def main(data_dir):
ureg = UnitRegistry()
food_description_file = os.path.join(data_dir, "FOOD_DES.txt")
nutrient_definition_file = os.path.join(data_dir, "NUTR_DEF.txt")
nutrition_data_file = os.path.join(data_dir, "NUT_DATA.txt")
weight_data_file = os.path.join(data_dir, "WEIGHT.txt")
food_descriptions = pandas.read_csv(food_description_file, quotechar='~', delimiter='^', encoding='latin-1',
header=None, names=food_description_columns)
nutrient_definitions = pandas.read_csv(nutrient_definition_file, quotechar='~', delimiter='^', encoding='latin-1',
header=None, names=nutrient_definition_columns)
nutrition_data = pandas.read_csv(nutrition_data_file, quotechar='~', delimiter='^', encoding='latin-1', header=None,
names=nutrition_data_columns)
weight_data = pandas.read_csv(weight_data_file, quotechar='~', delimiter='^', encoding='latin-1', header=None,
names=weight_data_columns)
# Pandas is retarded when it comes to handling text in csv files...
food_descriptions.fillna('', inplace=True)
nutrient_definitions.fillna('', inplace=True)
nutrition_data.fillna('', inplace=True)
weight_data.fillna('', inplace=True)
with app.test_request_context():
ingredients = {}
ingredient_preparations = []
nutrients = {}
ingredient_nutrients = []
for entry in food_descriptions.itertuples():
if entry.food_group in ignored_food_groups:
continue
ingredient = models.Ingredient(ingredient_id=int(entry.ndb_id))
ingredients[entry.ndb_id] = ingredient
ingredient.names.append(models.IngredientName(name=entry.description, canonical=True))
for entry in nutrient_definitions.itertuples():
nutrient_id = int(entry.nutrient_id)
display_name = nutrient_display_names.get(nutrient_id)
scientific_name = nutrient_scientific_names.get(nutrient_id)
recommended_daily_intake = nutrient_rdi.get(nutrient_id)
display = display_nutrient.get(nutrient_id, False)
nutrient = models.Nutrient(nutrient_id=nutrient_id, display_name=display_name,
scientific_name=scientific_name, measurement_unit=entry.units,
recommended_daily_intake=recommended_daily_intake, display=display)
nutrients[nutrient_id] = nutrient
# Most entries in the weights file conform to this pattern
weight_re = re.compile(r"([\w\s]+?)(?:\s+\(.*\))?(?:,\s+(.*))?\Z")
for weight_entry in weight_data.itertuples():
if weight_entry.ndb_id not in ingredients:
continue
# Pint thinks fl oz is femtolitre ounces
if weight_entry.measure_description == "fl oz":
description = "fluid ounces"
# US regulation defines a fluid ounce as equivalent to 30mL for nutrition labeling purposes
volume = weight_entry.amount * ureg.parse_expression("30 ml").to_base_units()
# Convert the gram weight to kilograms so density is in standard units
mass = weight_entry.gram_weight / 1000 * ureg.kilogram
density = float((mass / volume).magnitude)
# Special case, as pat matches a unit, but in this context should not be interpreted as such
elif weight_entry.measure_description.startswith("pat "):
description = weight_entry.measure_description
density = None
else:
match = weight_re.match(weight_entry.measure_description)
if match:
(unit_name, preparation) = match.groups()
# First determine that this weight contains units rather than something nebulous like a "serving"
try:
quantity = weight_entry.amount * ureg.parse_expression(unit_name)
description = preparation
# Convert to base units so volume measurements are in cubic meters
volume = quantity.to_base_units()
# Discard entries with non-volume measurements
if not volume.units.get("meter") == 3:
continue
# Convert the gram weight to kilograms so density is in standard units
mass = weight_entry.gram_weight / 1000 * ureg.kilogram
density = float((mass / volume).magnitude)
except UndefinedUnitError:
description = weight_entry.measure_description
density = None
else:
description = weight_entry.measure_description
density = None
ingredient_preparation = models.IngredientMeasure(ingredient_id=int(weight_entry.ndb_id),
description=description, density=density,
amount=float(weight_entry.amount),
weight=float(weight_entry.gram_weight))
ingredient_preparations.append(ingredient_preparation)
for entry in nutrition_data.itertuples():
if entry.ndb_id not in ingredients:
continue
ingredient_nutrient = models.IngredientNutrient(nutrient_id=int(entry.nutrient_id),
ingredient_id=int(entry.ndb_id),
quantity=float(entry.nutrient_value))
ingredient_nutrients.append(ingredient_nutrient)
db.session.add_all(ingredients.values())
db.session.add_all(nutrients.values())
db.session.commit()
db.session.add_all(ingredient_preparations)
db.session.add_all(ingredient_nutrients)
db.session.commit()
import metarecipe.models as m
from metarecipe.app import app, db
if __name__ == "__main__":
with app.test_request_context():
db.metadata.drop_all()
db.metadata.create_all()
# m.Ingredient.query.order_by(db.desc(db.func.similarity("chicken", m.Ingredient.ingredient_name))).limit(10).all()
def main(data_dir):
ureg = UnitRegistry()
food_description_file = os.path.join(data_dir, "FOOD_DES.txt")
nutrient_definition_file = os.path.join(data_dir, "NUTR_DEF.txt")
nutrition_data_file = os.path.join(data_dir, "NUT_DATA.txt")
weight_data_file = os.path.join(data_dir, "WEIGHT.txt")
food_descriptions = pandas.read_csv(food_description_file,
quotechar='~',
delimiter='^',
encoding='latin-1',
header=None,
names=food_description_columns)
nutrient_definitions = pandas.read_csv(nutrient_definition_file,
quotechar='~',
delimiter='^',
encoding='latin-1',
header=None,
names=nutrient_definition_columns)
nutrition_data = pandas.read_csv(nutrition_data_file,
quotechar='~',
delimiter='^',
encoding='latin-1',
header=None,
names=nutrition_data_columns)
weight_data = pandas.read_csv(weight_data_file,
quotechar='~',
delimiter='^',
encoding='latin-1',
header=None,
names=weight_data_columns)
