Streamlit client source code

- streamlit frontend
- Api backend url=https://gabcares-sepsis-fastapi.hf.space/api/v1/prediction?model

.streamlit/config.toml

@@ -0,0 +1,2 @@
+[client]
+showSidebarNavigation = false

Dockerfile

@@ -0,0 +1,25 @@
+FROM python:3.11.9-slim
+
+# Copy requirements file
+COPY requirements.txt .
+
+# Update pip
+RUN pip --timeout=3000 install --no-cache-dir --upgrade pip
+
+# Install dependecies
+RUN pip --timeout=3000 install --no-cache-dir -r requirements.txt
+
+# Make project directory
+RUN mkdir -p /src/client/
+
+# Set working directory
+WORKDIR /src/client
+
+# Copy client frontend
+COPY . .
+
+# Expose app port
+EXPOSE 8501
+
+# Start application
+CMD ["streamlit", "run", "app.py"]

app.py

@@ -0,0 +1,484 @@
+import os
+import time
+import httpx
+import string
+import random
+import datetime as dt
+from dotenv import load_dotenv
+
+import streamlit as st
+import extra_streamlit_components as stx
+
+import asyncio
+from aiocache import cached, Cache
+
+import pandas as pd
+from typing import Optional, Callable
+
+from config import ENV_PATH, BEST_MODELS, TEST_FILE, TEST_FILE_URL, HISTORY_FILE, markdown_table_all
+
+from utils.navigation import navigation
+from utils.footer import footer
+from utils.janitor import Janitor
+
+
+# Load ENV
+load_dotenv(ENV_PATH)  # API_URL
+
+# Set page configuration
+st.set_page_config(
+    page_title="Homepage",
+    page_icon="🤖",
+    layout="wide",
+    initial_sidebar_state='auto'
+)
+
+
+@cached(ttl=10, cache=Cache.MEMORY, namespace='streamlit_savedataset')
+# @st.cache_data(show_spinner="Saving datasets...") # Streamlit cache is yet to support async functions
+async def save_dataset(df: pd.DataFrame, filepath, csv=True) -> None:
+    async def save(df: pd.DataFrame, file):
+        return df.to_csv(file, index=False) if csv else df.to_excel(file, index=False)
+
+    async def read(file):
+        return pd.read_csv(file) if csv else pd.read_excel(file)
+
+    async def same_dfs(df: pd.DataFrame, df2: pd.DataFrame):
+        return df.equals(df2)
+
+    if not os.path.isfile(filepath):  # Save if file does not exists
+        await save(df, filepath)
+    else:  # Save if data are not same
+        df_old = await read(filepath)
+        if not await same_dfs(df, df_old):
+            await save(df, filepath)
+
+
+@cached(ttl=10, cache=Cache.MEMORY, namespace='streamlit_testdata')
+async def get_test_data():
+    try:
+        df_test_raw = pd.read_csv(TEST_FILE_URL)
+        await save_dataset(df_test_raw, TEST_FILE, csv=True)
+    except Exception:
+        df_test_raw = pd.read_csv(TEST_FILE)
+
+    # Some house keeping, clean df
+    df_test = df_test_raw.copy()
+    janitor = Janitor()
+    df_test = janitor.clean_dataframe(df_test)  # Cleaned
+
+    return df_test_raw, df_test
+
+
+# Function for selecting models
+async def select_model() -> str:
+    col1, _ = st.columns(2)
+    with col1:
+        selected_model = st.selectbox(
+            'Select a model', options=BEST_MODELS, key='selected_model')
+
+    return selected_model
+
+
+async def endpoint(model: str) -> str:
+    api_url = os.getenv("API_URL")
+    model_endpoint = f"{api_url}={model}"
+    return model_endpoint
+
+
+# Function for making prediction
+async def make_prediction(model_endpoint) -> Optional[pd.DataFrame]:
+
+    test_data = await get_test_data()
+    _, df_test = test_data
+
+    df: pd.DataFrame = None
+    search_patient = st.session_state.get('search_patient', False)
+    search_patient_id = st.session_state.get('search_patient_id', False)
+    manual_patient_id = st.session_state.get('manual_patient_id', False)
+    if isinstance(search_patient_id, str) and search_patient_id:  # And not empty string
+        search_patient_id = [search_patient_id]
+    if search_patient and search_patient_id:  # Search Form df and a patient was selected
+        mask = df_test['id'].isin(search_patient_id)
+        df_form = df_test[mask]
+        df = df_form.copy()
+    elif not (search_patient or search_patient_id) and manual_patient_id:  # Manual form df
+        columns = ['manual_patient_id', 'prg', 'pl', 'pr', 'sk',
+                   'ts', 'm11', 'bd2', 'age', 'insurance']
+        data = {c: [st.session_state.get(c)] for c in columns}
+        data['insurance'] = [1 if i == 'Yes' else 0 for i in data['insurance']]
+
+        # Make a DataFrame
+        df = pd.DataFrame(data).rename(
+            columns={'manual_patient_id': 'id'})
+        columns_int = ['prg', 'pl', 'pr', 'sk', 'ts', 'age']
+        columns_float = ['m11', 'bd2']
+
+        df[columns_int] = df[columns_int].astype(int)
+        df[columns_float] = df[columns_float].astype(float)
+    else:  # Form did not send a patient
+        message = 'You must choose valid patient(s) from the select box.'
+        icon = '😞'
+        st.toast(message, icon=icon)
+        st.warning(message, icon=icon)
+
+    if df is not None:
+        try:
+            # JSON data
+            data = df.to_dict(orient='list')
+
+            # Send POST request with JSON data using the json parameter
+            async with httpx.AsyncClient() as client:
+                response = await client.post(model_endpoint, json=data, timeout=30)
+                response.raise_for_status()  # Ensure we catch any HTTP errors
+
+            if (response.status_code == 200):
+                pred_prob = (response.json()['result'])
+                prediction = pred_prob['prediction'][0]
+                probability = pred_prob['probability'][0]
+
+                # Store results in session state
+                st.session_state['prediction'] = prediction
+                st.session_state['probability'] = probability
+                df['prediction'] = prediction
+                df['probability (%)'] = probability
+                df['time_of_prediction'] = pd.Timestamp(dt.datetime.now())
+                df['model_used'] = st.session_state['selected_model']
+
+                df.to_csv(HISTORY_FILE, mode='a',
+                          header=not os.path.isfile(HISTORY_FILE))
+        except Exception as e:
+            st.error(f'😞 Unable to connect to the API server. {e}')
+
+    return df
+
+
+async def convert_string(df: pd.DataFrame, string: str) -> str:
+    return string.upper() if all(col.isupper() for col in df.columns) else string
+
+
+async def make_predictions(model_endpoint, df_uploaded=None, df_uploaded_clean=None) -> Optional[pd.DataFrame]:
