在Pandas中与类方法一起使用pipe（）？

Question

我创建了一个类（存储在dataanalysis.py）来存储我将用于数据分析的大多数方法。以下是存储在此文件中的代码：

import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt

class Trening:
    def __init__(self, Path):
        self.path = Path
        self.df = None

    def getDF(self):
        return self.df

    def loadData(self, fileName= "pilot1.csv"):
        filePath = str(self.path + fileName)
        df = pd.read_csv(filePath, skiprows=2, decimal=".")
        #return df
        self.df = df

    def dnfCountandReplace(self, Return=False):
        filt = self.df['FINISH'] == 'DNF'
        dnf = self.df[filt]

        dnf = dnf.replace('DNF', 1)
        dnf.to_csv('faenta.csv')
        self.df.replace('DNF', np.NaN, inplace=True)
        self.df.dropna(subset=['FINISH'], inplace=True)
        if Return:
            return dnf

    def changeDataType(self, Return=False):
        self.df["FINISH"] = self.df["FINISH"].str.replace(',', '.').astype(float)
        self.df["INTER 1"] = self.df["INTER 1"].str.replace(',', '.').astype(float)
        self.df["SECTION IM4-FINISH"] = self.df["SECTION IM4-FINISH"].str.replace(',', '.').astype(float)
        self.df["COMMENT"] = self.df['COMMENT'].astype(str)
        self.df["COMMENT"] = self.df['COMMENT'].str.replace('1', 'COURSE 1')
        self.df["COMMENT"] = self.df['COMMENT'].str.replace('2', 'COURSE 2')
        self.df["COMMENT"] = self.df['COMMENT'].str.replace('9', 'STRAIGHT-GLIDING')
        pd.to_numeric(self.df['FINISH'], downcast='float', errors='raise')
        pd.to_numeric(self.df['INTER 1'], downcast='float', errors='raise')
        pd.to_numeric(self.df['SECTION IM4-FINISH'], downcast='float', errors='raise')
        if Return:
            return self.df


    def renameCommentToCourse(self, Return=False):
        self.df.rename(columns={'COMMENT': 'COURSE'}, inplace=True)
        if Return:
            return self.df

    def groupData(self, Return=False):
        self.df.groupby(['BIB#', 'COURSE'])['FINISH']
        if Return:
            return self.df

    def findTwoFastestRunsbyGroup(self, Return=False):
        self.df['FINISH'].nsmallest(2)
        if Return:
            return self.df

    def parseToCSV(self):
        self.df.to_csv('cool.csv')

    def calculateSpeed(self, Return=False):
        # (x2 - x1) / (t2 - t1)
        x2 = 2
        x1 = 0
        t1 = 0
        for i in self.df['INTER 1']:
            self.df['ENTRANCESPEED'] = (x2 - x1) / (self.df['INTER 1'] - t1)
        if Return:
            return self.df

在我读博士期间，我可能会运行这些类方法>；300次，但有时以不同的顺序和组合，这取决于我试图获得的结果。所以我的问题是：有没有一种方法可以将类/链方法分组在一起，以便在需要时调用相关的方法？我知道Pandas有一个有趣的pipe（）函数，但我不知道如何使用它。在main.py文件中，我有以下代码：

from models import Trening
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt


pilot1 = Trening("/Users/cmagelssen/Desktop/DataAnalyse/data/pilot1/trening1/")
pilot1.loadData("PILOT1_SESSION2.csv")
pilot1.dnfCountandReplace(Return=True)
X = pilot1.changeDataType(Return=True)
pilot1.renameCommentToCourse(Return=True)
pilot1.groupData(Return=True)
pilot1.findTwoFastestRunsbyGroup(Return=True)
df = pilot1.calculateSpeed(Return=True)
plot = pilot1.getDF()

有人知道我如何使用管道（）将这些方法链接/组合在一起吗

最好的， 基督教徒

Answer 1

您可以尝试下面的代码。具有您喜欢的任何功能组合。确保如果最后使用的是GetDF，那么管道中调用的所有函数也会编辑self.df

pilot1 = Trening("/Users/cmagelssen/Desktop/DataAnalyse/data/pilot1/trening1/")
df = pilot1.loadData("PILOT1_SESSION2.csv")



df =df.pipe(pilot1.dnfCountandReplace, Return =True)\
   .pipe(pilot1.changeDataType, Return =True)\
   .pipe(pilot1.renameCommentToCourse, Return =True)\
   .pipe(pilot1.groupData, Return =True)\
   .pipe(pilot1.findTwoFastestRunsbyGroup, Return =True)\
   .pipe(pilot1.findTwoFastestRunsbyGroup, Return =True)\
   .pipe(pilot1.calculateSpeed, Return =True)\
   .pipe(pilot1.getDF)