Code for "Learn Amazon Transcribe"

transcribe_demo.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "71e1dd2c-b819-4535-a811-01ce5b0ef83a",
   "metadata": {},
   "source": [
    "## Amazon Transcribe Demo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "3c233341-9a88-41a6-bbf9-14c032f8c26b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-07-09T02:30:57.424543Z",
     "iopub.status.busy": "2024-07-09T02:30:57.423709Z",
     "iopub.status.idle": "2024-07-09T02:30:58.461834Z",
     "shell.execute_reply": "2024-07-09T02:30:58.461290Z",
     "shell.execute_reply.started": "2024-07-09T02:30:57.424518Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: boto3 in /Users/asimjaws/.pyenv/versions/3.11.6/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages (1.34.129)\n",
      "Requirement already satisfied: botocore<1.35.0,>=1.34.129 in /Users/asimjaws/.pyenv/versions/3.11.6/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages (from boto3) (1.34.129)\n",
      "Requirement already satisfied: jmespath<2.0.0,>=0.7.1 in /Users/asimjaws/.pyenv/versions/3.11.6/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages (from boto3) (1.0.1)\n",
      "Requirement already satisfied: s3transfer<0.11.0,>=0.10.0 in /Users/asimjaws/.pyenv/versions/3.11.6/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages (from boto3) (0.10.1)\n",
      "Requirement already satisfied: python-dateutil<3.0.0,>=2.1 in /Users/asimjaws/.pyenv/versions/3.11.6/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages (from botocore<1.35.0,>=1.34.129->boto3) (2.9.0.post0)\n",
      "Requirement already satisfied: urllib3!=2.2.0,<3,>=1.25.4 in /Users/asimjaws/.pyenv/versions/3.11.6/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages (from botocore<1.35.0,>=1.34.129->boto3) (1.26.18)\n",
      "Requirement already satisfied: six>=1.5 in /Users/asimjaws/.pyenv/versions/3.11.6/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages (from python-dateutil<3.0.0,>=2.1->botocore<1.35.0,>=1.34.129->boto3) (1.16.0)\n"
     ]
    }
   ],
   "source": [
    "!pip install boto3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "3906e38a-2cdd-49fa-a1e5-af51d7510c63",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-07-09T02:30:59.933588Z",
     "iopub.status.busy": "2024-07-09T02:30:59.933195Z",
     "iopub.status.idle": "2024-07-09T02:31:00.995033Z",
     "shell.execute_reply": "2024-07-09T02:31:00.994434Z",
     "shell.execute_reply.started": "2024-07-09T02:30:59.933559Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "upload: ./demo1.mp3 to s3://asimj-test/x-transcribe-demo/demo1.mp3  \n"
     ]
    }
   ],
   "source": [
    "!aws s3 cp demo1.mp3 s3://asimj-test/x-transcribe-demo/demo1.mp3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "030399e0-4f3c-4629-9d35-9704ca8d37a3",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-07-09T02:31:02.780958Z",
     "iopub.status.busy": "2024-07-09T02:31:02.780366Z",
     "iopub.status.idle": "2024-07-09T02:31:02.871916Z",
     "shell.execute_reply": "2024-07-09T02:31:02.871647Z",
     "shell.execute_reply.started": "2024-07-09T02:31:02.780916Z"
    }
   },
   "outputs": [],
   "source": [
    "import sys, os, time, boto3, json\n",
    "from datetime import datetime\n",
    "\n",
    "# Exponentially backoff with sleep\n",
    "def exp_backoff_sleep(start_seconds, initial_backoff=4.0):\n",
    "    elapsed_seconds = time.time() - start_seconds\n",
    "    if elapsed_seconds < initial_backoff:\n",
    "        backoff = initial_backoff\n",
    "    else:\n",
    "        backoff = elapsed_seconds * 0.25\n",
    "    time.sleep(backoff)\n",
    "    app_log(f\"Waiting ... Elapsed time: {elapsed_seconds :.2f} seconds\")\n",
    "\n",
    "# Log status and debugging messages \n",
    "def app_log(msg,val_dict={}):\n",
    "  timestamp_string = datetime.now().strftime(\"[%Y-%m-%d %H:%M:%S.\") + \\\n",
    "      f\"{datetime.now().microsecond // 10000:02d}\" + \"]\"\n",
    "  val_string = ' '.join(f\"{key}={value}\" for key, value in val_dict.items())\n",
    "  output_string = \" \".join([timestamp_string, msg, val_string])\n",
    "  print(output_string, file=sys.stderr)\n",
    "  sys.stderr.flush()\n",
    "\n",
    "def convert_s3_audio_to_transcript(s3_audio_path):\n",
    "    # Initialize the boto3 clients for S3 and Transcribe\n",
    "    s3 = boto3.client('s3')\n",
    "    transcribe = boto3.client('transcribe')\n",
    "\n",
    "    # Extract the bucket and object key from the S3 audio file path\n",
    "    bucket, s3_key = s3_audio_path.replace(\"s3://\", \"\").split('/', 1)\n",