+
+    df: pd.DataFrame = None
+    search_patient = st.session_state.get('search_patient', False)
+    patient_id_bulk = st.session_state.get('patient_id_bulk', False)
+    upload_bulk_predict = st.session_state.get('upload_bulk_predict', False)
+    if search_patient and patient_id_bulk:  # Search Form df and a patient was selected
+        _, df_test = await get_test_data()
+        mask = df_test['id'].isin(patient_id_bulk)
+        df_bulk: pd.DataFrame = df_test[mask]
+        df = df_bulk.copy()
+
+    elif not (search_patient or patient_id_bulk) and upload_bulk_predict:  # Upload widget df
+        df = df_uploaded_clean.copy()
+    else:  # Form did not send a patient
+        message = 'You must choose valid patient(s) from the select box.'
+        icon = '😞'
+        st.toast(message, icon=icon)
+        st.warning(message, icon=icon)
+
+    if df is not None:  # df should be set by form input or upload widget
+        try:
+            # JSON data
+            data = df.to_dict(orient='list')
+
+            # Send POST request with JSON data using the json parameter
+            async with httpx.AsyncClient() as client:
+                response = await client.post(model_endpoint, json=data, timeout=30)
+                response.raise_for_status()  # Ensure we catch any HTTP errors
+
+            if (response.status_code == 200):
+                pred_prob = (response.json()['result'])
+                predictions = pred_prob['prediction']
+                probabilities = pred_prob['probability']
+
+                # Add columns sepsis, probability, time, and model used to uploaded df and form df
+
+                async def add_columns(df):
+                    df[await convert_string(df, 'sepsis')] = predictions
+                    df[await convert_string(df, 'probability_(%)')] = probabilities
+                    df[await convert_string(df, 'time_of_prediction')
+                       ] = pd.Timestamp(dt.datetime.now())
+                    df[await convert_string(df, 'model_used')
+                       ] = st.session_state['selected_model']
+
+                    return df
+
+                # Form df if search patient is true or df from Uploaded data
+                if search_patient:
+                    df = await add_columns(df)
+
+                    df.to_csv(HISTORY_FILE, mode='a', header=not os.path.isfile(
+                        HISTORY_FILE))  # Save only known patients
+
+                else:
+                    df = await add_columns(df_uploaded)  # Raw, No cleaning
+
+                # Store df with prediction results in session state
+                st.session_state['bulk_prediction_df'] = df
+        except Exception as e:
+            st.error(f'😞 Unable to connect to the API server. {e}')
+
+    return df
+
+
+def on_click(func: Callable, model_endpoint: str):
+    async def handle_click():
+        await func(model_endpoint)
+
+    loop = asyncio.new_event_loop()
+    asyncio.set_event_loop(loop)
+    loop.run_until_complete(handle_click())
+    loop.close()
+
+
+async def search_patient_form(model_endpoint: str) -> None:
+    test_data = await get_test_data()
+    _, df_test = test_data
+
+    patient_ids = df_test['id'].unique().tolist()+['']
+    if st.session_state['sidebar'] == 'single_prediction':
+        with st.form('search_patient_id_form'):
+            col1, _ = st.columns(2)
+            with col1:
+                st.write('#### Patient ID 🤒')
+                st.selectbox(
+                    'Search a patient', options=patient_ids, index=len(patient_ids)-1, key='search_patient_id')
+            st.form_submit_button('Predict', type='primary', on_click=on_click, kwargs=dict(
+                func=make_prediction, model_endpoint=model_endpoint))
+    else:
+        with st.form('search_patient_id_bulk_form'):
+            col1, _ = st.columns(2)
+            with col1:
+                st.write('#### Patient ID 🤒')
+                st.multiselect(
+                    'Search a patient', options=patient_ids, default=None, key='patient_id_bulk')
+            st.form_submit_button('Predict', type='primary', on_click=on_click, kwargs=dict(
+                func=make_predictions, model_endpoint=model_endpoint))
+
+
+async def gen_random_patient_id() -> str:
+    numbers = ''.join(random.choices(string.digits, k=6))
+    letters = ''.join(random.choices(string.ascii_lowercase, k=4))
+    return f"ICU{numbers}-gen-{letters}"
+
+
+async def manual_patient_form(model_endpoint) -> None:
+    with st.form('manual_patient_form'):
+
+        col1, col2, col3 = st.columns(3)
+
+        with col1:
+            st.write('### Patient Demographics 🛌')
+            st.text_input(
+                'ID', value=await gen_random_patient_id(), key='manual_patient_id')
+            st.number_input('Age: patients age (years)', min_value=0,
+                            max_value=100, step=1, key='age')
+            st.selectbox('Insurance: If a patient holds a valid insurance card', options=[
+                'Yes', 'No'], key='insurance')
+
+        with col2:
+            st.write('### Vital Signs 🩺')
+            st.number_input('BMI (weight in kg/(height in m)^2', min_value=10.0,
+                            format="%.2f", step=1.00, key='m11')
+            st.number_input(
+                'Blood Pressure (mm Hg)', min_value=10.0, format="%.2f", step=1.00, key='pr')
+            st.number_input(
+                'PRG (plasma glucose)', min_value=10.0, format="%.2f", step=1.00, key='prg')
+
+        with col3:
+            st.write('### Blood Work 💉')
+            st.number_input(
+                'PL: Blood Work Result-1 (mu U/ml)', min_value=10.0, format="%.2f", step=1.00, key='pl')
+            st.number_input(
+                'SK: Blood Work Result 2 (mm)', min_value=10.0, format="%.2f", step=1.00, key='sk')
+            st.number_input(
+                'TS: Blood Work Result-3 (mu U/ml)', min_value=10.0, format="%.2f", step=1.00, key='ts')
+            st.number_input(
+                'BD2: Blood Work Result-4 (mu U/ml)', min_value=10.0, format="%.2f", step=1.00, key='bd2')
+
+        st.form_submit_button('Predict', type='primary', on_click=on_click, kwargs=dict(
+            func=make_prediction, model_endpoint=model_endpoint))
+
+
+async def do_single_prediction(model_endpoint: str) -> None:
+    if st.session_state.get('search_patient', False):
