Btukfyl

I'm very new to OOP, and want to start getting in the habit of writing modular, reusable code. Sorry for the wall of code and text, just trying to give as much context and be as clear as possible.

I have defined a series of functions, all which act on a pandas DataFrame, and I wish to create a class from this series of functions. I thought it would be a good idea to create a class because I will be using this code on DataFrames that are pretty standard going forward.

My question is twofold:
1. Should I be using OOP concepts and design for this problem
2. What the hell is causing my darn code to break?!

Thanks in advance everyone!

Below are the functions I have defined:

def __describe(df, col):

    df_desc = df[col].describe()

    return df_desc



def __sort_by_character_study_day(df):        

    new_index_values = df.index.levels[0].str.split().str[-1].astype(int)

    df.index = df.index.set_levels(new_index_values, level='study_day')

    df = df.sort_index()

    return df



def change_from_baseline(df, col):

    df_sorted = sort_df(df = df, by = ["Animal_id", "ord"])

    groupedf = grouper(df=df_sorted, by='Animal_id')

    df_sorted['Change From Baseline'] = df_sorted[col] - groupedf[col].transform('first')

    df_sorted['Percent Change From Baseline'] = (df_sorted['Change From Baseline'] / groupedf[col].transform('first')).round(4) * 100

    return df_sorted



def grouper(df, by):

    df = df.groupby(by)

    return df



def sns_box_plot(df, col, y_lab):   

    plt.subplots(figsize=(12,8))

    sns.set_style("whitegrid")

    g = sns.boxplot(x="study_day", y = col, hue="group_c", 

                    data = df, palette='rainbow', 

                    hue_order=["Group 1", "Group 2", "Group 3"])

    plt.ylabel(y_lab, fontsize=20)

    plt.xlabel("Study Day", fontsize=20)

    plt.tick_params('both', labelsize='14')

    plt.show()



def sns_bar_plot(df, col, y_lab):       

    plt.subplots(figsize=(12,8))

    sns.set_style("whitegrid")

    g = sns.barplot(x="study_day", y = col, hue="group_c", 

                    data = df, palette='rainbow')

    plt.ylabel(y_lab, fontsize=20)

    plt.xlabel("Study Day", fontsize=20)

    plt.tick_params('both', labelsize='14')

    plt.show()



def sort_df(df, by):

    df = df.sort_values(by)

    return df   



def summary_stats(df, by, col):

    dfgrouped = grouper(df, by)

    df_desc = __describe(dfgrouped, col)

    df_summ = pd.DataFrame()

    df_summ["count"] = df_desc['count']

    df_summ["mean"] = df_desc['mean'].round(2).astype(str)

    df_summ["std"] = df_desc['std'].round(2).astype(str)

    df_summ["25%"] = df_desc['25%'].round(2).astype(str)

    df_summ["50%"] = df_desc['50%'].round(2).astype(str)

    df_summ["75%"] = df_desc['75%'].round(2).astype(str)

    df_summ["min"] = df_desc['min'].round(2).astype(str)

    df_summ["max"] = df_desc['max'].round(2).astype(str)

    return __sort_by_character_study_day(df_summ)



def summary_stat_formatted(df, by, col):

    data_summ = summary_stats(df, by, col)

    formatted_summ = pd.DataFrame()

    formatted_summ["Number of Observations"] = data_summ['count'].astype(int)

    formatted_summ["Mean (SD)"] = data_summ["mean"] +' (' + data_summ["std"] +')'

    formatted_summ["Median (25th - 75th %ile)"] = data_summ["50%"] + ' (' + data_summ["25%"] +' ,' + data_summ["75%"] +')'  

    formatted_summ["Min, Max"] = data_summ["min"] +' ,' + data_summ["max"]

    return formatted_summ



def write_to_excel(df, outfile, sheetname):

    writer = pd.ExcelWriter(outfile, engine='xlsxwriter')

    df.to_excel(writer, sheet_name=sheetname)

When I call one of the functions, say:

summary_stats(df=bw, by= ["study_day", "group_c"], col='Body_Weight')

Everything works as expected, I get the following output:

Now, since I will be using these same functions on nearly identical DataFrames (only thing different is the name of the dataframe and the analysis column), I thought it might be a good idea to create a class (please let me know if this is the right approach). My class looks like this:

class PreClinicalData(object):



    def __init__(self, df):

        self.df = df



    def __describe(self, col):

        df_desc = self.df[col].describe()

        return df_desc



    def __sort_by_character_study_day(self):

        new_index_values = self.df.index.levels[0].str.split().str[-1].astype(int)

        self.df.index = self.df.index.set_levels(new_index_values, level='study_day')

        self.df = self.df.sort_index()

        return self.df



    def change_from_baseline(self,  col):

        df_sorted = self.sort_df(df = self.df, by = ["Animal_id", "ord"])

        groupedf = self.grouper(df=df_sorted, by='Animal_id')

