utils.py

from collections import namedtuple
import datetime  as dt
import pandas    as pd
import numpy     as np

long_nan = 9223372036854775807

def is_long_nan(v):
    if v == long_nan:
        return True
    else :
        return False
    
def to_nan(x):
    if is_long_nan(x):    
        return np.nan
    else:
        return x
        
def _to_date(row):
    date = int(row['DATE'])
    return pd.datetime( year=date/10000, month = date/ 100 % 100, day = date%100)

def _to_datetime(row):
    date = int(row['DATE'])
    time = int(row['TIME']) / 1000
    return pd.datetime( year=date/10000, month=date/ 100 % 100, day = date%100,
                        hour=time/10000, minute=time/100%100, second =time%100)

def _to_dataframe(cloumset, index_func = None, index_column = None):
    df = pd.DataFrame(cloumset)
    for col in df.columns:
        if df.dtypes.loc[col] == np.int64:
            df.loc[:,col] = df.loc[:,col].apply(to_nan)            
    if index_func:
        df.index = df.apply(index_func, axis = 1)
    elif index_column:
        df.index = df[index_column]
        del df.index.name

    return df

def _error_to_str(error):
    if error:
        if 'message' in error:
            return str(error['error']) + "," + error['message']
        else:
            return str(error['error']) + ","
    else:
        return ","


def to_obj(class_name, data):
    try:
        if type(data) == list or type(data) == tuple:
            result = []
            for d in data:
                result.append( namedtuple(class_name, d.keys())(*d.values()) )
            return result

        elif type(data) == dict :
            result = namedtuple(class_name, data.keys())(*data.values())
            return result
        else:
            return data
    except Exception as e:
        print(class_name, data, e)
        return data

def to_date_int(date):
    if isinstance(date, (str, unicode)):
        t = dt.datetime.strptime(date, "%Y-%m-%d")
        date_int = t.year * 10000 + t.month * 100 + t.day
        return date_int
    elif isinstance(date, (int, np.integer)):
        return date
    else:
        return -1

def to_time_int(time):
    if isinstance(time, (str, unicode)):
        t = dt.datetime.strptime(time, "%H:%M:%S")
        time_int = t.hour * 10000 + t.minute * 100 + t.second
        return time_int
    elif isinstance(time, (int, np.integer)):
        return time
    else:
        return -1 

def extract_result(cr, data_format="", index_column=None, class_name=""):
    """
        format supports pandas, obj.
    """
    
    err = _error_to_str(cr['error']) if 'error' in cr else None
    if 'result' in cr:
        if data_format == "pandas":
            if index_column :
                return (_to_dataframe(cr['result'], None, index_column), err)
            # if 'TIME' in cr['result']:
            #     return (_to_dataframe(cr['result'], _to_datetime), err)
            # elif 'DATE' in cr['result']:
            #     return (_to_dataframe(cr['result'], _to_date), err)
            else:
                return (_to_dataframe(cr['result']), err)

        elif data_format == "obj" and cr['result'] and class_name:
            r = cr['result']
            if type(r) == list or type(r) == tuple:
                result = []
                for d in r:
                    result.append( namedtuple(class_name, d.keys())(*d.values()) )
            elif type(r) == dict :
                result = namedtuple(class_name, r.keys())(*r.values())
            else:
                result = r

            return (result, err)
        else:            
            return (cr['result'], err)
    else:
        return (None, err)