+        await search_patient_form(model_endpoint)
+    else:
+        await manual_patient_form(model_endpoint)
+
+
+async def show_prediction() -> None:
+    final_prediction = st.session_state.get('prediction', None)
+    final_probability = st.session_state.get('probability', None)
+
+    if final_prediction is None:
+        st.markdown('#### Prediction will show below! 🔬')
+        st.divider()
+    else:
+        st.markdown('#### Prediction! 🔬')
+        st.divider()
+        if final_prediction.lower() == 'positive':
+            st.toast("Sepsis alert!", icon='🦠')
+            message = f"It is **{final_probability:.2f} %** likely that the patient will develop **sepsis.**"
+            st.warning(message, icon='😞')
+            time.sleep(5)
+            st.toast(message)
+        else:
+            st.toast("Continous monitoring", icon='🔬')
+            message = f"The patient will **not** develop sepsis with a likelihood of **{final_probability:.2f}%**."
+            st.success(message, icon='😊')
+            time.sleep(5)
+            st.toast(message)
+
+    # Set prediction and probability to None
+    st.session_state['prediction'] = None
+    st.session_state['probability'] = None
+
+
+# @st.cache_data(show_spinner=False) Caching results from async functions buggy
+async def convert_df(df: pd.DataFrame):
+    return df.to_csv(index=False)
+
+
+async def bulk_upload_widget(model_endpoint: str) -> None:
+    uploaded_file = st.file_uploader(
+        "Choose a CSV or Excel File", type=['csv', 'xls', 'xlsx'])
+
+    uploaded = uploaded_file is not None
+
+    upload_bulk_predict = st.button('Predict', type='primary',
+                                    help='Upload a csv/excel file to make predictions', disabled=not uploaded, key='upload_bulk_predict')
+    df = None
+    if upload_bulk_predict and uploaded:
+        df_test_raw, _ = await get_test_data()
+        # Uploadfile is a "file-like" object is accepted
+        try:
+            try:
+                df = pd.read_csv(uploaded_file)
+            except Exception:
+                df = pd.read_excel(uploaded_file)
+
+            df_columns = set(df.columns)
+            df_test_columns = set(df_test_raw.columns)
+            df_schema = df.dtypes
+            df_test_schema = df_test_raw.dtypes
+
+            if df_columns != df_test_columns or not df_schema.equals(df_test_schema):
+                df = None
+                raise Exception
+            else:
+                # Clean dataframe
+                janitor = Janitor()
+                df_clean = janitor.clean_dataframe(df)
+
+                df = await make_predictions(
+                    model_endpoint, df_uploaded=df, df_uploaded_clean=df_clean)
+
+        except Exception:
+            st.subheader('Data template')
+            data_template = df_test_raw[:3]
+            st.dataframe(data_template)
+            csv = await convert_df(data_template)
+            message_1 = 'Upload a valid csv or excel file.'
+            message_2 = f"{message_1.split('.')[0]} with the columns and schema of the above data template."
+            icon = '😞'
+            st.toast(message_1, icon=icon)
+
+            st.download_button(
+                label='Download template',
+                data=csv,
+                file_name='Data template.csv',
+                mime="text/csv",
+                type='secondary',
+                key='download-data-template'
+            )
+            st.info('Download the above template for use as a baseline structure.')
+
+            # Display explander to show the data dictionary
+            with st.expander("Expand to see the data dictionary", icon="💡"):
+                st.subheader("Data dictionary")
+                st.markdown(markdown_table_all)
+            st.warning(message_2, icon=icon)
+
+    return df
+
+
+async def do_bulk_prediction(model_endpoint: str) -> None:
+    if st.session_state.get('search_patient', False):
+        await search_patient_form(model_endpoint)
+    else:
+        # File uploader
+        await bulk_upload_widget(model_endpoint)
+
+
+async def show_bulk_predictions(df: pd.DataFrame) -> None:
+    if df is not None:
+        st.subheader("Bulk predictions 🔮", divider=True)
+        st.dataframe(df.astype(str))
+
+        csv = await convert_df(df)
+        message = 'The predictions are ready for download.'
+        icon = '⬇️'
+        st.toast(message, icon=icon)
+        st.info(message, icon=icon)
+        st.download_button(
+            label='Download predictions',
+            data=csv,
+            file_name='Bulk prediction.csv',
+            mime="text/csv",
+            type='secondary',
+            key='download-bulk-prediction'
+        )
+
+        # Set bulk prediction df to None
+        st.session_state['bulk_prediction_df'] = None
+
+
+async def sidebar(sidebar_type: str) -> st.sidebar:
+    return st.session_state.update({'sidebar': sidebar_type})
+
+
+async def main():
+    st.title("🤖 Predict Sepsis 🦠")
+
+    # Navigation
+    await navigation()
+
+    st.sidebar.toggle("Looking for a patient?", value=st.session_state.get(
+        'search_patient', False), key='search_patient')
+
+    selected_model = await select_model()
+    model_endpoint = await endpoint(selected_model)
+
+    selected_predict_tab = st.session_state.get('selected_predict_tab')
+    default = 1 if selected_predict_tab is None else selected_predict_tab
+
+    with st.spinner('A little house keeping...'):
+        time.sleep(st.session_state.get('sleep', 1.5))
+        chosen_id = stx.tab_bar(data=[
+            stx.TabBarItemData(id=1, title='🔬 Predict', description=''),
+            stx.TabBarItemData(id=2, title='🔮 Bulk predict',
+                               description=''),
+        ], default=default)
+        st.session_state['sleep'] = 0
+
+    if chosen_id == '1':
+        await sidebar('single_prediction')
+        await do_single_prediction(model_endpoint)
+        await show_prediction()
+
+    elif chosen_id == '2':
+        await sidebar('bulk_prediction')
+        df_with_predictions = await do_bulk_prediction(model_endpoint)
+        if df_with_predictions is None:
+            df_with_predictions = st.session_state.get(
+                'bulk_prediction_df', None)
+        await show_bulk_predictions(df_with_predictions)
+
+    # Add footer
+    await footer()
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

