import argparse
import datetime
import pickle
import math
import logging
import numpy as np
import pandas as pd
import dateutil
from scipy.special import erf
from epa.models import EpaStationTextarRawData, EpaStationTextarAbnormal
from epa.logging import get_logger
logger = get_logger(__name__)
def extract_mark(df):
df = df.fillna("NoValue")
s = df.stack(df.columns.names)
regex = "(-?\d+\.?\d*|NR)?([\*#x]|NoValue)?"
df_value_mark = s.astype(str).str.extract(regex, expand=False)
df_value = df_value_mark.iloc[:, 0].unstack(["station", "item"])
df_mark = df_value_mark.iloc[:, 1].unstack(["station", "item"])
return df_value, df_mark
def extract_features(df, main_feature="PM10"):
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
import time
import importlib
from inspect import isclass
from functools import wraps
from epa import logging
logger = logging.get_logger(__name__)
class ClassNotFoundError(Exception):
"""Class not found in module"""
def __init__(self, module_name, class_name):
self.module_name = module_name
self.class_name = class_name
def get_class_with_name(module_name, class_name):
module = importlib.import_module(module_name)
attrs = [a for a in dir(module) if not a.startswith("__")]
for attr in attrs:
x = getattr(module, attr)
if isclass(x) and x.__name__ == class_name:
return x
raise ClassNotFoundError(module_name, class_name)
def timer(f):