foodscope

A closer look at food, based on data from foodb

Download the database (also available here)

curl https://foodb.ca/public/system/downloads/foodb_2020_4_7_csv.tar.gz -O
tar xf foodb_2020_4_7_csv.tar.gz

Usage (see demo)

import foodscope as fs


# global view
db = fs.FooDb()
db.food.select("milk")

# case insensitive match across permutations
db.compound.equiv(fs.expand_greeks("beta-carotene"))

# base data frames
food = fs.Food()
compound = fs.Compound()

# unfiltered
nutrient = fs.Nutrient(cols="all")

#simple filters
food.select("chicken")
nutrient.select("protein")
fs.filter_content(source_type="Nutrient")

# fancy stuff
fs.foods_by_compound("sulf")
fs.composition("apple")
fs.health_effects("artichoke")

# compounds
c = fs.Compound()
c.equiv(fs.vitamin_a)
c.health_effects(fs.vitamin_a)

# misc
fs.expand_greeks("beta-Carotene")

https://foodb.ca/compounds/FDB003717#references

TODO

• Bin foods by nutrient (quartiles)
• get health effects of a given meal
• preprocess data to minimize file size, or import to postgres?
• fuzzy vs exact match - could use global option?
• ask foodb about nan cols
• use CompoundSynonym.csv
• convert to hdf5
• filter against config file with compounds of interest
• check compound list for capital greek letters
• cache results from complex functions
• document columns in return values
• add aliases for other vitamins

look up health effects of garlic and compare with this paper