config.py

@@ -0,0 +1,46 @@
+import os
+from pathlib import Path
+
+# Paths
+BASE_DIR = './'
+DATA = os.path.join(BASE_DIR, 'data/')
+TEST_FILE = os.path.join(DATA, 'Paitients_Files_Test.csv')
+HISTORY = os.path.join(DATA, 'history/')
+HISTORY_FILE = os.path.join(HISTORY, 'history.csv')
+
+# Urls
+TEST_FILE_URL = "https://raw.githubusercontent.com/D0nG4667/sepsis_prediction_full_stack/model_development/dev/data/Paitients_Files_Test.csv"
+
+
+# ENV when using standalone streamlit server
+ENV_PATH = Path('../../env/online.env')
+
+ALL_MODELS = [
+    "AdaBoostClassifier",
+    "CatBoostClassifier",
+    "DecisionTreeClassifier",
+    "KNeighborsClassifier",
+    "LGBMClassifier",
+    "LogisticRegression",
+    "RandomForestClassifier",
+    "SupportVectorClassifier",
+    "XGBoostClassifier",
+]
+
+BEST_MODELS = ["RandomForestClassifier", "XGBoostClassifier"]
+
+markdown_table_all = """    
+| Column   Name                | Attribute/Target | Description                                                                                                                                                                                                  |
+|------------------------------|------------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
+| ID                           | N/A              | Unique number to represent patient ID                                                                                                                                                                        |
+| PRG           | Attribute1       |  Plasma glucose|
+| PL               | Attribute 2     |   Blood Work Result-1 (mu U/ml)                                                                                                                                                |
+| PR              | Attribute 3      | Blood Pressure (mm Hg)|
+| SK              | Attribute 4      | Blood Work Result-2 (mm)|
+| TS             | Attribute 5      |     Blood Work Result-3 (mu U/ml)|                                                                                  
+| M11     | Attribute 6    |  Body mass index (weight in kg/(height in m)^2|
+| BD2             | Attribute 7     |   Blood Work Result-4 (mu U/ml)|
+| Age              | Attribute 8      |    patients age  (years)|
+| Insurance | N/A     | If a patient holds a valid insurance card|
+| Sepsis                 | Target           | Positive: if a patient in ICU will develop a sepsis , and Negative: otherwise |
+"""