# Pandas is retarded when it comes to handling text in csv files...
food_descriptions.fillna('', inplace=True)
nutrient_definitions.fillna('', inplace=True)
nutrition_data.fillna('', inplace=True)
weight_data.fillna('', inplace=True)
with app.test_request_context():
ingredients = {}
ingredient_preparations = []
nutrients = {}
ingredient_nutrients = []
for entry in food_descriptions.itertuples():
if entry.food_group in ignored_food_groups:
continue
ingredient = models.Ingredient(ingredient_id=int(entry.ndb_id))
ingredients[entry.ndb_id] = ingredient
ingredient.names.append(
models.IngredientName(name=entry.description, canonical=True))
for entry in nutrient_definitions.itertuples():
nutrient_id = int(entry.nutrient_id)
display_name = nutrient_display_names.get(nutrient_id)
scientific_name = nutrient_scientific_names.get(nutrient_id)
recommended_daily_intake = nutrient_rdi.get(nutrient_id)
display = display_nutrient.get(nutrient_id, False)
nutrient = models.Nutrient(
nutrient_id=nutrient_id,
display_name=display_name,
scientific_name=scientific_name,
measurement_unit=entry.units,
recommended_daily_intake=recommended_daily_intake,
display=display)
nutrients[nutrient_id] = nutrient
# Most entries in the weights file conform to this pattern
weight_re = re.compile(r"([\w\s]+?)(?:\s+\(.*\))?(?:,\s+(.*))?\Z")
for weight_entry in weight_data.itertuples():
if weight_entry.ndb_id not in ingredients:
continue
# Pint thinks fl oz is femtolitre ounces
if weight_entry.measure_description == "fl oz":
description = "fluid ounces"
# US regulation defines a fluid ounce as equivalent to 30mL for nutrition labeling purposes
volume = weight_entry.amount * ureg.parse_expression(
"30 ml").to_base_units()
# Convert the gram weight to kilograms so density is in standard units
mass = weight_entry.gram_weight / 1000 * ureg.kilogram
density = float((mass / volume).magnitude)
# Special case, as pat matches a unit, but in this context should not be interpreted as such
elif weight_entry.measure_description.startswith("pat "):
description = weight_entry.measure_description
density = None
else:
match = weight_re.match(weight_entry.measure_description)
if match:
(unit_name, preparation) = match.groups()
# First determine that this weight contains units rather than something nebulous like a "serving"
try:
quantity = weight_entry.amount * ureg.parse_expression(
unit_name)
description = preparation
# Convert to base units so volume measurements are in cubic meters
volume = quantity.to_base_units()
# Discard entries with non-volume measurements
if not volume.units.get("meter") == 3:
continue
# Convert the gram weight to kilograms so density is in standard units
mass = weight_entry.gram_weight / 1000 * ureg.kilogram
density = float((mass / volume).magnitude)
except UndefinedUnitError:
description = weight_entry.measure_description
density = None
else:
description = weight_entry.measure_description
density = None
ingredient_preparation = models.IngredientMeasure(
ingredient_id=int(weight_entry.ndb_id),
description=description,
density=density,
amount=float(weight_entry.amount),
weight=float(weight_entry.gram_weight))
ingredient_preparations.append(ingredient_preparation)
for entry in nutrition_data.itertuples():
if entry.ndb_id not in ingredients:
continue
ingredient_nutrient = models.IngredientNutrient(
nutrient_id=int(entry.nutrient_id),
ingredient_id=int(entry.ndb_id),
quantity=float(entry.nutrient_value))
ingredient_nutrients.append(ingredient_nutrient)
db.session.add_all(ingredients.values())
db.session.add_all(nutrients.values())
db.session.commit()
db.session.add_all(ingredient_preparations)
db.session.add_all(ingredient_nutrients)
db.session.commit()