
+    "واحد": "1",
+    "واحدة": "1",
+    "١": "1",
+
+    # Two
+    "اتنين": "2",
+    "اتنين": "2",
+    "اثنين": "2",
+    "اثنان": "2",
+    "٢": "2",
+
+    # Three
+    "تلاتة": "3",
+    "تلات": "3",
+    "ثلاثة": "3",
+    "ثلاث": "3",
+    "٣": "3",
+
+    # Four
+    "اربعة": "4",
+    "أربعة": "4",
+    "٤": "4",
+
+    # Five
+    "خمسة": "5",
+    "خمسه": "5",
+    "٥": "5",
+
+    # Six
+    "ستة": "6",
+    "ست": "6",
+    "٦": "6",
+
+    # Seven
+    "سبعة": "7",
+    "سبعه": "7",
+    "٧": "7",
+
+    # Eight
+    "تمانية": "8",
+    "تمنية": "8",
+    "ثمانية": "8",
+    "ثماني": "8",
+    "ثمان": "8",
+    "ثمانيّة": "8",
+    "ثماني": "8",
+    "٨": "8",
+
+    # Nine
+    "تسعة": "9",
+    "تسعه": "9",
+    "٩": "9"
+}
+
+# ===== Helpers =====
+def normalize_arabic(text: str) -> str:
+    diacritics = re.compile(r'[\u0617-\u061A\u064B-\u0652]')
+    text = re.sub(diacritics, '', text)
+    text = re.sub(r'[إأآا]', 'ا', text)
+    text = re.sub(r'ى', 'ي', text)
+    text = re.sub(r'ؤ', 'و', text)
+    text = re.sub(r'ئ', 'ي', text)
+    text = re.sub(r'ة', 'ه', text)
+    return text
+
+def replace_arabic_numbers(text: str) -> str:
+    for word, digit in arabic_numbers.items():
+        text = re.sub(fr"(?:^|\s){word}(?:$|\s)", f" {digit} ", text, flags=re.IGNORECASE)
+    return " ".join(text.split())
+
+def join_digit_sequences(text: str) -> str:
+    tokens = text.split()
+    out, buffer = [], []
+    for tok in tokens:
+        if tok.isdigit() and len(tok) == 1:
+            buffer.append(tok)
+        else:
+            if buffer:
+                out.append("".join(buffer))  # join sequences like 8 5 -> 85
+                buffer = []
+            out.append(tok)
+    if buffer:
+        out.append("".join(buffer))
+    return " ".join(out)
+
+def ensure_16k_wav(input_path, output_path):
+    y, sr = librosa.load(input_path, sr=16000, mono=True)
+    sf.write(output_path, y, 16000)
+
+# ===== FastAPI app =====
+app = FastAPI(title="Arabic ASR API", description="ASR API with NeMo and Arabic/English digit conversion")
+
+@app.on_event("startup")
+def load_model():
+    global asr_model
+    model_path = "C:/Users/thegh/Python_Projects/Expertflow/UnderProgress/Peter_Projects/NP_Detection_Nvidia_conformer/stt_ar_fastconformer_hybrid_large_pc_v1.0.nemo"
+    asr_model = nemo_asr.models.EncDecCTCModel.restore_from(model_path)
+
+@app.post("/transcribe")
+async def transcribe_audio(file: UploadFile = File(...)):
+    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp:
+        tmp.write(await file.read())
+        tmp_path = tmp.name
+
+    resampled_path = tmp_path.replace(".wav", "_16k.wav")
+    ensure_16k_wav(tmp_path, resampled_path)
+
+    try:
+        result = asr_model.transcribe([resampled_path])
+        raw_text = result[0].text
+        print(raw_text)
+        raw_text = normalize_arabic(raw_text)
+        cleaned_text = replace_arabic_numbers(raw_text)
+        cleaned_text = join_digit_sequences(cleaned_text)
+
+        return JSONResponse(content={"transcription": cleaned_text})
+
+    finally:
+        os.remove(tmp_path)
+        if os.path.exists(resampled_path):
+            os.remove(resampled_path)
+
+@app.post("/transcribe-bytes")
+async def transcribe_audio_bytes(audio_bytes: bytes = File(...)):
+    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp:
+        tmp.write(audio_bytes)
+        tmp_path = tmp.name
+
+    resampled_path = tmp_path.replace(".wav", "_16k.wav")
+    ensure_16k_wav(tmp_path, resampled_path)
+
+    try:
+        result = asr_model.transcribe([resampled_path])
+        raw_text = result[0].text
+
+        raw_text = normalize_arabic(raw_text)
+        cleaned_text = replace_arabic_numbers(raw_text)
+        cleaned_text = join_digit_sequences(cleaned_text)
+
+        return JSONResponse(content={"transcription": cleaned_text})
+
+    finally:
+        os.remove(tmp_path)
+        if os.path.exists(resampled_path):
+            os.remove(resampled_path)
+
+if __name__ == "__main__":
+    uvicorn.run(app, host="0.0.0.0", port=8000, reload=True)