data/Paitients_Files_Test.csv

@@ -0,0 +1,170 @@
+ID,PRG,PL,PR,SK,TS,M11,BD2,Age,Insurance
+ICU200609,1,109,38,18,120,23.1,0.407,26,1
+ICU200610,1,108,88,19,0,27.1,0.4,24,1
+ICU200611,6,96,0,0,0,23.7,0.19,28,1
+ICU200612,1,124,74,36,0,27.8,0.1,30,1
+ICU200613,7,150,78,29,126,35.2,0.692,54,0
+ICU200614,4,183,0,0,0,28.4,0.212,36,1
+ICU200615,1,124,60,32,0,35.8,0.514,21,1
+ICU200616,1,181,78,42,293,40.0,1.258,22,1
+ICU200617,1,92,62,25,41,19.5,0.482,25,0
+ICU200618,0,152,82,39,272,41.5,0.27,27,0
+ICU200619,1,111,62,13,182,24.0,0.138,23,1
+ICU200620,3,106,54,21,158,30.9,0.292,24,1
+ICU200621,3,174,58,22,194,32.9,0.593,36,1
+ICU200622,7,168,88,42,321,38.2,0.787,40,1
+ICU200623,6,105,80,28,0,32.5,0.878,26,1
+ICU200624,11,138,74,26,144,36.1,0.557,50,1
+ICU200625,3,106,72,0,0,25.8,0.207,27,1
+ICU200626,6,117,96,0,0,28.7,0.157,30,1
+ICU200627,2,68,62,13,15,20.1,0.257,23,0
+ICU200628,9,112,82,24,0,28.2,1.282,50,1
+ICU200629,0,119,0,0,0,32.4,0.141,24,1
+ICU200630,2,112,86,42,160,38.4,0.246,28,1
+ICU200631,2,92,76,20,0,24.2,1.698,28,1
+ICU200632,6,183,94,0,0,40.8,1.461,45,0
+ICU200633,0,94,70,27,115,43.5,0.347,21,1
+ICU200634,2,108,64,0,0,30.8,0.158,21,1
+ICU200635,4,90,88,47,54,37.7,0.362,29,1
+ICU200636,0,125,68,0,0,24.7,0.206,21,1
+ICU200637,0,132,78,0,0,32.4,0.393,21,1
+ICU200638,5,128,80,0,0,34.6,0.144,45,1
+ICU200639,4,94,65,22,0,24.7,0.148,21,1
+ICU200640,7,114,64,0,0,27.4,0.732,34,1
+ICU200641,0,102,78,40,90,34.5,0.238,24,0
+ICU200642,2,111,60,0,0,26.2,0.343,23,0
+ICU200643,1,128,82,17,183,27.5,0.115,22,0
+ICU200644,10,92,62,0,0,25.9,0.167,31,0
+ICU200645,13,104,72,0,0,31.2,0.465,38,0
+ICU200646,5,104,74,0,0,28.8,0.153,48,0
+ICU200647,2,94,76,18,66,31.6,0.649,23,0
+ICU200648,7,97,76,32,91,40.9,0.871,32,0
+ICU200649,1,100,74,12,46,19.5,0.149,28,0
+ICU200650,0,102,86,17,105,29.3,0.695,27,1
+ICU200651,4,128,70,0,0,34.3,0.303,24,1
+ICU200652,6,147,80,0,0,29.5,0.178,50,1
+ICU200653,4,90,0,0,0,28.0,0.61,31,1
+ICU200654,3,103,72,30,152,27.6,0.73,27,1
+ICU200655,2,157,74,35,440,39.4,0.134,30,1
+ICU200656,1,167,74,17,144,23.4,0.447,33,0
+ICU200657,0,179,50,36,159,37.8,0.455,22,0
+ICU200658,11,136,84,35,130,28.3,0.26,42,0
+ICU200659,0,107,60,25,0,26.4,0.133,23,1
+ICU200660,1,91,54,25,100,25.2,0.234,23,0
+ICU200661,1,117,60,23,106,33.8,0.466,27,0
+ICU200662,5,123,74,40,77,34.1,0.269,28,0
+ICU200663,2,120,54,0,0,26.8,0.455,27,1
+ICU200664,1,106,70,28,135,34.2,0.142,22,1
+ICU200665,2,155,52,27,540,38.7,0.24,25,0
+ICU200666,2,101,58,35,90,21.8,0.155,22,1
+ICU200667,1,120,80,48,200,38.9,1.162,41,1
+ICU200668,11,127,106,0,0,39.0,0.19,51,1
+ICU200669,3,80,82,31,70,34.2,1.292,27,1
+ICU200670,10,162,84,0,0,27.7,0.182,54,0
+ICU200671,1,199,76,43,0,42.9,1.394,22,1
+ICU200672,8,167,106,46,231,37.6,0.165,43,1
+ICU200673,9,145,80,46,130,37.9,0.637,40,1
+ICU200674,6,115,60,39,0,33.7,0.245,40,1
+ICU200675,1,112,80,45,132,34.8,0.217,24,1
+ICU200676,4,145,82,18,0,32.5,0.235,70,1
+ICU200677,10,111,70,27,0,27.5,0.141,40,1
+ICU200678,6,98,58,33,190,34.0,0.43,43,1
+ICU200679,9,154,78,30,100,30.9,0.164,45,1
+ICU200680,6,165,68,26,168,33.6,0.631,49,1
+ICU200681,1,99,58,10,0,25.4,0.551,21,1
+ICU200682,10,68,106,23,49,35.5,0.285,47,1
+ICU200683,3,123,100,35,240,57.3,0.88,22,0
+ICU200684,8,91,82,0,0,35.6,0.587,68,0
+ICU200685,6,195,70,0,0,30.9,0.328,31,0
+ICU200686,9,156,86,0,0,24.8,0.23,53,1
+ICU200687,0,93,60,0,0,35.3,0.263,25,1
+ICU200688,3,121,52,0,0,36.0,0.127,25,1
+ICU200689,2,101,58,17,265,24.2,0.614,23,1
+ICU200690,2,56,56,28,45,24.2,0.332,22,0
+ICU200691,0,162,76,36,0,49.6,0.364,26,1
