streamlit_UI_code.py

import streamlit as st
import pandas as pd
import os
from glob import glob


from pymongo import MongoClient
import pandas as pd
from glob import glob

def ibis_ngs_db_connection():
    username_ibis = ''
    password_ibis = #in url format
    #database 
    connection_string_ibis = f'database://{username_ibis}:{password_ibis}@ipaddress/'
    client_ibis = MongoClient(connection_string_ibis)
    ibis_ngs_db = client_ibis['database name']
    return ibis_ngs_db

def common_uat_db_connection():
    username_ibis = ''
    password_ibis = #in url format
    connection_string_ibis = f'database://{username_ibis}:{password_ibis}@ipaddress/'
    client_ibis = MongoClient(connection_string_ibis)
    ibis_ngs_db = client_ibis['database name']
    return ibis_ngs_db


def aon_ngs_db_connection():
    username_aon = ''
    password_aon = #in url format
    connection_string_aon = f'database://{username_aon}:{password_aon}@ipaddress/'
    client_aon = MongoClient(connection_string_aon)
    aon_ngs_db = client_aon['database name']
    return aon_ngs_db


def doc_textNLP(docID, db):
    temp_entity = db.documents.find_one({'docID':{'$in':[f"{docID}"]}})
    doc_text_df = pd.DataFrame(temp_entity['docTextNLP'])
    return doc_text_df

def low_quality_report(docID, db):
    temp_entity = db.documents.find_one({'docID':{'$in':[f"{docID}"]}})
    low_quality_df = pd.DataFrame(temp_entity['docTextNLPLowQuality'])
    return low_quality_df

def db_selection(doc_id):
    if f'{doc_id}'.isdigit():
        db = ibis_ngs_db_connection()
    else:
        db = ibis_ngs_db_connection()
    return db


# Function to get list of document IDs
def get_document_ids():
    # Assuming 'abc' is the folder containing documents
    document_ids = [file.split('_')[0] for file in os.listdir('abc') if file.endswith('.png')]
    return list(set(document_ids))

# Function to load image based on selected document ID and page number
def load_image(image_path, document_id, page_number):
    im_path = f"{image_path}{document_id}-{page_number-1}.png"
    if os.path.exists(im_path):
        return im_path
    else:
        return None

# Function to load dataframe based on selected document ID
def load_dataframe(auto_csv_path, document_id, page_number):
    csv_path = glob(f'{auto_csv_path}*{document_id}*auto.csv')
    print(csv_path)
    if len(csv_path)>0:
        auto_df = pd.read_csv(csv_path[0])
        auto_df_page = auto_df[auto_df['Page#']==page_number]
        return auto_df_page
    else:
        return None
    
def path_setting(PhaseData_path, Inbound_CSV_path):
    auto_csv_path = f'{PhaseData_path}/Batch1/NLP_batch/'
    image_path = f'{PhaseData_path}/Data/output/images/'
    inbound_df = pd.read_csv(Inbound_CSV_path)
    pif_list = list(inbound_df.pif_key.values)
    return pif_list, image_path, auto_csv_path

def main():
    st.set_page_config(layout="wide")
    st.write("### Input Paths")
    PhaseData_path = st.text_input("Enter PhaseData Path:")
    Inbound_CSV_path = st.text_input("Enter Inbound_CSV Path:")
    
    if not PhaseData_path or not Inbound_CSV_path:
        st.warning("Please enter both PhaseData Path and Inbound_CSV Path.")
        return
    
    inbound_df_path = 'inbound_issues_tempus_2_q2_new.csv'
    pif_list, image_path, auto_csv_path = path_setting(PhaseData_path, Inbound_CSV_path)
    
    col1_width = st.sidebar.slider("Width of First Column", 0.1, 10.0, 2.0, 0.1)
    col2_width = st.sidebar.slider("Width of Second Column", 0.1, 10.0, 6.5, 0.1)
    col3_width = st.sidebar.slider("Width of Third Column", 0.1, 10.0, 5.0, 0.1)

    col1, col2, col3 = st.columns([col1_width, col2_width, col3_width])

    with col1:
        st.write("### Document Selection")
        global doc_index
        doc_index = st.number_input("Select Document Index", min_value=1, max_value=len(pif_list)+1, step=1, value=1)
        document_id = pif_list[doc_index-1]
        st.write("Current Document ID: ", document_id)
        pages = [int(i.split('-')[-1].split('.')[0]) for i in glob(f"{image_path}{document_id}*.png")]
        page_number = st.number_input("Page Number", min_value=1, max_value=len(pages), step=1, value=1)
        if f'{document_id}'.isdigit():
            db = ibis_ngs_db_connection()
        else:
            db = aon_ngs_db_connection()
        db_df = doc_textNLP(f'{document_id}', db)
        df_tmp = pd.read_csv(glob(f"{auto_csv_path}*{document_id}*auto.csv")[0])
        reason_for_onhold = st.text_area("Reason for On-hold: ", value="Add a reason for onhold column into inbound CSV")
        comment_pipeline_db = f"#BM DB Page:{db_df[db_df['Page#']==page_number].shape[0]}\n#BM Pipeline Page:{df_tmp[df_tmp['Page#']==page_number].shape[0]}\n-----------------------------\n#BM DB Total:{db_df.shape[0]}\n#BM Pipeline Total:{df_tmp.shape[0]}"
        pipeline_db_stats = st.text_area("Biomarker Stats#: ", value=comment_pipeline_db, height=150)

    with col2:
        st.write("### Display Image")
        im_path = load_image(image_path, document_id, page_number)
        if im_path:
            st.image(im_path)
        else:
            st.write("Image not found")

    with col3:
        st.write("### Display DataFrame")
        df = load_dataframe(auto_csv_path, document_id, page_number, )
        if df is not None:
            columns_to_display = st.multiselect("Select Columns to Display", df.columns)
            if len(columns_to_display) > 0:
                st.write(df[columns_to_display])
                st.subheader("Add Comments on NLP Output")
                reco_list = ['Regular NLP working', 'Wrong Report type', 'Report not Found',
                             'Dev work required', 'NLP not supported (non 5 labs)', 
                             'Limited Regular NLP Support and Manual NLP (ROI) is working', 
                             'Poor quality report- NLP not supported']
                comment = st.selectbox("Select Comment", options=reco_list)
                add_comment = st.text_area("Add your additional comment here")
                if st.button("Submit"):
                    global comments_df
                    data = {
                        "Document ID": [document_id],
                        "Page": [page_number],
                        "Comment": [comment],
                        "Additional Comment": [add_comment]
                    }
                    comments_df = comments_df.append(pd.DataFrame(data), ignore_index=True)
                    comments_csv = "comments.csv"
                    if os.path.exists(comments_csv):
                        comments_df.to_csv(comments_csv, mode="a", header=False, index=False)
                    else:
                        comments_df.to_csv(comments_csv, index=False)
                    
            else:
                st.write("No columns selected")
        else:
            st.write("DataFrame not found")

if __name__ == "__main__":
    main()