gsva_analysis

import os
import pandas as pd
import numpy as np
from pandas import DataFrame, Series
from pandas_datareader import data, wb
import re

#standard imports as well as "re" for regex and "os" to maniuplate the directory location.

def period_remover(dataframe): 
    df=dataframe
    df.loc[:,"target_id"]=df.loc[:, "target_id"].replace(regex=True, to_replace=r"[.]\d*", value="")
    #If you didn't use the [.] it would replace everything because "." is a catch-all.
    return df

def expression_frame(patient_ids_from_tpm_dataframe,tpm_counts_df):
    iplist=0
    ptid=patient_ids_from_tpm_dataframe
    tpmdf=tpm_counts_df
    df=pd.DataFrame(index=hgncid)
    while iplist<=len(ptid)-1:
        t=tpmdf.loc[tpmdf["sample"] == ptid[iplist]].sort_values(by=["tpm","ext_gene","sample"], ascending=False, kind="mergesort").drop_duplicates("ext_gene")
        t=t.pivot(columns="sample", index="ext_gene", values="tpm")
        df=df.join(t)
        iplist+=1
    return df

#By sorting by tpm values , all the highest values are at the "top" of the frame
#Then organize by ext_gene (which is the hugo-ID) and then the sample name (which is the patient samples)
#When removing duplicates it takes only the first unique entity it sees, thereby reducing a gene per sample to a single tpm value.
#"mergesort" is the only stable sorting algorithm at this time.
#The use of a pivot table essentially creates the expression table.
#The .join function adds to the dataframe a column at a time.