import pandas as pd
import numpy as np
import scipy.stats as stats
def read_csv(path, sep=None, feat_drop=None, date_key=None, dropna=False, drop_axis=None):
"""
Wraps the pandas.read_csv function adding extra features.
:path: string
Path to the csv file to be read.
:sep: char
Same argument as pandas.read_csv function.
:feat_drop: list of strings
Features to drop
:date_key: string
The name of the key representing the date in the dataset.
:returns: Pandas DataFrame
The pandas dataframe created from the data.
"""
df = pd.read_csv(path, sep=sep, index_col=date_key)
if dropna:
for d in drop_axis:
df = df.dropna(axis=d)
if feat_drop is not None:
df = df.drop(feat_drop, axis=1)
return df
def read_network(path, names=['mean', 'var', 'skew', 'kurt'], index=None):
"""
A specific reader
"""
data = np.loadtxt(path)
mean = data.mean(axis=1)
variance = data.var(axis=1)
skew = stats.skew(data, axis=1)
kustosis = stats.kurtosis(data, axis=1)
df = pd.DataFrame({names[0] : mean, names[1]: variance, names[2]: skew, names[3]: kustosis}, index=index)
return df