arabic_recording.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69e69f1a8dfd83a548a609c1b1f0b82a112447c598e71a7120c50cf599f5982a
+size 320044

creating env.txt

@@ -0,0 +1,26 @@
+python -m venv asr_env
+Set-ExecutionPolicy -Scope Process -ExecutionPolicy Bypass
+asr_env\Scripts\activate
+pip install nemo_toolkit sounddevice scipy
+pip install hydra-core==1.3.2
+pip install lightning==2.2.1
+pip install lhotse
+pip install einops
+pip install transformers
+pip install sentencepiece
+pip install pandas
+pip install jiwer
+pip install librosa
+pip install webdataset
+pip install webdataset
+pip install pyannote.core
+pip install pyannote.audio
+pip install datasets
+pip install editdistance
+pip install ipython
+pip install huggingface_hub==0.34.4
+
+$env:NEMO_MODEL_PATH="C:\Users\Administrator\Projects\NP_nvidia_conformer\stt_ar_fastconformer_hybrid_large_pcd_v1.0.nemo"
+
+python -m streamlit run app.py
+python -m uvicorn app_api:app --host 0.0.0.0 --port 8070 --reload

main.py

@@ -0,0 +1,24 @@
+import sounddevice as sd
+import scipy.io.wavfile as wav
+import nemo.collections.asr as nemo_asr
+
+# ===== SETTINGS =====
+SAMPLE_RATE = 16000
+DURATION = 10  # seconds
+OUTPUT_FILE = "arabic_recording.wav"
+
+# ===== STEP 1: Record audio =====
+print("🎙️ Recording... Speak Arabic now!")
+audio = sd.rec(int(SAMPLE_RATE * DURATION), samplerate=SAMPLE_RATE, channels=1, dtype='int16')
+sd.wait()
+wav.write(OUTPUT_FILE, SAMPLE_RATE, audio)
+print(f"✅ Recording finished. Saved as {OUTPUT_FILE}")
+
+# ===== STEP 2: Load ASR model =====
+print("📥 Loading Arabic ASR model...")
+asr_model = nemo_asr.models.EncDecCTCModel.restore_from("C:/Users/thegh/Python_Projects/Expertflow/UnderProgress/Peter_Projects/nvidia_asr_eg_conformer_better_than_whisper/stt_ar_fastconformer_hybrid_large_pcd_v1.0.nemo")
+
+# ===== STEP 3: Transcribe =====
+print("🔍 Transcribing...")
+transcription = asr_model.transcribe([OUTPUT_FILE])
+print("📝 Transcription:", transcription[0])

nid.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:28196f306c4203691a675aba248cee8b1b6086e25cc6b1b0f6e80b6f15f6d203
+size 480044

pn.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e27d4d2bbfdb49ddb5258ef3f69c670de01c165ae5311b0d1b7562a66109cd0d
+size 480044

recordings/case_name_20250909_144019.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e4ce60742dcc7e7cf5dd5471a658c96bb3a56b70fe87ab6d41cb045d063f63e0
+size 480044

recordings/case_name_20250909_144453.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69e656cf7ac6c438217ee245b5dce11a7de1770c86f3ec177bca741ae9dbd834
+size 480044

recordings/case_name_20250909_155450.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d602dce673c8040b7067e7d4e025e74c5f85a066a872539e0c97cae14ae045ae
+size 320044

recordings/case_name_20250909_155812.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:593675815f80e5378f9dc7570ae0b54723c74579f1c8e704db1377210cea1c8b
+size 320044

recordings/case_name_20250909_155921.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:deccab7fd347d90086229df8a76f102e729dc18e77ddc9ef8dd2a6118a735c58
+size 320044

recordings/case_name_20250909_170323.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35c686a2d39bde06e7946d1c62f0cbd4f38db3621d198ea3ab8caaa80ff8fd24
+size 320044

recordings/case_name_20250909_170702.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d5b1f0f2e7d90b7b543f1a73c01381cae3dcd13d53d4ba7602c8015b9c31c740
+size 320044

recordings/case_name_20250924_200610.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a1cce3c759d3c8110677f55c305b03cf32f2513bc87d4e264236df917c1421d1
+size 320044

recordings/name_20250909_143929.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1cb4526915aa263eb641995eda73f3f97d8fa64ef392eed3d7c78ec26de28979
+size 480044

recordings/name_20250909_143955.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c18f7c0ddc66ee22faeb556e19685644f6fdba42da151951917d12066e12ca4b
+size 480044

