{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "71d69be2",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torchaudio\n",
    "import numpy as np \n",
    "import torch\n",
    "import pandas as pd\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "eb5c6da2",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Salah1.wav',\n",
       " 'Salah2.wav',\n",
       " 'Salah3.wav',\n",
       " 'Salah4.wav',\n",
       " 'Salah5.wav',\n",
       " 'Salah6.wav',\n",
       " 'Salah7.wav',\n",
       " 'Salah8.wav',\n",
       " 'Salah9.wav',\n",
       " 'Salah10.wav']"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "files = os.listdir(\"./\")\n",
    "files = [x for x in files if \".wav\" in x]\n",
    "files = [f\"Salah{i}.wav\" for i in range(1,11)]\n",
    "files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "b2be1d8e",
   "metadata": {},
   "outputs": [],
   "source": [
    "words = {}\n",
    "words[1] =  \"نحب ماكلة بنينة كسكروت نظيف و رخيص\"\n",
    "words[2]=  \"باهي وقتاش نمشيو ال تونس\"\n",
    "words[3] = \"اعطيني خمسة الاف و خمسة ميا بلاهي\"\n",
    "words[4] = \"تعبت هاني راكش في الدار\"\n",
    "words[5] = \"نهار السبت ماشي نقرى ان شاء الله\"\n",
    "words[6]= \"زعما نلقى أحمد في الستاد ولا ماهوش هوني\"\n",
    "words[7]= \"نحب نمشي ال بنزرت نرتاح شوية\"\n",
    "words[8] = \"حكيت مع لولاد قالولي كل شي مريقل نهار السبت\"\n",
    "words[9] = \"ناكل كفتاجي و نجم نشري شوية حوت زادة\"\n",
    "words[10] = \"انتي خويا و عشيري صالح نحبك\"\n",
    "words = [words[i] for i in range(1,11)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "c46588ba",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([1, 238080])\n",
      "torch.Size([1, 184320])\n",
      "torch.Size([1, 207360])\n",
      "torch.Size([1, 168960])\n",
      "torch.Size([1, 168960])\n",
      "torch.Size([1, 192000])\n",
      "torch.Size([1, 184320])\n",
      "torch.Size([1, 199680])\n",
      "torch.Size([1, 230400])\n",
      "torch.Size([1, 192000])\n"
     ]
    }
   ],
   "source": [
    "durations= []\n",
    "path_jz = \"samples/\"\n",
    "paths  = [os.path.join(path_jz,x) for x in files]\n",
    "srs= [48000 for x in paths]\n",
    "IDs=[]\n",
    "for f in files: \n",
    "    x,sr = torchaudio.load(f)\n",
    "    new_audio = torch.mean(x, dim=0).unsqueeze(0)\n",
    "    print(new_audio.shape)\n",
    "    torchaudio.save(os.path.join(\"monoaudiotun\", f), new_audio, sr)\n",
    "    duration = float(x.shape[1]) / sr\n",
    "    durations.append(duration)\n",
    "    IDs.append(f.split(\".\")[0])\n",
    "    \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b71db098",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_salah = pd.DataFrame(\n",
    "    {'ID': IDs,\n",
    "     'duration': durations,\n",
    "     'wav': paths,\n",
    "     \"sr\": srs,\n",
    "     \"wrd\": words\n",
    "    })\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "b3fdd365",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_salah.to_csv(\"test_salah_local.csv\", index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "f6ac8451",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "%WER 45.59 [ 31 / 68, 3 ins, 7 del, 21 sub ]\n",
      "%SER 90.00 [ 9 / 10 ]\n",
      "Scored 10 sentences, 0 not present in hyp.\n",
      "================================================================================\n",
      "ALIGNMENTS\n",
      "\n",
      "Format:\n",
      "<utterance-id>, WER DETAILS\n",
      "<eps> ; reference  ; on ; the ; first ;  line\n",
      "  I   ;     S      ; =  ;  =  ;   S   ;   D  \n",
      " and  ; hypothesis ; on ; the ; third ; <eps>\n",
