LSHTM_analysis/scripts/combining_FIXME.py

179 lines
6.5 KiB
Python
Executable file

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
'''
Created on Tue Aug 6 12:56:03 2019
@author: tanu
'''
# FIXME: change filename 2(mcsm normalised data)
# to be consistent like (pnca_complex_mcsm_norm.csv) : changed manually, but ensure this is done in the mcsm pipeline
#=======================================================================
# Task: combine 2 dfs on comm_valson cols by detecting them
# includes sainity checks
#=======================================================================
#%% load packages
import sys, os
import pandas as pd
import numpy as np
import re
#from varname import nameof
#%% end of variable assignment for input and output files
#=======================================================================
#%% function/methd to combine dfs
def detect_common_cols (df1, df2):
"""
Detect comm_valson cols
@param df1: df
@type df1: pandas df
@param df2: df
@type df2: pandas df
@return: comm_valson cols
@type: list
"""
common_cols = np.intersect1d(df1.columns, df2.columns).tolist()
print('Length of comm_cols:', len(common_cols)
, '\nMerging column/s:', common_cols
, '\n---------------------------------------------------------------'
, '\nType:', type(common_cols)
, '\n\ndtypes in merging columns:\n', df1[common_cols].dtypes
, '\n---------------------------------------------------------------')
return common_cols
#%% Function to combine 2 dfs by detecting commom cols and performing
# sanity checks on the output df
def combine_dfs_with_checks(df1, df2, my_join = 'outer'):
"""
Combine 2 dfs by finding merging columns automatically
@param df1: data frame
@type df1: pandas df
@param df2: data frame
@type df2: pandas df
@my_join: join type for merging
@type my_join: string
@return: combined_df
@type: pandas df
"""
print('Finding comm_cols and merging cols:'
,'\n=========================================================')
common_cols = np.intersect1d(df1.columns, df2.columns).tolist()
print('Length of comm_cols:', len(common_cols)
, '\nmerging column/s:', common_cols
, '\ntype:', type(common_cols))
#print('\ndtypes in merging columns:\n', df1[common_cols].dtypes)
print('selecting consistent dtypes for merging (object i.e string)')
#merging_cols = df1[comm_valson_cols].select_dtypes(include = [object]).columns.tolist()
#merging_cols = df1[comm_valson_cols].select_dtypes(include = ['int64']).columns.tolist()
merging_cols = common_cols.copy()
nmerging_cols = len(merging_cols)
print(' length of merging cols:', nmerging_cols
, '\nmerging cols:', merging_cols, 'type:', type(merging_cols)
, '\n=========================================================')
#========================
# merge 1 (combined_df)
# concatenating 2dfs:
# df1, df2
#========================
# checking cross-over of mutations in the two dfs to merge
ndiff_1 = df1[merging_cols].squeeze().isin(df2[merging_cols].squeeze()).sum()
ndiff1 = df1.shape[0] - ndiff_1
print('There are', ndiff1, 'unmatched mutations in left df')
#missing_mutinfo = df1[~left_df['mutationinformation'].isin(df2['mutationinformation'])]
#missing_mutinfo.to_csv('infoless_muts.csv')
ndiff_2 = df2[merging_cols].squeeze().isin(df1[merging_cols].squeeze()).sum()
ndiff2 = df2.shape[0] - ndiff_2
print('There are', ndiff2, 'unmatched mutations in right_df')
#comm_vals = np.intersect1d(df1[merging_cols], df2[merging_cols])
#comm_vals_count = len(comm_vals)
#print('length of comm_valson values:', comm_vals_count , '\ntype:', type(comm_vals_count))
#========================
# merging dfs & sanity checks
#========================
fail = False
print('combing with:', my_join)
comb_df = pd.merge(df1, df2, on = merging_cols, how = my_join)
expected_cols = df1.shape[1] + df2.shape[1] - nmerging_cols
if my_join == 'right':
df2_nd = df2.drop_duplicates(merging_cols, keep = 'first')
expected_rows = df2_nd.shape[0]
if my_join == 'left':
expected_rows = df1.shape[0]
#if my_join == 'inner':
# expected_rows = comm_vals_count
#if my_join == 'outer':
# df1_nd = df1.drop_duplicates(merging_cols, keep = 'first')
# df2_nd = df2.drop_duplicates(merging_cols, keep = 'first')
# expected_rows = df1_nd.shape[0] + df2_nd.shape[0] - comm_vals_count
if my_join == ('inner' or 'outer') and len(merging_cols) > 1:
#comm_vals = np.intersect1d(df1['mutationinformation'], df2['mutationinformation'])
print('length of merging_cols > 1, therefore omitting row checks')
combined_df = comb_df.copy()
expected_rows = len(combined_df)
else:
comm_vals = np.intersect1d(df1[merging_cols], df2[merging_cols])
print('length of merging_cols == 1, calculating expected rows in merged_df')
combined_df = comb_df.drop_duplicates(subset = merging_cols, keep ='first')
if my_join == 'inner':
expected_rows = len(comm_vals)
if my_join == 'outer':
df1_nd = df1.drop_duplicates(merging_cols, keep = 'first')
df2_nd = df2.drop_duplicates(merging_cols, keep = 'first')
expected_rows = df1_nd.shape[0] + df2_nd.shape[0] - len(comm_vals)
if len(combined_df) == expected_rows and len(combined_df.columns) == expected_cols:
print('PASS: successfully combined dfs with:', my_join, 'join')
else:
print('FAIL: combined_df\'s expected rows and cols not matched')
fail = True
print('\nExpected no. of rows:', expected_rows
, '\nGot:', len(combined_df)
, '\nExpected no. of cols:', expected_cols
, '\nGot:', len(combined_df.columns))
if fail:
sys.exit()
#if clean:
#foo = combined_df2.filter(regex = r'.*_x|_y', axis = 1)
#print(foo.columns)
#print('Detected duplicate cols with suffix: _x _y'
# , '\Dropping duplicate cols and cleaning')
# drop position col containing suffix '_y' and then rename col without suffix
combined_df_clean = combined_df.drop(combined_df.filter(regex = r'.*_y').columns, axis = 1)
combined_df_clean.rename(columns=lambda x: re.sub('_x$','', x), inplace = True)
return combined_df_clean
#%% end of function
#=======================================================================