Solutions to the takehome test by Elsa

Please note that solution to problem 2 requires the file "keys.py" to be modified with valid credentials in order for the main script to run correctly.

To do so, please open keys.py and provide said key and secret key to execute the script.

---

💠 Folder 1: P1 Solution

---

Contains:

▪️ haarcascade_frontalface_default.xml -> the cascade filter .xml

▪️ IMG_9721.jpg example image with multiple faces

▪️ Problem_1_Elsa.py -> solution script

▪️ results -> folder containg cropped headshots

---

💠 Folder 2: P2 Solution

---

Contains:

▪️ Problem_2_Elsa.py -> solution script

▪️ keys.py -> file imported by the main script with the keys to access s3.

---

Original description:

Data Engineer Takehome Test

Please clone this repo in your account, send us the link of the solution in your GitHub account together with your application.

Problem 1:

Build a Python script that detects faces in an image using OpenCV, and saves the headshots of the detected faces to a specified directory. The script should take as input a file path to an image, a directory path to save the headshots, and output the number of faces detected in the image. Instructions:

• Use OpenCV's Haar Cascade classifier for face detection
• The script should be written in Python and use the following libraries: OpenCV, Numpy, and PIL (Python Imaging Library)
• The script should be well commented and easy to understand
• The script should be able to handle a variety of image types (e.g. jpeg, png, etc.)
• The script should be able to handle images with multiple faces
• The script should save the headshots in the specified directory with the file name in the format "face_1.jpg", "face_2.jpg", etc.

Problem 2:

Move all image files from one S3 bucket to another S3 bucket, but only if the image has no transparent pixels.

Objective: Write a Python script that uses the Boto3 library to accomplish the following:

• List all the image files in a given S3 bucket
• Check if each image file has transparent pixels
• If an image file has no transparent pixels, copy it to a different S3 bucket
• If an image file has transparent pixels, log it in a separate file

Guidelines:

• Your script should take the name of the source and destination buckets as input
• You should use the Boto3 library to interact with S3
• You should use the Pillow library to check for transparent pixels in an image
• Your script should handle any errors that may occur during the opening of image file, copy process and anywhere else you deem necessary
• Your script should be well commented and easy to understand
• Your script should be executed from the command line