recordings/name_20250909_144405.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:156343600d42c617bc6ff3e0cf8d9a5b2496c669c3df5dd6198da2edc8789918
+size 480044

recordings/name_20250909_144432.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:17111b91c15186023b981d90f78e019ba5a08732f55829152b95e908de85e731
+size 480044

recordings/name_20250909_155412.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b3e235ff08dd58ac4473dcb61830b3efadaa592c3d8e119db6afe2292a5cde6b
+size 320044

recordings/name_20250909_155430.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e1d997cb4c861a5c85d0496bd40519b60af65fd46ae0c78fbcc7114a6e6fe6d3
+size 320044

recordings/name_20250909_155755.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6019ed92a9f018cf134d2a71977c7ee4ab9606254877c74b025a2f0e8a423ab2
+size 320044

recordings/name_20250909_155905.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20406885f1d603e292978be8f3ed00e04033c60f9470aa6b0d885bad475155e3
+size 320044

recordings/name_20250909_170307.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a5d0089543ca5d0a4ef73966f5489e328aae57b68e2a06de4d52e3536e654307
+size 320044

recordings/name_20250909_170647.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f7f0f07b624e56ac63207a23d2e54acf0b8d674b00777874001a8878657e360
+size 320044

recordings/name_20250924_200542.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:581257c00e8111835069ddb373793339857b4033735e1409b955e06b82d776bc
+size 320044

recordings/name_20250924_202837.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:06fe7a6dff2e570391f4e6821c77e5963813dfd8767b37ecc9caf21d144b25ac
+size 320044

recordings/nid_20250909_142059.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38ed48894135e6f92e0e428f27c176a79cd7519c83c3d685321cd22bb42e5b58
+size 480044

recordings/nid_20250909_143513.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c30475287c68770f2b71157faaf1d5ea707e15ceda672133f89f87c0b4756b7
+size 480044

recordings/nid_20250909_143739.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fae5fac2e1507ca414e7d49ac2a0160ec2aa4899729ba489e101569c26498103
+size 480044

recordings/nid_20250909_143835.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0effbb67fb0a569061feaf7cc7aad9ab765a8de3ce79db6c0d2ffd753456ecdb
+size 480044

recordings/nid_20250909_143901.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e04c0a855fa4daae2f3f9217279a03618f8285dedddd3b245463e67789e4d0f0
+size 480044

recordings/nid_20250909_144232.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:66f4b88a1e24a6e7e1f40858c1cf47bb6713ea628b8490e7734fc2571f799d9b
+size 480044

recordings/nid_20250909_144254.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a84fd4e3de934c50349878b256eb444ae495ff4a00ae3469612cace2065f0e56
+size 480044

recordings/nid_20250909_144321.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3678e8e832a55cd6e26f0ff60e2b981fb930aab9ccd17b4c9b1cff812bfa408c
+size 480044

recordings/nid_20250909_144751.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:059c44a3e513b9bcf2f291332cac08d85916eda6d7f9e9255aaa9ed6039dc397
+size 480044

recordings/nid_20250909_144821.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9f22ca10b2d76de478d19059d28dd069b7ec5ccacfeacedbf72428fa5ba5097
+size 480044

recordings/nid_20250909_144845.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3032493cce45f28cafa105889849b8f36297bce8b26f3f7efe12b2f672c41db
+size 480044

recordings/nid_20250909_155231.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e955511fc29e4b2f79bc965500953a21c5c7783ce0d056c9a31691011eb86be6
+size 320044

recordings/nid_20250909_155548.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:16f2c641707582dd46b23bb5a9052ccb26c9bdc037c5598cc0f08f4890caf0fa
+size 320044

recordings/nid_20250909_155608.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbabc607909fdcb37e038bb219fac5cd7ecf04e58ff5ea2e8c9c24e89911d210
+size 320044

recordings/nid_20250909_155738.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:17e8c822b743d63feac8d1f73569f11c5e8c4c194570821d153dfdc04631e6e2
+size 320044

recordings/nid_20250909_155849.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:874c046138aa8c4d027cf0ce7264be41eb9651073c68494ac97b4c9f13e646f5
+size 320044

recordings/nid_20250909_165246.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b34efe57fc87502dee62228601118b6deec6259dad5d87e66e4f5f2eec94106
+size 320044

recordings/nid_20250909_165306.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:099771f6f816cb407a270cfad86ce6a95b6b9e7fc3ba9d0cca017adb462ac396
+size 320044

recordings/nid_20250909_165751.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5634375424dc750832e26d3d163d4417e2700165d676e3a50d2ff00205cbddc
+size 320044

recordings/nid_20250909_165824.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e70a01a94969390a5f471606dcf8fd09717eda39f24aff14a48178a7b33cded6
+size 320044