machine_learning/predict_dimensions_from_annotations.py

import pandas as pd
from fuzzywuzzy import process

def predict_dimensions_from_annotations(annotations):
    # Load the data from CSV into a DataFrame
    data = pd.read_csv("/content/output.csv")

    # Initialize a list to store the predicted dimensions
    predicted_dimensions = []

    # Iterate over the annotations
    for annotation in annotations:
        # Extract object type name
        object_type_name = annotation['object_type_name']
        
        # Use fuzzy matching to find the closest matching product name in the CSV data
        closest_match = process.extractOne(object_type_name, data['Product Family'])
        
        # If a match is found above a certain threshold
        if closest_match[1] >= 50:
            # Get the index of the matched product family
            index = data[data['Product Family'] == closest_match[0]].index[0]
            
            # Retrieve the dimensions of the matched product
            length, height, width = data.loc[index, ['LENGTH (Inches)', 'HEIGHT (Inches)', 'WIDTH (Inches)']]
            
            # Append the predicted dimensions to the list
            predicted_dimensions.append((length, height, width))
        else:
            # If no match is found above the threshold, append None
            predicted_dimensions.append(None)

    return predicted_dimensions