+ICU200692,0,95,64,39,105,44.6,0.366,22,1
+ICU200693,4,125,80,0,0,32.3,0.536,27,1
+ICU200694,5,136,82,0,0,0.0,0.64,69,1
+ICU200695,2,129,74,26,205,33.2,0.591,25,1
+ICU200696,3,130,64,0,0,23.1,0.314,22,1
+ICU200697,1,107,50,19,0,28.3,0.181,29,1
+ICU200698,1,140,74,26,180,24.1,0.828,23,1
+ICU200699,1,144,82,46,180,46.1,0.335,46,1
+ICU200700,8,107,80,0,0,24.6,0.856,34,1
+ICU200701,13,158,114,0,0,42.3,0.257,44,1
+ICU200702,2,121,70,32,95,39.1,0.886,23,1
+ICU200703,7,129,68,49,125,38.5,0.439,43,0
+ICU200704,2,90,60,0,0,23.5,0.191,25,0
+ICU200705,7,142,90,24,480,30.4,0.128,43,0
+ICU200706,3,169,74,19,125,29.9,0.268,31,1
+ICU200707,0,99,0,0,0,25.0,0.253,22,1
+ICU200708,4,127,88,11,155,34.5,0.598,28,1
+ICU200709,4,118,70,0,0,44.5,0.904,26,1
+ICU200710,2,122,76,27,200,35.9,0.483,26,0
+ICU200711,6,125,78,31,0,27.6,0.565,49,1
+ICU200712,1,168,88,29,0,35.0,0.905,52,1
+ICU200713,2,129,0,0,0,38.5,0.304,41,1
+ICU200714,4,110,76,20,100,28.4,0.118,27,1
+ICU200715,6,80,80,36,0,39.8,0.177,28,1
+ICU200716,10,115,0,0,0,0.0,0.261,30,1
+ICU200717,2,127,46,21,335,34.4,0.176,22,1
+ICU200718,9,164,78,0,0,32.8,0.148,45,1
+ICU200719,2,93,64,32,160,38.0,0.674,23,1
+ICU200720,3,158,64,13,387,31.2,0.295,24,1
+ICU200721,5,126,78,27,22,29.6,0.439,40,1
+ICU200722,10,129,62,36,0,41.2,0.441,38,1
+ICU200723,0,134,58,20,291,26.4,0.352,21,0
+ICU200724,3,102,74,0,0,29.5,0.121,32,0
+ICU200725,7,187,50,33,392,33.9,0.826,34,0
+ICU200726,3,173,78,39,185,33.8,0.97,31,1
+ICU200727,10,94,72,18,0,23.1,0.595,56,1
+ICU200728,1,108,60,46,178,35.5,0.415,24,1
+ICU200729,5,97,76,27,0,35.6,0.378,52,1
+ICU200730,4,83,86,19,0,29.3,0.317,34,0
+ICU200731,1,114,66,36,200,38.1,0.289,21,0
+ICU200732,1,149,68,29,127,29.3,0.349,42,0
+ICU200733,5,117,86,30,105,39.1,0.251,42,0
+ICU200734,1,111,94,0,0,32.8,0.265,45,0
+ICU200735,4,112,78,40,0,39.4,0.236,38,0
+ICU200736,1,116,78,29,180,36.1,0.496,25,0
+ICU200737,0,141,84,26,0,32.4,0.433,22,0
+ICU200738,2,175,88,0,0,22.9,0.326,22,0
+ICU200739,2,92,52,0,0,30.1,0.141,22,1
+ICU200740,3,130,78,23,79,28.4,0.323,34,1
+ICU200741,8,120,86,0,0,28.4,0.259,22,1
+ICU200742,2,174,88,37,120,44.5,0.646,24,1
+ICU200743,2,106,56,27,165,29.0,0.426,22,1
+ICU200744,2,105,75,0,0,23.3,0.56,53,1
+ICU200745,4,95,60,32,0,35.4,0.284,28,1
+ICU200746,0,126,86,27,120,27.4,0.515,21,1
+ICU200747,8,65,72,23,0,32.0,0.6,42,1
+ICU200748,2,99,60,17,160,36.6,0.453,21,1
+ICU200749,1,102,74,0,0,39.5,0.293,42,1
+ICU200750,11,120,80,37,150,42.3,0.785,48,1
+ICU200751,3,102,44,20,94,30.8,0.4,26,1
+ICU200752,1,109,58,18,116,28.5,0.219,22,1
+ICU200753,9,140,94,0,0,32.7,0.734,45,1
+ICU200754,13,153,88,37,140,40.6,1.174,39,1
+ICU200755,12,100,84,33,105,30.0,0.488,46,0
+ICU200756,1,147,94,41,0,49.3,0.358,27,0
+ICU200757,1,81,74,41,57,46.3,1.096,32,1
+ICU200758,3,187,70,22,200,36.4,0.408,36,1
+ICU200759,6,162,62,0,0,24.3,0.178,50,1
+ICU200760,4,136,70,0,0,31.2,1.182,22,1
+ICU200761,1,121,78,39,74,39.0,0.261,28,1
+ICU200762,3,108,62,24,0,26.0,0.223,25,1
+ICU200763,0,181,88,44,510,43.3,0.222,26,1
+ICU200764,8,154,78,32,0,32.4,0.443,45,1
+ICU200765,1,128,88,39,110,36.5,1.057,37,1
+ICU200766,7,137,90,41,0,32.0,0.391,39,0
+ICU200767,0,123,72,0,0,36.3,0.258,52,1
+ICU200768,1,106,76,0,0,37.5,0.197,26,1
+ICU200769,6,190,92,0,0,35.5,0.278,66,0
+ICU200770,2,88,58,26,16,28.4,0.766,22,1
+ICU200771,9,170,74,31,0,44.0,0.403,43,1
+ICU200772,9,89,62,0,0,22.5,0.142,33,1
+ICU200773,10,101,76,48,180,32.9,0.171,63,1
+ICU200774,2,122,70,27,0,36.8,0.34,27,1
+ICU200775,5,121,72,23,112,26.2,0.245,30,1
+ICU200776,1,126,60,0,0,30.1,0.349,47,1
+ICU200777,1,93,70,31,0,30.4,0.315,23,1

