Services Layer¶

The following API reference is generated directly from the project's Python docstrings.

Data loading¶

Data loading and preprocessing service for the Mangetamain application.

This module centralizes all CSV file reading operations from the data/ folder. All loading functions use standardized paths and consistent parameters.

filter_recipes ¶

filter_recipes(
    recipes_df,
    prep_range=(0, 180),
    ingredients_range=(1, 30),
    calories_range=(0, 1000),
    vegetarian_only=False,
    nutrition_grades=None,
)

Filters recipes according to specified criteria.

Parameters:

Name	Type	Description	Default
`recipes_df`	`DataFrame`	Recipes DataFrame	required
`prep_range`	`tuple`	Tuple (min, max) for preparation time	`(0, 180)`
`ingredients_range`	`tuple`	Tuple (min, max) for number of ingredients	`(1, 30)`
`calories_range`	`tuple`	Tuple (min, max) for calories	`(0, 1000)`
`vegetarian_only`	`bool`	Filter only vegetarian recipes	`False`
`nutrition_grades`	`list`	List of accepted nutritional grades (A, B, C, D, E)	`None`

Returns:

Type	Description
`DataFrame`	Filtered DataFrame

get_recipe_details ¶

get_recipe_details(recipe_id, recipes_df)

Retrieves all details of a recipe.

Parameters:

Name	Type	Description	Default
`recipe_id`	`int`	Recipe ID	required
`recipes_df`	`DataFrame`	Recipes DataFrame	required

Returns:

Type	Description
`dict`	Dictionary with recipe details

get_recipe_name ¶

get_recipe_name(recipe_id, recipes_df)

Retrieves the readable name of a recipe from its tokens.

Parameters:

Name	Type	Description	Default
`recipe_id`	`int`	Recipe ID	required
`recipes_df`	`DataFrame`	Recipes DataFrame	required

Returns:

Type	Description
`str`	Recipe name (approximate based on tokens)

get_recipe_stats ¶

get_recipe_stats(recipes_df)

Calculates global statistics on recipes.

Parameters:

Name	Type	Description	Default
`recipes_df`	`DataFrame`	Recipes DataFrame	required

Returns:

Type	Description
`dict`	Dictionary with statistics

load_recipes ¶

load_recipes(data_dir=None)

Loads and preprocesses recipe data from preprocessed_recipes.csv

The file now includes popularity statistics (review_count, average_rating, popularity_score) calculated from RAW_interactions.csv.

Returns:

Type	Description
`DataFrame`	DataFrame with preprocessed columns including popularity statistics

read_csv_file ¶

read_csv_file(
    filename,
    data_dir=None,
    usecols=None,
    dtype=None,
    nrows=None,
    **kwargs,
)

Centralized function to read any CSV file from the data/ folder. In production mode (ENV=prod), reads from Google Drive. In development mode, reads from local data/ directory.

This function standardizes CSV reading with error handling and integrated Streamlit caching.

Parameters:

Name	Type	Description	Default
`filename`	`str`	Name of the CSV file (e.g., "preprocessed_recipes.csv")	required
`data_dir`	`Optional[str]`	Directory containing the data (default: "data/")	`None`
`usecols`	`Optional[List[str]]`	List of columns to load (default: all)	`None`
`dtype`	`Optional[dict]`	Dictionary of column types	`None`
`nrows`	`Optional[int]`	Maximum number of rows to read	`None`
`**kwargs`		Additional arguments for pd.read_csv()	`{}`

Returns:

Type	Description
`DataFrame`	pd.DataFrame: DataFrame containing the data

Raises:

Type	Description
`FileNotFoundError`	If the file does not exist

read_pickle_file ¶

read_pickle_file(filename, data_dir=None)

Centralized function to read any pickle file from the data/ folder. In production mode (ENV=prod), reads from Google Drive. In development mode, reads from local data/ directory.

This function standardizes pickle reading with error handling and integrated Streamlit caching.

Parameters:

Name	Type	Description	Default
`filename`	`str`	Name of the pickle file (e.g., "similarity_matrix.pkl")	required
`data_dir`	`Optional[str]`	Directory containing the data (default: "data/")	`None`

Returns:

Type	Description
`Dict[str, Any]`	Dict[str, Any]: Dictionary containing the pickled data

Raises:

Type	Description
`FileNotFoundError`	If the file does not exist

Recommendation engine¶

Système de recommandations basé sur une matrice de similarité pré-calculée.

RecipeRecommender ¶

RecipeRecommender(recipes_df)

Système de recommandations de recettes utilisant une matrice de similarité pré-calculée.

Initialise le système de recommandations.

