def recommend(dish_id: int, recommendation_count: int = 5) -> Optional[dict]:
"""Recommends similar dishes.
Args:
dish_id (int): Dish id which koolsla recommends similar ones.
recommendation_count (int): Max recommendation count given by user.
Returns:
indice, similarity value (dictionary): Indice & similarity value or None.
"""
is_valid = data.validate_dish_id(dish_id)
if is_valid:
# Import the data
imported_data = data.import_data()
# Get the metadatasets
dataset = data.split_data(imported_data)
# Train the recommendation engine
tfidf_matrix = tfidf.train_engine(dataset['names'])
# Generate recommendations
recomended_dishes = tfidf.find_similarities(tfidf_matrix, dish_id, recommendation_count)
# Input dish
color_print.print_green('Given Dish')
color_print.print_dish(
name=dataset['names'][dish_id],
dish_id=dish_id,
color=color_print.green)
# Display the recommended dishes
color_print.print_yellow('Top ' + str(recommendation_count) + ' Recommendations')
for i, _ in recomended_dishes:
color_print.print_green(i)
color_print.print_dish(
name=dataset['names'][i],
dish_id=i,
color=color_print.yellow)
return recomended_dishes
return None
def test_train_engine(self):
dish_dataset = data.import_data(data.dish_dataset_path)
split_data = data.split_data(dish_dataset)
tfidf_matrix = tfidf.train_engine(split_data['names'])
self.assertIsNotNone(tfidf_matrix)
def test_find_similarities(self):
dish_dataset = data.import_data(data.dish_dataset_path)
split_data = data.split_data(dish_dataset)
tfidf_matrix = tfidf.train_engine(split_data['names'])
similar_dishes = tfidf.find_similarities(tfidf_matrix, 25, 5)
self.assertIsNotNone(similar_dishes)