def extractRecipesFromJSON(allRecipes):
jsonFilePath = os.path.join(c.PATH_TO_RESOURCES, "aliasdata", "miniAliases.json")
myDict = util.loadJSONDict(jsonFilePath)
# Fill allRecipes with just the values for each JSON member,
# as the values actually contain the keys as a member
for _, val in myDict.items():
allRecipes.append(val)
def createAliasDatapoints(jsonDataFilePath, featureList):
# sparse matrix, shape = (n_samples, n_features)
allPoints = []
countToAliasName = {}
allAlias = util.loadJSONDict(jsonDataFilePath)
count = 0
for aliasName, aliasData in allAlias.items():
point = createAliasPoint(aliasData, featureList)
allPoints.append(point)
countToAliasName[count] = aliasName
count += 1
vec = DictVectorizer()
dataMatrix = vec.fit_transform(allPoints)
return dataMatrix, countToAliasName, allPoints
def createRecipeDatapoints(jsonDataFilePath, featureList):
# sparse matrix, shape = (n_samples, n_features)
allPoints = []
countToRecipeName = {}
allRecipes = util.loadJSONDict(jsonDataFilePath)
count = 0
for recipeName, recipe in allRecipes.items():
point = createRecipePoint(recipe, featureList)
allPoints.append(point)
countToRecipeName[count] = recipeName
count += 1
vec = DictVectorizer()
dataMatrix = vec.fit_transform(allPoints)
return dataMatrix, countToRecipeName, allPoints
def initializeTraits(verbose):
traitsDataPath = os.path.join(constants.PATH_TO_ROOT, "res", "csp_defaultTraits.json")
traits = util.loadJSONDict(traitsDataPath)
traits["verbose"] = verbose
traits["ND"] = nutrientdatabase.NutrientDatabase()
return traits
def main(argv):
global ingredientMassDict
global validAliasDict
global unitCountDict
validIngredientsFilePath = os.path.join(c.PATH_TO_RESOURCES, "validIngredients.json")
validAliasDict = util.loadJSONDict(validIngredientsFilePath)
conversionDict = util.createWaterConversionDict()
allRecipes = []
# Each alias has 3 main fields:
# "count"
# "aliasBuddies"
# "lines"
aliasData = {}
ingredientLineDict = {}
#ingredientMassDict = {}
#unitCountDict = {}
# Read in and parse recipe data structures (dictionaries) from a json file.
extractRecipesFromJSON(allRecipes)
# Convert all string data to lowercase.
lowerAllStrings(allRecipes)
#ndb = ndb.NutrientDatabase()
#Let's f**k around.
unmatched = float(0)
tried = float(0)
for recipe in allRecipes:
# print "Ingredient Lines: " + str(len(recipe['ingredientLines']))
# print recipe['ingredientLines']
# print "\nIngredients: " + str(len(recipe['ingredients']))
# print recipe['ingredients']
for ingredientLineIndex in range(0, len(recipe['ingredientLines'])):
if ingredientLineIndex == len(recipe['ingredients']):
break
ingredientLine = recipe['ingredientLines'][ingredientLineIndex].encode('ascii', errors='ignore')
ingredient = recipe['ingredients'][ingredientLineIndex].encode('ascii', errors='ignore')
if ingredient not in validAliasDict:
continue
if ingredient not in ingredientLineDict:
ingredientLineDict[ingredient] = []
ingredientLineDict[ingredient].append(ingredientLine)
#print ingredientLineDict
for ingredient in ingredientLineDict:
for ingredientLine in ingredientLineDict[ingredient]:
#TIME TO PARSE.
words = ingredientLine.split()
potentialStart = removeHyphen(words[0])
#If the first token is a number, try the next few.
if isPossibleAmount(words[0]):
if '/' in potentialStart:
tokens = potentialStart.split('/')
first = float(tokens[0])
second = float(tokens[1])
potentialStart = first/second
amount = float(potentialStart)
potentialUnit, foundUnit = extractUnit(words, ingredient, conversionDict)
if potentialUnit != None:
#Add both the mass and the unit count
if foundUnit:
massInGrams = amount*ndb.getConversionFactor(ingredient, potentialUnit)
else:
massInGrams = amount*conversionDict[potentialUnit]
if ingredient not in ingredientMassDict:
ingredientMassDict[ingredient] = []
ingredientMassDict[ingredient].append(massInGrams)
if ingredient not in unitCountDict:
unitCountDict[ingredient] = Counter()
unitCountDict[ingredient][potentialUnit] += 1
#print "Amount: " + str(amount) + " Unit: " + potentialUnit
else:
unmatched += 1
#print "Couldn't match unit for ingredient: " + ingredient
#print words
#print
elif not hasAnAmount(words):
if ingredient not in unitCountDict:
unitCountDict[ingredient] = Counter()
unitCountDict[ingredient]['unitless'] += 1
tried += 1
print "Missed amounts for " + str(unmatched) + " / " + str(tried) + " ingredients."
print str((tried-unmatched)/tried*100) + "% Success rate!"
# Get the counts of ingredient short names.
# Create a dictionary storing relationships between the various aliases.
# Create a dictionary with aliases as keys and lists of lines they've been
# associated with as values.
fillAliasData(allRecipes, aliasData)
#Temporarily removed to test.
dumpAliasDataToJSONFiles(aliasData)
#Now create small files
smallAliasData = {}
for _ in range(250):
item = aliasData.popitem()
smallAliasData[item[0]] = item[1]
smallFilePath = os.path.join(c.PATH_TO_RESOURCES, "aliasData_small.json")
util.dumpJSONDict(smallFilePath, smallAliasData)
for _ in range(250):
item = aliasData.popitem()
smallAliasData[item[0]] = item[1]
mediumFilePath = os.path.join(c.PATH_TO_RESOURCES, "aliasData_medium.json")
util.dumpJSONDict(mediumFilePath, smallAliasData)
for _ in range(500):
item = aliasData.popitem()
smallAliasData[item[0]] = item[1]
largeFilePath = os.path.join(c.PATH_TO_RESOURCES, "aliasData_large.json")
util.dumpJSONDict(largeFilePath, smallAliasData)
def createNutrientDataJSON(): nutrientIDsFilePath = os.path.join(c.PATH_TO_RESOURCES, "allNutrientIDs.json") nutrientIDs = util.loadJSONDict(nutrientIDsFilePath) buildNutritionalDatabase(nutrientIDs)