Parameters:

Name	Type	Description	Default
`recipes_df`	`DataFrame`	DataFrame contenant les recettes	required

get_similar_recipes ¶

get_similar_recipes(recipe_id, k=10)

Trouve les k recettes les plus similaires à une recette donnée. Utilise uniquement la matrice de similarité pré-calculée.

Parameters:

Name	Type	Description	Default
`recipe_id`	`int`	ID de la recette de référence	required
`k`	`int`	Nombre de recommandations à retourner	`10`

Returns:

Type	Description
`List[Tuple[Series, float]]`	Liste de tuples (recette, score de similarité)

recommend_by_filters ¶

recommend_by_filters(
    prep_range,
    ingredients_range,
    calories_range,
    vegetarian_only=False,
    k=12,
)

Recommande des recettes basées sur des filtres.

Parameters:

Name	Type	Description	Default
`prep_range`	`Tuple[int, int]`	Tuple (min, max) pour le temps de préparation	required
`ingredients_range`	`Tuple[int, int]`	Tuple (min, max) pour le nombre d'ingrédients	required
`calories_range`	`Tuple[int, int]`	Tuple (min, max) pour les calories	required
`vegetarian_only`	`bool`	Filtrer uniquement les recettes végétariennes	`False`
`k`	`int`	Nombre de recommandations	`12`

Returns:

Type	Description
`List[Series]`	Liste de recettes recommandées

format_recommendations_for_display ¶

format_recommendations_for_display(recommendations)

Formate les recommandations pour l'affichage.

Parameters:

Name	Type	Description	Default
`recommendations`	`List[Tuple[Series, float]]`	Liste de tuples (recette, score)	required

Returns:

Type	Description
`List[dict]`	Liste de dictionnaires avec les informations formatées

get_recommender ¶

get_recommender(recipes_df)

Crée et met en cache le système de recommandations.

Parameters:

Name	Type	Description	Default
`recipes_df`	`DataFrame`	DataFrame des recettes	required

Returns:

Type	Description
`RecipeRecommender`	Instance de RecipeRecommender

Search utilities¶

Moteur de recherche optimisé pour les recettes. Utilise la recherche textuelle et le ranking par pertinence.

format_description ¶

format_description(recipe, max_length=180)

Génère une description courte pour une recette.

Parameters:

Name	Type	Description	Default
`recipe`	`Series`	Série pandas de la recette	required
`max_length`	`int`	Longueur maximale de la description	`180`

Returns:

Type	Description
`str`	Description formatée

format_recipe_title ¶

format_recipe_title(recipe)

Génère un titre lisible pour une recette.

Parameters:

Name	Type	Description	Default
`recipe`	`Series`	Série pandas de la recette	required

Returns:

Type	Description
`str`	Titre formaté

get_recipe_by_id ¶

get_recipe_by_id(recipes_df, recipe_id)

Récupère une recette par son ID.

Parameters:

Name	Type	Description	Default
`recipes_df`	`DataFrame`	DataFrame des recettes	required
`recipe_id`	`int`	ID de la recette	required

Returns:

Type	Description
`Optional[Series]`	Série pandas de la recette ou None

get_trending_recipes ¶

get_trending_recipes(recipes_df, limit=20)

Récupère les recettes "tendances" (fallback quand pas de recherche). Tri par health score ou random.

Parameters:

Name	Type	Description	Default
`recipes_df`	`DataFrame`	DataFrame des recettes	required
`limit`	`int`	Nombre de recettes à retourner	`20`

Returns:

Type	Description
`DataFrame`	DataFrame des recettes tendances

search_recipes ¶

search_recipes(
    recipes_df,
    query="",
    prep_time_max=180,
    ingredients_max=30,
    calories_max=1000,
    vegetarian_only=False,
    nutrition_grades=None,
    sort_by="relevance",
    page=1,
    page_size=20,
)

Recherche et filtre les recettes avec pagination.

Parameters:

Name	Type	Description	Default
`recipes_df`	`DataFrame`	DataFrame complet des recettes	required
`query`	`str`	Texte de recherche	`''`
`prep_time_max`	`int`	Temps de préparation maximum	`180`
`ingredients_max`	`int`	Nombre d'ingrédients maximum	`30`
`calories_max`	`int`	Calories maximum	`1000`
`vegetarian_only`	`bool`	Filtrer les recettes végétariennes uniquement	`False`
`nutrition_grades`	`List[str]`	Liste des grades nutritionnels acceptés	`None`
`sort_by`	`str`	Tri (relevance, health_score, prep_time)	`'relevance'`
`page`	`int`	Numéro de page	`1`
`page_size`	`int`	Nombre de résultats par page	`20`

Returns:

Type	Description
`Tuple[DataFrame, int]`	Tuple (DataFrame paginé, nombre total de résultats)