import streamlit as st
import json
import pandas as pd
import plotly.express as px
import seaborn as sns
import matplotlib.pyplot as plt
import streamlit.components.v1 as components
# Global variable to hold selected row index
selected_row_index = None
# Initialize an empty DataFrame
filtered_data = pd.DataFrame()
# Function to load JSONL file into a DataFrame
def load_jsonl(file_path):
data = []
with open(file_path, 'r') as f:
for line in f:
data.append(json.loads(line))
return pd.DataFrame(data)
# Function to filter DataFrame by keyword
def filter_by_keyword(df, keyword):
return df[df.apply(lambda row: row.astype(str).str.contains(keyword).any(), axis=1)]
# Streamlit App
st.title("Medical Licensing Exam Explorer with Speech Synthesis, Plotly and Seaborn π")
# Dropdown for file selection
file_option = st.selectbox("Select file:", ["small_file.jsonl", "large_file.jsonl"])
st.write(f"You selected: {file_option}")
# Load the data
small_data = load_jsonl("usmle_16.2MB.jsonl")
large_data = load_jsonl("usmle_2.08MB.jsonl")
# Show filtered data grid
if file_option == "small_file.jsonl":
data = small_data
else:
data = large_data
# Text input for search keyword
search_keyword = st.text_input("Enter a keyword to filter data (e.g., Heart, Lung, Pain, Memory):")
# Button to trigger search
if st.button("Search"):
filtered_data = filter_by_keyword(data, search_keyword)
st.write(f"Filtered Dataset by '{search_keyword}'")
selected_data = st.dataframe(filtered_data)
def generate_html_with_textarea(text_to_speak):
return f'''
Read It Aloud
π Read It Aloud
'''
# Define your text passage
text_passage = "A 60-year-old man is brought to the emergency department by police officers because he was acting strangely in public. The patient was found talking nonsensically to characters on cereal boxes in the store. Past medical history is significant for multiple hospitalizations for alcohol-related injuries and seizures. The patientβs vital signs are within normal limits. Physical examination shows a disheveled male who is oriented to person, but not time or place. Neurologic examination shows nystagmus and severe gait ataxia. A T1/T2 MRI is performed and demonstrates evidence of damage to the mammillary bodies. The patient is given the appropriate treatment for recovering most of his cognitive functions. However, significant short-term memory deficits persist. The patient remembers events from his past such as the school and college he attended, his current job, and the names of family members quite well. Which of the following is the most likely diagnosis in this patient?"
# Generate HTML code
documentHTML5 = generate_html_with_textarea(text_passage)
# Button to read all filtered rows
if st.button("Read All Rows"):
if not filtered_data.empty:
html_blocks = []
for idx, row in filtered_data.iterrows():
question_text = row.get("question", "No question field")
documentHTML5 = generate_html(question_text, "", idx)
html_blocks.append(documentHTML5)
all_html = ''.join(html_blocks)
components.html(all_html, width=1280, height=1024)
else:
st.warning("No rows to read.")
# Insert the HTML into Streamlit
# Button to read all filtered rows
if st.button("Read Aloud Text"):
components.html(documentHTML5, width=1280, height=1024)
# Plotly and Seaborn charts for EDA
if st.button("Generate Charts"):
st.subheader("Plotly Charts π")
# 1. Scatter Plot
fig = px.scatter(data, x=data.columns[0], y=data.columns[1])
st.plotly_chart(fig)
# 2. Line Plot
fig = px.line(data, x=data.columns[0], y=data.columns[1])
st.plotly_chart(fig)
# 3. Bar Plot
fig = px.bar(data, x=data.columns[0], y=data.columns[1])
st.plotly_chart(fig)
# 4. Histogram
fig = px.histogram(data, x=data.columns[0])
st.plotly_chart(fig)
# 5. Box Plot
fig = px.box(data, x=data.columns[0], y=data.columns[1])
st.plotly_chart(fig)
st.subheader("Seaborn Charts π")
# 6. Violin Plot
fig, ax = plt.subplots()
sns.violinplot(x=data.columns[0], y=data.columns[1], data=data)
st.pyplot(fig)
# 7. Swarm Plot
fig, ax = plt.subplots()
sns.swarmplot(x=data.columns[0], y=data.columns[1], data=data)
st.pyplot(fig)
# 8. Pair Plot
fig = sns.pairplot(data)
st.pyplot(fig)
# 9. Heatmap
fig, ax = plt.subplots()
sns.heatmap(data.corr(), annot=True)
st.pyplot(fig)
# 10. Regplot (Regression Plot)
fig, ax = plt.subplots()
sns.regplot(x=data.columns[0], y=data.columns[1], data=data)
st.pyplot(fig)