DragonFruit - Software Engineering Challenge

Q1-DataStructureRepresentation.txt

For the microscope images, we can use a binary matrix where each pixel 
corresponds to either black or white.

For the dye sensor images, we can use a sparse representation. 
We’ll only store the coordinates (x, y) where the dye is detected, instead of storing the entire image.
We can use a list of tuples or a dictionary

Microscope image: 100,000 x 100,000 bits
Dye sensor image (sparse representation): Depends on the number of dye-detected pixels.

Q2-simulatedImages.py

import numpy as np

# random microscope image with blobs
microImage = np.random.choice([0, 1], size=(100000, 100000), p=[0.75, 0.25])

# random dye sensor image (sparse)
dyePixels = 1000  # Adjust as needed
dyeSensorImage = {}
for _ in range(dyePixels):
    x, y = np.random.randint(0, 100000), np.random.randint(0, 100000)
    dyeSensorImage[(x, y)] = True

print(list(dye_sensor_image.keys())[:10])

Q3-cancerDetection.py

def hasCancer(microImage, dyeSensorImage):
    parasiteArea = np.sum(microImage)  # Total area
    dyeDetectedArea = len(dyeSensorImage)  # Total dye detected area
    return dyeDetectedArea > 0.1 * parasiteArea

isCancerous = hasCancer(microImage, dyeSensorImage)
print(f"Parasite has cancer: {isCancerous}")

Q4-optimized.py

def has_cancer_optimized(microImage, dyeSensorImage):
    parasitePixels = set(zip(*np.where(microImage == 1)))  # Parasite pixels
    dyeDetectedPixels = set(dyeSensorImage.keys())  # Dye-lit pixels
    intersection = parasitePixels & dyeDetectedPixels
    return len(intersection) > 0.1 * len(parasitePixels)

isCancerousOpt = hasCancerOpt(microImage, dyeSensorImage)
print(f"Parasite has cancer (optimized): {isCancerousOpt}")

Q5-other.txt

For both types of images, we can consider lossless compression technique like Huffman coding.
Huffman coding assigns shorter codes to frequently occurring pixel values.

Q6-tools.txt

I used Python for code examples and explanations.
Used ChatGPT LLM for clarification and approaches