timotta

from sklearn.preprocessing import OneHotEncoder
from sklearn.base import BaseEstimator, TransformerMixin


class OneHotEconderByColumn(TransformerMixin, BaseEstimator):
    def __init__(self, *, columns):
        self.columns = columns
        self.ohe = OneHotEncoder(drop="first", sparse=False)

    def fit(self, X, y=None):

timotta

def calculate_distance_one_point(city, position, cache, value=-1):
    row, col = position

    if (row, col) in cache:
        print( f"Already {row} {col}" )
        return None
    cache.add((row, col))

    print("doing", row, col)

timotta

# Needs EMBED_SIZE ans SENTENCE_SIZE
# df.text_array is a column with list os word in each cell

w2v_model = Word2Vec(
    sentences=df.text_array, 
    vector_size=EMBED_SIZE, 
    window=5, 
    min_count=1, 
    workers=4, 
    seed=1982,

timotta

# Needs gensim w2v_model trained
# EMBED_SIZE defined
# X should be a Matrix with examples on rows, and word indexes in sequence as columns

class LSTMClassification(nn.Module):
    def __init__(self, embed_vector, hidden_dim=100, dropout=0.5):
        super(LSTMClassification, self).__init__()
        self.n_layers = 1
        self.output_size = 1
        self.hidden_dim = hidden_dim

timotta

def color_the_best(x):
    styles = []
    bigger = max(x)
    for i in x:
        if i == bigger:
            styles.append('background-color: lightgreen')
        else:
            styles.append('')
    return styles
df.style.apply(color_the_best, axis=1)

timotta

timotta

timotta

timotta

timotta

rDCG^{k,n} =
\frac {
\sum_{i=1}^{n}\frac{rel_{i}}{log_{2}^{i+1}}
}{
\sum_{i=1}^{k}\frac{rel_{max}}{log_{2}^{i+1}}
}