# !pip install streamlit
# !pip install pandas

import pandas as pd
import streamlit as st
import base64
import io
import base64

# Functions
def map_data_to_template(mapping_df, template_df, data_df):
    # Initialize the final output dataframe with the template columns, filled with NaN
    final_output_df = pd.DataFrame(columns=template_df.columns)

    # Prepare a dictionary to hold the mapping from MEDLab to NDA variables
    variable_mapping = mapping_df.set_index('MEDLab Variable')['NDA Variable'].to_dict()

    # Iterate over each NDA variable to map the data
    for nda_var in final_output_df.columns:
        medlab_vars = [medlab_var for medlab_var, nda_mapped_var in variable_mapping.items() if nda_mapped_var == nda_var]
        
        # Initialize the column with None
        final_output_df[nda_var] = [None] * len(data_df)
        
        # Go through each potential MEDLab variable until we find one that's present and has data
        for medlab_var in medlab_vars:
            if medlab_var in data_df.columns and not data_df[medlab_var].isnull().all():
                # If a date column, convert to the specified format
                if 'date' in medlab_var:
                    final_output_df[nda_var] = pd.to_datetime(data_df[medlab_var], errors='coerce').dt.strftime('%m/%d/%Y')
                else:
                    final_output_df[nda_var] = data_df[medlab_var]
                break  # Stop checking once we've mapped one

    return final_output_df


# Streamlit app
def main():
    st.markdown("<h1 style='text-align: center; color: #E694FF;'>Data Transformer</h1>", unsafe_allow_html=True)

    # File Uploader for each CSV
    st.subheader("Upload Files")
    nimh_template_file = st.file_uploader("Choose NIMH Template CSV", type=['csv'])
    redcap_data_file = st.file_uploader("Choose REDCap Data CSV", type=['csv'])
    conversion_key_file = st.file_uploader("Choose Conversion Key CSV", type=['csv'])

    if nimh_template_file and redcap_data_file and conversion_key_file:
        # Convert the file objects to DataFrames
        nimh_template_df = pd.read_csv(io.StringIO(nimh_template_file.getvalue().decode('utf-8')), skiprows=1)
        redcap_data_df = pd.read_csv(io.StringIO(redcap_data_file.getvalue().decode('utf-8')))
        conversion_key_df = pd.read_csv(io.StringIO(conversion_key_file.getvalue().decode('utf-8')))
        
        transformed_data_df = map_data_to_template(
            conversion_key_df, 
            nimh_template_df, 
            redcap_data_df
        )
        
        # Display transformed data
        st.subheader("Transformed Data")
        st.write(transformed_data_df)

        # Download button for transformed data
        st.subheader("Download Transformed Data")
        csv = transformed_data_df.to_csv(index=False)
        b64 = base64.b64encode(csv.encode()).decode()  # some strings <-> bytes conversions necessary here
        href = f'<a href="data:file/csv;base64,{b64}" download="transformed_data.csv">Download CSV File</a>'
        st.markdown(href, unsafe_allow_html=True)


if __name__ == '__main__':
    main()