        df_sorted['Change From Baseline'] = df_sorted[col] - groupedf[col].transform('first')

        df_sorted['Percent Change From Baseline'] = (df_sorted['Change From Baseline'] / groupedf[col].transform('first')).round(4) * 100

        return df_sorted



    def summary_stats(self, by, col):

        dfgrouped = self.grouper(df=self.df, by=by)

        df_desc = self.__describe(dfgrouped, col=col)

        df_summ = pd.DataFrame()

        df_summ["count"] = df_desc['count']

        df_summ["mean"] = df_desc['mean'].round(2).astype(str)

        df_summ["std"] = df_desc['std'].round(2).astype(str)

        df_summ["25%"] = df_desc['25%'].round(2).astype(str)

        df_summ["50%"] = df_desc['50%'].round(2).astype(str)

        df_summ["75%"] = df_desc['75%'].round(2).astype(str)

        df_summ["min"] = df_desc['min'].round(2).astype(str)

        df_summ["max"] = df_desc['max'].round(2).astype(str)

        return self.__sort_by_character_study_day(df_summ)



    def summary_stat_formatted(self, by, col):

        data_summ = self.summary_stats(self, by, col)

        formatted_summ = pd.DataFrame()

        formatted_summ["Number of Observations"] = data_summ['count'].astype(int)

        formatted_summ["Mean (SD)"] = data_summ["mean"] +' (' + data_summ["std"] +')'

        formatted_summ["Median (25th - 75th %ile)"] = data_summ["50%"] + ' (' + data_summ["25%"] +' ,' + data_summ["75%"] +')'  

        formatted_summ["Min, Max"] = data_summ["min"] +' ,' + data_summ["max"]

        return formatted_summ



    def write_to_excel(self, outfile, sheetname):

        writer = pd.ExcelWriter(outfile, engine='xlsxwriter')

        self.df.to_excel(writer, sheet_name=sheetname)    



    def grouper(self, by):

        return self.df.groupby(by)



    def sns_box_plot(self, col, y_lab):   

        plt.subplots(figsize=(12,8))

        sns.set_style("whitegrid")

        g = sns.boxplot(x="study_day", y = col, hue="group_c", 

                        data = self.df, palette='rainbow', 

                        hue_order=["Group 1", "Group 2", "Group 3"])

        plt.ylabel(y_lab, fontsize=20)

        plt.xlabel("Study Day", fontsize=20)

        plt.tick_params('both', labelsize='14')

        plt.show()



    def sns_bar_plot(self, col, y_lab):       

        plt.subplots(figsize=(12,8))

        sns.set_style("whitegrid")

        g = sns.barplot(x="study_day", y = col, hue="group_c", 

                        data = self.df, palette='rainbow')

        plt.ylabel(y_lab, fontsize=20)

        plt.xlabel("Study Day", fontsize=20)

        plt.tick_params('both', labelsize='14')

        plt.show()



    def sort_df(self, by):

        return self.df.sort_values(by)   



    def summary_stats(self, by, col):

        dfgrouped = self.grouper(self, by)

        df_desc = self.__describe(dfgrouped, col)

        df_summ = pd.DataFrame()

        df_summ["count"] = df_desc['count']

        df_summ["mean"] = df_desc['mean'].round(2).astype(str)

        df_summ["std"] = df_desc['std'].round(2).astype(str)

        df_summ["25%"] = df_desc['25%'].round(2).astype(str)

        df_summ["50%"] = df_desc['50%'].round(2).astype(str)

        df_summ["75%"] = df_desc['75%'].round(2).astype(str)

        df_summ["min"] = df_desc['min'].round(2).astype(str)

        df_summ["max"] = df_desc['max'].round(2).astype(str)

        return self.__sort_by_character_study_day(df_summ)



    def summary_stat_formatted(self, by, col):

        data_summ = self.summary_stats(self, by, col)

        formatted_summ = pd.DataFrame()

        formatted_summ["Number of Observations"] = data_summ['count'].astype(int)

        formatted_summ["Mean (SD)"] = data_summ["mean"] +' (' + data_summ["std"] +')'

        formatted_summ["Median (25th - 75th %ile)"] = data_summ["50%"] + ' (' + data_summ["25%"] +' ,' + data_summ["75%"] +')'  

        formatted_summ["Min, Max"] = data_summ["min"] +' ,' + data_summ["max"]

        return formatted_summ



    def write_to_excel(self, outfile, sheetname):

        writer = pd.ExcelWriter(outfile, engine='xlsxwriter')

        self.df.to_excel(writer, sheet_name=sheetname)

I can instantiate an instance of the class as follows:

bodyweight = PreClinicalData(chickv_data['bodyweight'])

And I am able to call the DataFrame and use the head() method just fine.:

bodyweight.df.head()

Now, when I call the same method as shown above...I get the following:

bodyweight.summary_stats(by = ["study_day", "group_c"], col = "Body_Weight")

which yields the following error:

I have no clue where these alleged three arguments are coming from?!