mLogsFunctions/rmOutliers.py

import numpy as np
import pandas as pd
import streamlit as st 
import altair as alt
from streamlit_vega_lite import altair_component

from .fx import *
from mLogsFunctions import *

def rmOutliers(df):
    _o1, _o2 = st.columns([1,8])
    with _o1:
        st.session_state = selection_info(df,"method", "option_w", "option_x", "option_y", "option_c")

    #Crossplot and bar plot-----------------------------------------------------------------------
    with _o2:
        def rm_outliers(data):
            interval = interval_define()
            col21, col22 = st.columns(2)
            with col21:
                selected_points = altair_component(make_selection(data, 
                                                                interval, 
                                                                st.session_state.option_x, 
                                                                st.session_state.option_y, 
                                                                st.session_state.option_c,
                                                                )
                                                )
                if len(selected_points) > 0:
                    del[selected_points['name']]
            
            with col22:
                selected_df = None
                if len(selected_points) != 0:
                    query = ' & '.join(
                    f'{crange[0]} <= `{col}` <= {crange[1]}'
                    for col, crange in selected_points.items())
                    selected_df = data.query(query)
                    st.write(f"Total selected points: {len(selected_df)}")
                    st.dataframe(selected_df, width=800, height=260,use_container_width=False)
                else:
                    st.write("No Selection")
                
                if selected_df is not None:
                    st.write("Histogram of selected data:")
                    histogram_x = bar_plot(selected_df, st.session_state.option_x)
                    histogram_y = bar_plot(selected_df, st.session_state.option_y)
                    st.write(alt.hconcat(histogram_x,histogram_y))
                else:
                    st.write("Histogram of entire data:")
                    histogram_x = bar_plot(data, st.session_state.option_x)
                    histogram_y = bar_plot(data, st.session_state.option_y)
                    st.write(alt.hconcat(histogram_x,histogram_y))

        #Outlier Removal-----------------------------------------------------------------------
            st.write('---')
            df_nomarlized = data.copy()
            curve_editting = st.selectbox("Select curve to edit:",
                                        key="selected_curve",
                                        options=columns_list(data, no_depth=True, no_well=True),
                                        )
            n_value = int(st.text_input("Number of rows for Mean calculation ", "5"))

            def normalize_outlier(df_nomarlized, selected_df, curve, n_value):
                n=n_value//2
                for i in selected_df.index:
                    df_nomarlized.loc[[i],curve.upper()] = df_nomarlized.loc[i-n:i+n,curve.upper()].mean()
                return df_nomarlized
            def remove_data_point(df_nomarlized, selected_df, curve):
                for i in selected_df.index:
                    df_nomarlized[i, curve] = 0                   #ERROR ALARM!!!!
                    # df_nomarlized = df_nomarlized.drop(index=i)     #ERROR ALARM!!!!
                return df_nomarlized
                
            if st.button("Outliers Processing"):
                st.session_state.fdata = normalize_outlier(df_nomarlized, selected_df, curve_editting, n_value)
                _well = "".join((st.session_state.fdata.WELL.unique()).tolist())
                st.session_state.loc_data = pd.concat([df[(df["WELL"] != _well)],st.session_state.fdata], axis=0)
                selected_df = None
            if st.button("Remove"):
                st.session_state.fdata = remove_data_point(df_nomarlized, selected_df, curve_editting)
                _well = "".join((st.session_state.fdata.WELL.unique()).tolist())
                st.write(_well)
                st.write(type(_well))
                st.session_state.loc_data = pd.concat([df[(df["WELL"] != _well)],st.session_state.fdata], axis=0)
                selected_df = None

        #Curve View-----------------------------------------------------------------------
            def plt_curs(data, option_w):
                data_plt = data[data["WELL"] == option_w]
                if plotting_curves != []:
                    for i, c in enumerate(plotting_curves):
                        charts_dict[i] = curve_plot(data=data_plt,filted_data=selected_df, x_column=c)
            
            # with col2:
            charts_dict={}
            plotting_curves = st.multiselect("Select curves to plot:", key="curvs_plt", options=columns_list(data, no_depth=True, no_well=True))
            
            if st.session_state.option_w is not None:
                if 'loc_data' not in st.session_state:
                    plt_curs(df_nomarlized, st.session_state.option_w)
                else:
                    plt_curs(st.session_state.loc_data, st.session_state.option_w)
                    
            #Show Curve-----------------------------------------------------------------------
                st.write(alt.concat(*charts_dict.values()).configure(autosize='fit'))#.configure_concat(spacing=0))
                
        #------------------------
        def check_method(df):
            if st.session_state.method == "Single Well":
                data = df[df.WELL == st.session_state.option_w]
                data = data.sort_values(by=['DEPTH'])
                data = data.reset_index().drop(["index"], axis=1)
            else:
                data = df
            return data
        #------------------------
        
        if 'loc_data' not in st.session_state:
            data = check_method(df)
        else:
            data = check_method(st.session_state.loc_data)       
        
        rm_outliers(data)
            
    # # Download --------------------------------------------------------------
    st.write('---')
    st.write("Download final result to csv file")
    if "loc_data" not in st.session_state:
        saving_df = df
    else:
        saving_df = st.session_state.loc_data
    st.download_button(label='Download',
                    data = saving_df.to_csv(),
                    file_name='Query_data.csv',
                    mime='text/csv')