      "================================================================================\n",
      "Salah4, %WER 0.00 [ 0 / 5, 0 ins, 0 del, 0 sub ]\n",
      "تعبت ; هاني ; راكش ; في ; الدار\n",
      " =   ;  =   ;  =   ; =  ;   =  \n",
      "تعبت ; هاني ; راكش ; في ; الدار\n",
      "================================================================================\n",
      "Salah5, %WER 57.14 [ 4 / 7, 0 ins, 1 del, 3 sub ]\n",
      "نهار ; السبت ; ماشي ; نقرى ; ان ;   شاء   ;  الله\n",
      " =   ;   =   ;  =   ;  S   ; S  ;    S    ;   D  \n",
      "نهار ; السبت ; ماشي ; نقرا ; إن ; شاءالله ; <eps>\n",
      "================================================================================\n",
      "Salah2, %WER 60.00 [ 3 / 5, 0 ins, 1 del, 2 sub ]\n",
      "باهي ; وقتاش ; نمشيو  ;   ال  ;  تونس\n",
      " =   ;   =   ;   S    ;   S   ;   D  \n",
      "باهي ; وقتاش ; نمشيوا ; لتونس ; <eps>\n",
      "================================================================================\n",
      "Salah7, %WER 33.33 [ 2 / 6, 0 ins, 1 del, 1 sub ]\n",
      "نحب ; نمشي ;   ال   ; بنزرت ; نرتاح ; شوية\n",
      " =  ;  =   ;   S    ;   D   ;   =   ;  =  \n",
      "نحب ; نمشي ; لبنزرت ; <eps> ; نرتاح ; شوية\n",
      "================================================================================\n",
      "Salah6, %WER 37.50 [ 3 / 8, 0 ins, 0 del, 3 sub ]\n",
      "زعما ; نلقى ; أحمد ; في ; الستاد ; ولا  ; ماهوش ; هوني\n",
      " S   ;  =   ;  =   ; =  ;   S    ;  S   ;   =   ;  =  \n",
      "زعمة ; نلقى ; أحمد ; في ; السعد  ; وإلا ; ماهوش ; هوني\n",
      "================================================================================\n",
      "Salah10, %WER 83.33 [ 5 / 6, 1 ins, 1 del, 3 sub ]\n",
      "انتي ; <eps> ; خويا ;   و    ; عشيري ;  صالح ; نحبك\n",
      " S   ;   I   ;  =   ;   S    ;   S   ;   D   ;  =  \n",
      "إنت  ;   ي   ; خويا ; وعشيلي ;  صلاح ; <eps> ; نحبك\n",
      "================================================================================\n",
      "Salah8, %WER 44.44 [ 4 / 9, 2 ins, 0 del, 2 sub ]\n",
      "حكيت ; مع ;  لولاد  ; قالولي ; كل ; شي ; مريقل ; <eps> ; <eps> ; نهار ; السبت\n",
      " =   ; =  ;    S    ;   =    ; =  ; =  ;   S   ;   I   ;   I   ;  =   ;   =  \n",
      "حكيت ; مع ; الأولاد ; قالولي ; كل ; شي ;   مر  ;   ي   ;   ل   ; نهار ; السبت\n",
      "================================================================================\n",
      "Salah3, %WER 85.71 [ 6 / 7, 0 ins, 1 del, 5 sub ]\n",
      "اعطيني ; خمسة ; الاف ;   و   ; خمسة ;   ميا   ; بلاهي\n",
      "  S    ;  =   ;  S   ;   S   ;  S   ;    S    ;   D  \n",
      "أعطيني ; خمسة ; آلاف ; وخمسة ; ملا  ; باللاهي ; <eps>\n",
      "================================================================================\n",
      "Salah9, %WER 25.00 [ 2 / 8, 0 ins, 1 del, 1 sub ]\n",
      "ناكل ; كفتاجي ;   و   ;  نجم  ; نشري ; شوية ; حوت ; زادة\n",
      " =   ;   =    ;   S   ;   D   ;  =   ;  =   ;  =  ;  =  \n",
      "ناكل ; كفتاجي ; وننجم ; <eps> ; نشري ; شوية ; حوت ; زادة\n",
      "================================================================================\n",
      "Salah1, %WER 28.57 [ 2 / 7, 0 ins, 1 del, 1 sub ]\n",
      "نحب ; ماكلة ; بنينة ; كسكروت ; نظيف ;   و   ;  رخيص\n",
      " =  ;   =   ;   =   ;   =    ;  =   ;   S   ;   D  \n",
      "نحب ; ماكلة ; بنينة ; كسكروت ; نظيف ; ورخيص ; <eps>\n"
     ]
    }
   ],
   "source": [
    "filein = \"wer_test_salah.txt\"\n",
    "with open(filein, \"r\") as wer : \n",
    "    lines = wer.read().splitlines()\n",
    "    print(\"\\n\".join(lines))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}