def read_coocs():
"""
Read the file "Cococcs of ingredients"
"""
raw = dict(hp.dumpRead("Cococcs of ingredients"))
coocs = dict()
for key in raw:
coocs[(key[0][2:-1].lower(), key[1][2:-1].lower())] = raw[key]
return dict(sorted(coocs.items(), key=lambda item: item[1], reverse=True))
def generate_ingredients_h(il=[], init=random.choice(hp.dumpRead("model/inglist_final"))):
"""
Helper function for generate_ingredients()
"""
coocs = read_coocs()
selected = il
for key in coocs:
if len(il) < 20 and key[0] == init and key[0] not in selected:
selected.append(key[1])
return selected
def generate_ingredients():
"""
Generate a list of related ingredients
It is possible to comment out this code and add a list of ingredients manually.
For example:
selected = ["pasta", "tomato", "garlic", "oil", "onions", "salt", "pepper"]
return selected
This way you can use custom ingredients
"""
first = ""
selected = []
# Choose a random, valid first ingredient
while first == "" or len(first) > 30:
first = random.choice(hp.dumpRead("model/inglist_final"))
# Get a list of 20 ingredients that are related to the previous ones
# Not all 20 are used, this is just to make sure that the model does not run out of ingredients.
# 20 is deduced by:
# [maximum length of a sentence] x [maximum percentage of ingredients in a sentence]
# 50 x 0.40 = 20
while len(selected) < 20:
selected += generate_ingredients_h(init=random.choice(hp.dumpRead("model/inglist_final")))
return selected
def init_vars():
"""
Prepare the data for training
"""
data = hp.dumpRead(DATANAME)
words = sorted(list(set(data)))
ewords = dict((c, i) for i, c in enumerate(words))
enumbs = dict((i, c) for i, c in enumerate(words))
seq_length = 100
dataX = []
dataY = []
for i in range(0, len(data) - seq_length, 1):
seq_in = data[i:i + seq_length]
seq_out = data[i + seq_length]
dataX.append([ewords[char] for char in seq_in])
dataY.append(ewords[seq_out])
n_patterns = len(dataX)
X = np.reshape(dataX, (n_patterns, seq_length, 1))
X = X / float(len(words))
y = np_utils.to_categorical(dataY)
return (X, y, dataX, dataY, words, ewords, enumbs)
# -*- coding: utf-8 -*- import helpers as hp a = hp.dumpRead()