coconut / pages /4 Sunburst.py
faizhalas's picture
Update pages/4 Sunburst.py
19b9cbc verified
#===import module===
import streamlit as st
import pandas as pd
import plotly.express as px
import numpy as np
import sys
#===config===
st.set_page_config(
page_title="Coconut",
page_icon="πŸ₯₯",
layout="wide",
initial_sidebar_state="collapsed"
)
hide_streamlit_style = """
<style>
#MainMenu
{visibility: hidden;}
footer {visibility: hidden;}
[data-testid="collapsedControl"] {display: none}
</style>
"""
st.markdown(hide_streamlit_style, unsafe_allow_html=True)
with st.popover("πŸ”— Menu"):
st.page_link("https://www.coconut-libtool.com/", label="Home", icon="🏠")
st.page_link("pages/1 Scattertext.py", label="Scattertext", icon="1️⃣")
st.page_link("pages/2 Topic Modeling.py", label="Topic Modeling", icon="2️⃣")
st.page_link("pages/3 Bidirected Network.py", label="Bidirected Network", icon="3️⃣")
st.page_link("pages/4 Sunburst.py", label="Sunburst", icon="4️⃣")
st.page_link("pages/5 Burst Detection.py", label="Burst Detection", icon="5️⃣")
st.page_link("pages/6 Keywords Stem.py", label="Keywords Stem", icon="6️⃣")
st.header("Sunburst Visualization", anchor=False)
st.subheader('Put your file here...', anchor=False)
#===clear cache===
def reset_all():
st.cache_data.clear()
#===check type===
@st.cache_data(ttl=3600)
def get_ext(extype):
extype = uploaded_file.name
return extype
@st.cache_data(ttl=3600)
def upload(extype):
papers = pd.read_csv(uploaded_file)
#lens.org
if 'Publication Year' in papers.columns:
papers.rename(columns={'Publication Year': 'