handler.py

from boto3 import client as boto3_client
import face_recognition
import urllib.parse
import os
import pickle
import boto3
from boto3.dynamodb.conditions import Key
import csv
import json
import io

def findInDynamoTable(table, key):
    print("Searching DynamoDB table")
    response = table.scan(FilterExpression=Key('name').eq(key))
    data = response['Items']
    for item in data:
        print(item)
    return data

def generate_encoding(filename):
    with open(filename, "rb") as file:
        data = pickle.load(file)
        print(data)
    return data

def pushToS3(data, key, bucket_name):
    print("Pushing data to S3")
    file_name = os.path.splitext(key)[0] + ".csv"
    header = ["year", "major", "name"]
    try:
        with open("/tmp/" + file_name, 'w', encoding='utf-8-sig', newline='') as f:
            writer = csv.DictWriter(f, fieldnames=header)
            writer.writeheader()
            writer.writerow({'year': data[0]['year'], 'major': data[0]['major'], 'name': data[0]['name']})
        with open("/tmp/" + file_name, "rb") as f:
            s3 = boto3.resource('s3')
            bucket = s3.Bucket(bucket_name)
            bucket.upload_fileobj(f, file_name)
    except Exception as e:
        print(e)
        print('Error in creating and uploading csv to bucket {}.'.format(bucket_name))
        raise e

def faceRecognitionFromVideo(video_path, data):
    print("Recognizing face from video")
    path = "/tmp/"

    os.system("ffmpeg -i " + os.path.join(path, video_path) + " -r 1 " + os.path.join(path, "image-%3d.jpeg"))
    all_files = os.listdir(path)
    image_files = [fname for fname in all_files if fname.endswith('.jpeg')]

    for image_file in image_files:
        found_image = face_recognition.load_image_file(os.path.join(path, image_file))
        image_encoding = face_recognition.face_encodings(found_image)

        if not image_encoding:
            continue

        image_encoding = image_encoding[0]

        for i, known_encoding in enumerate(data['encoding']):
            results = face_recognition.compare_faces([image_encoding], known_encoding)

            if results[0]:
                ans = data['name'][i]
                searched_data = findInDynamoTable(table, ans)
                pushToS3(searched_data, video_path, output_bucket)
                return

    return "no_face_found"

def face_recognition_handler(event, context):
    print("Handling face recognition event")
    print(event)
    bucket_name = event['Records'][0]['s3']['bucket']['name']
    key = urllib.parse.unquote_plus(event['Records'][0]['s3']['object']['key'], encoding='utf-8')
    print("Bucket name: {}, Key: {}".format(bucket_name, key))

    s3 = boto3.client('s3')
    s3.download_file(bucket_name, key, "/tmp/" + key)
    data = generate_encoding('encoding')

    return faceRecognitionFromVideo(key, data)

if __name__ == "__main__":
    input_bucket = ""
    output_bucket = ""
    aws_access_key_id = ""
    aws_secret_access_key = ""
    region_name = "us-east-1"
    s3_client = boto3.client('s3', aws_access_key_id= aws_access_key_id, aws_secret_access_key=aws_secret_access_key, region_name=region_name)
    s3 = boto3.resource(
    service_name='s3',
    region_name=region_name,
    aws_access_key_id=aws_access_key_id,
    aws_secret_access_key=aws_secret_access_key
    )
    TABLE_NAME = ""
    dynamodb = boto3.resource('dynamodb', region_name="us-east-2", aws_access_key_id=aws_access_key_id, aws_secret_access_key=aws_secret_access_key)
    table = dynamodb.Table(TABLE_NAME)