    "\n",
    "    # Formatting the current timestamp for job name and output key\n",
    "    timestamp = datetime.strftime(datetime.now(), '%Y-%m-%d-%H-%M-%S-%f')\n",
    "    output_key = f\"{s3_key.rsplit('.', 1)[0]}-{timestamp}.json\"\n",
    "    s3_output_path = f\"s3://{bucket}/{output_key}\"\n",
    "    \n",
    "    # Setting up the transcription job\n",
    "    job_name = f\"TranscriptionJob-{timestamp}\"\n",
    "    job_uri = f\"s3://{bucket}/{s3_key}\"\n",
    "\n",
    "    # Start the transcription job\n",
    "    target_language = 'en-US'\n",
    "    transcribe.start_transcription_job(\n",
    "        TranscriptionJobName=job_name,\n",
    "        Media={'MediaFileUri': job_uri},\n",
    "        MediaFormat=s3_key.split('.')[-1],\n",
    "        LanguageCode=target_language,  \n",
    "        OutputBucketName=bucket,\n",
    "        OutputKey=output_key)\n",
    "\n",
    "    # Wait for the transcription job to complete\n",
    "    app_log(\"Transcription started\")\n",
    "    start_seconds = time.time()\n",
    "    while True:\n",
    "        status = transcribe.get_transcription_job(TranscriptionJobName=job_name)\n",
    "        if status['TranscriptionJob']['TranscriptionJobStatus'] \\\n",
    "                in ['COMPLETED', 'FAILED']:\n",
    "            break\n",
    "        exp_backoff_sleep(start_seconds)\n",
    "    app_log(\"Transcription completed\")\n",
    "\n",
    "    # Get the output file from S3\n",
    "    response = s3.get_object(Bucket=bucket, Key=output_key)\n",
    "\n",
    "    # Read and extract the transcript text\n",
    "    transcript_text = json.load(\n",
    "        response['Body'])['results']['transcripts'][0]['transcript']\n",
    "\n",
    "    # Infer transcript local path and write transcript to file\n",
    "    transcript_id = output_key.split('/')[-1].rsplit('.', 1)[0]\n",
    "    if not os.path.exists('output'): os.makedirs('output')\n",
    "    transcript_path = f\"output/{transcript_id}-Transcript-Raw.txt\"\n",
    "    with open(transcript_path, 'w') as file:\n",
    "        file.write(transcript_text)\n",
    "\n",
    "    app_log(f\"Transcript saved locally at {transcript_path}\")\n",
    "\n",
    "    # Calculate elapsed time\n",
    "    elapsed_seconds = round(time.time() - start_seconds, 2)\n",
    "\n",
    "    # Return locals\n",
    "    return {\n",
    "        'elapsed_seconds': elapsed_seconds,\n",
    "        's3_audio_path': s3_audio_path,\n",
    "        's3_output_path': s3_output_path,\n",
    "        'transcript_id': transcript_id,\n",
    "        'transcript_path': transcript_path,\n",
    "        'transcript_chars': len(transcript_text)\n",
    "    }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "887dd86b-1fa6-4050-8445-a768882991fb",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-07-09T02:31:05.055479Z",
     "iopub.status.busy": "2024-07-09T02:31:05.054819Z",
     "iopub.status.idle": "2024-07-09T02:31:14.193048Z",
     "shell.execute_reply": "2024-07-09T02:31:14.192247Z",
     "shell.execute_reply.started": "2024-07-09T02:31:05.055452Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[2024-07-08 21:31:05.48] Transcription started \n",
      "[2024-07-08 21:31:09.57] Waiting ... Elapsed time: 0.09 seconds \n",
      "[2024-07-08 21:31:10.70] Waiting ... Elapsed time: 4.17 seconds \n",
      "[2024-07-08 21:31:12.12] Waiting ... Elapsed time: 5.31 seconds \n",
      "[2024-07-08 21:31:13.90] Waiting ... Elapsed time: 6.74 seconds \n",
      "[2024-07-08 21:31:13.99] Transcription completed \n",
      "[2024-07-08 21:31:14.17] Transcript saved locally at output/demo1-2024-07-08-21-31-05-188132-Transcript-Raw.txt \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'elapsed_seconds': 8.7,\n",
       " 's3_audio_path': 's3://asimj-test/x-transcribe-demo/demo1.mp3',\n",
       " 's3_output_path': 's3://asimj-test/x-transcribe-demo/demo1-2024-07-08-21-31-05-188132.json',\n",
       " 'transcript_id': 'demo1-2024-07-08-21-31-05-188132',\n",
       " 'transcript_path': 'output/demo1-2024-07-08-21-31-05-188132-Transcript-Raw.txt',\n",
       " 'transcript_chars': 105}"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "demo1_mp3 = \"s3://asimj-test/x-transcribe-demo/demo1.mp3\"\n",
    "convert_s3_audio_to_transcript(demo1_mp3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "064b24c1-165f-412f-806b-60eac6df85a8",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.11.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}