data/history/history.csv

@@ -0,0 +1,48 @@
+,id,prg,pl,pr,sk,ts,m11,bd2,age,insurance,prediction,probability (%),time_of_prediction,model_used
+0,ICU718842-gen-nlwl,10,10,10,10,10,10.0,10.0,0,1,Negative,72.0,2024-07-26 09:15:06.306895,RandomForestClassifier
+0,ICU639505-gen-yfab,10,10,10,10,10,10.0,10.0,0,1,Negative,72.0,2024-07-26 09:16:27.459841,RandomForestClassifier
+0,ICU195521-gen-hvkd,10,10,10,10,10,10.0,10.0,0,1,Negative,72.0,2024-07-26 09:16:31.500162,RandomForestClassifier
+0,ICU387670-gen-yumn,10,10,10,10,10,10.0,10.0,0,1,Negative,72.0,2024-07-26 09:17:16.749320,RandomForestClassifier
+0,ICU982466-gen-iczo,10,10,10,10,10,10.0,10.0,0,1,Negative,72.0,2024-07-26 09:17:53.997299,RandomForestClassifier
+0,ICU907863-gen-mrwx,10,10,10,10,10,10.0,10.0,0,1,Negative,72.0,2024-07-26 09:20:04.840719,RandomForestClassifier
+2,ICU200611,6,96,0,0,0,23.7,0.19,28,1,Negative,96.0,2024-07-26 09:23:42.879892,RandomForestClassifier
+0,ICU652368-gen-mczz,10,10,10,10,10,10.0,10.0,0,1,Negative,72.0,2024-07-26 09:38:19.827792,RandomForestClassifier
+0,ICU630077-gen-hmcu,10,10,10,10,10,10.0,10.0,0,1,Negative,72.0,2024-07-26 09:38:23.524249,RandomForestClassifier
+0,ICU860259-gen-gxvq,10,10,10,10,10,10.0,10.0,0,1,Negative,72.0,2024-07-26 17:13:49.802527,RandomForestClassifier
+6,ICU200615,1,124,60,32,0,35.8,0.514,21,1,Negative,79.0,2024-07-26 17:14:16.407693,RandomForestClassifier
+2,ICU200611,6,96,0,0,0,23.7,0.19,28,1,Negative,96.0,2024-07-26 17:14:35.174256,RandomForestClassifier
+4,ICU200613,7,150,78,29,126,35.2,0.692,54,0,Positive,63.0,2024-07-26 17:14:35.174256,RandomForestClassifier
+7,ICU200616,1,181,78,42,293,40.0,1.258,22,1,Positive,56.0,2024-07-26 17:14:35.174256,RandomForestClassifier
+8,ICU200617,1,92,62,25,41,19.5,0.482,25,0,Negative,99.0,2024-07-26 17:14:35.174256,RandomForestClassifier
+9,ICU200618,0,152,82,39,272,41.5,0.27,27,0,Negative,73.0,2024-07-26 17:14:35.174256,RandomForestClassifier
+10,ICU200619,1,111,62,13,182,24.0,0.138,23,1,Negative,100.0,2024-07-26 17:14:35.174256,RandomForestClassifier
+0,ICU857889-gen-akic,10,10,10,10,10,10.0,10.0,0,1,Negative,72.0,2024-07-26 17:36:06.901881,RandomForestClassifier
+1,ICU200610,1,108,88,19,0,27.1,0.4,24,1,Negative,88.0,2024-07-26 17:36:25.138350,RandomForestClassifier
+2,ICU200611,6,96,0,0,0,23.7,0.19,28,1,Negative,96.0,2024-07-26 17:36:51.337836,RandomForestClassifier
+4,ICU200613,7,150,78,29,126,35.2,0.692,54,0,Positive,63.0,2024-07-26 17:36:51.337836,RandomForestClassifier
+6,ICU200615,1,124,60,32,0,35.8,0.514,21,1,Negative,79.0,2024-07-26 17:36:51.337836,RandomForestClassifier
+7,ICU200616,1,181,78,42,293,40.0,1.258,22,1,Positive,56.0,2024-07-26 17:36:51.337836,RandomForestClassifier
+8,ICU200617,1,92,62,25,41,19.5,0.482,25,0,Negative,99.0,2024-07-26 17:36:51.337836,RandomForestClassifier
+9,ICU200618,0,152,82,39,272,41.5,0.27,27,0,Negative,73.0,2024-07-26 17:36:51.337836,RandomForestClassifier
+0,ICU335499-gen-perk,10,10,10,10,10,10.0,10.0,0,1,Negative,72.0,2024-07-26 18:06:45.472828,RandomForestClassifier
+0,ICU547969-gen-zvtk,10,10,10,10,10,10.0,10.0,0,1,Negative,72.0,2024-07-26 18:07:08.106437,RandomForestClassifier
+1,ICU200610,1,108,88,19,0,27.1,0.4,24,1,Negative,88.0,2024-07-26 18:07:28.070427,RandomForestClassifier
+3,ICU200612,1,124,74,36,0,27.8,0.1,30,1,Negative,69.0,2024-07-26 18:07:28.070427,RandomForestClassifier
+5,ICU200614,4,183,0,0,0,28.4,0.212,36,1,Positive,52.0,2024-07-26 18:07:28.070427,RandomForestClassifier
+7,ICU200616,1,181,78,42,293,40.0,1.258,22,1,Positive,56.0,2024-07-26 18:07:28.070427,RandomForestClassifier
+8,ICU200617,1,92,62,25,41,19.5,0.482,25,0,Negative,99.0,2024-07-26 18:07:28.070427,RandomForestClassifier
+9,ICU200618,0,152,82,39,272,41.5,0.27,27,0,Negative,73.0,2024-07-26 18:07:28.070427,RandomForestClassifier
+0,ICU930006-gen-yifq,10,10,10,10,10,10.0,10.0,0,1,Negative,72.0,2024-07-26 20:27:22.413639,RandomForestClassifier
+1,ICU200610,1,108,88,19,0,27.1,0.4,24,1,Negative,88.0,2024-07-26 20:29:38.351280,RandomForestClassifier
+1,ICU200610,1,108,88,19,0,27.1,0.4,24,1,Negative,88.0,2024-07-26 20:29:53.169096,RandomForestClassifier
+0,ICU766249-gen-ojbw,10,10,10,10,10,10.0,10.0,0,1,Negative,72.0,2024-07-26 20:30:20.584075,RandomForestClassifier
+1,ICU200610,1,108,88,19,0,27.1,0.4,24,1,Negative,88.0,2024-07-26 20:30:41.673970,RandomForestClassifier
+4,ICU200613,7,150,78,29,126,35.2,0.692,54,0,Positive,63.0,2024-07-26 20:30:41.673970,RandomForestClassifier
+6,ICU200615,1,124,60,32,0,35.8,0.514,21,1,Negative,79.0,2024-07-26 20:30:41.673970,RandomForestClassifier
+8,ICU200617,1,92,62,25,41,19.5,0.482,25,0,Negative,99.0,2024-07-26 20:30:41.673970,RandomForestClassifier
+3,ICU200612,1,124,74,36,0,27.8,0.1,30,1,Negative,69.0,2024-07-26 20:30:58.738631,RandomForestClassifier
+5,ICU200614,4,183,0,0,0,28.4,0.212,36,1,Positive,52.0,2024-07-26 20:30:58.738631,RandomForestClassifier
+6,ICU200615,1,124,60,32,0,35.8,0.514,21,1,Negative,79.0,2024-07-26 20:30:58.738631,RandomForestClassifier
+8,ICU200617,1,92,62,25,41,19.5,0.482,25,0,Negative,99.0,2024-07-26 20:30:58.738631,RandomForestClassifier
+0,ICU842996-gen-yvxc,10,10,10,10,10,10.0,10.0,0,1,Negative,72.0,2024-07-26 22:20:43.920677,RandomForestClassifier
+0,ICU453751-gen-bnlu,10,10,10,10,10,10.0,10.0,0,1,Negative,72.0,2024-07-26 22:21:25.044488,RandomForestClassifier

pages/01_🕰️_History.py

@@ -0,0 +1,60 @@
+import streamlit as st
+from streamlit_extras.dataframe_explorer import dataframe_explorer
+
+import asyncio
+from aiocache import cached, Cache
+
+import pandas as pd
+
+from utils.navigation import navigation
+from utils.footer import footer
+
+from config import HISTORY_FILE
+
+# Set page configuration
+st.set_page_config(
+    page_title='History Page',
+    page_icon='🕰️',
+    layout="wide",
+    initial_sidebar_state='auto'
+)
+
+
+# @st.cache_data(show_spinner="Getting history of predictions...")
+@cached(ttl=10, cache=Cache.MEMORY, namespace='streamlit_savedataset')
+async def get_history_data():
+    try:
+        df_history = pd.read_csv(HISTORY_FILE, index_col=0)
+        df_history['time_of_prediction'] = [timestamps[0]
+                                            for timestamps in df_history['time_of_prediction'].str.split('.')[0:]]
+
+        df_history['time_of_prediction'] = pd.to_datetime(
+            df_history['time_of_prediction'])
+    except Exception as e:
+        df_history = None
+
+    return df_history
+
+
+async def main():
+    st.title("Prediction History 🕰️")
+
+    # Navigation
+    await navigation()
+
+    df_history = await get_history_data()
+
+    if df_history is not None:
+        df_history_explorer = dataframe_explorer(df_history, case=False)
+
+        st.dataframe(df_history_explorer)
+    else:
+        st.info(
+            "There is no history file yet. Make a prediction.", icon='ℹ️')
+
+    # Add footer
+    await footer()
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

requirements.txt

@@ -0,0 +1,8 @@
+streamlit==1.36.0
+extra-streamlit-components==0.1.71
+streamlit-extras==0.4.3
+
+httpx==0.27.0
+aiocache==0.12.2
+
+python-dotenv==1.0.1

utils/footer.py

@@ -0,0 +1,26 @@
+import streamlit as st
+
+
+async def footer():
+    footer = """
+        <style>
+        .footer {
+            position: fixed;
+            bottom: 0;
+            left: 0;
+            width: 100%;
+            background-color: transparent;
+            color: #333;
+            text-align: center;
+            padding: 10px;
+            box-shadow: 0 -1px 5px rgba(0, 0, 0, 0.1);
+            z-index: 100; /* Prevent overlaying of page content on footer */
+        }
+        </style>
+        <div class="footer">
+                    &copy; 2024. Made with 💖 <a href="https://www.linkedin.com/in/dr-gabriel-okundaye" target="_blank" style="text-decoration: none;">Gabriel Okundaye</a>
+            <span style="color: #aaaaaa;">& Light ✨</span><br>
+        </div>
+        """
+
+    return st.markdown(footer, unsafe_allow_html=True)

utils/janitor.py

@@ -0,0 +1,49 @@
+import pandas as pd
+import re
+
+
+class Janitor:
+    def __init__(self):
+        pass
+
+    def clean_dataframe(self, df: pd.DataFrame) -> pd.DataFrame:
+        # Apply all cleaning procedure in sequence
+        df = df.copy()  # First make a copy to preserve integrity of the old df
+        df = self.drop_duplicates(df)
+        df = self.snake_case_columns(df)
+        df = self.fix_none(df)
+        df = self.fix_datatypes(df)
+        df = self.dropna_target(df)
+        df = df.reset_index(drop=True)  # Fix index
+        return df
+
+    def drop_duplicates(self, df):
+        return df.drop_duplicates() if df.duplicated().sum() > 0 else df
+
+    def snake_case_columns(self, df):
+        pattern = r'(?<!^)(?=[A-Z][a-z])|(?<=[a-z])(?=[A-Z])'
+        df.columns = [re.sub(pattern, '_', column).lower()
+                      for column in df.columns]
+        return df
+
+    def fix_none(self, df):
+        def replace_none(value):
+            like_nan = {'none', ''}
+            if pd.isnull(value) or (isinstance(value, str) and (value.lower().strip() in like_nan)):
+                value = pd.NA
+            return value
+
+        return df.map(replace_none)
+
+    def fix_datatypes(self, df):
+        columns_int = ['prg', 'pl', 'pr', 'sk', 'ts', 'age']
+        columns_float = ['m11', 'bd2']
+        col_to_fix = {col for col in columns_int+columns_float}
+        if col_to_fix.issubset(df.columns):
+            df[columns_int] = df[columns_int].astype(int)
+            df[columns_float] = df[columns_float].astype(float)
+        return df
+
+    # Drop rows with missing values in target column and reset index
+    def dropna_target(self, df):
+        return df.dropna(subset='sepsis') if 'sepsis' in df.columns else df

utils/navigation.py

@@ -0,0 +1,10 @@
+import streamlit as st
+
+
+async def navigation():
+    # Navigation
+    st.sidebar.page_link("app.py", label="Home", icon="🤖")
+    st.sidebar.page_link("pages/01_🕰️_History.py", label="History", icon="🕰️")
+
+    # Divider
+    st.sidebar.divider()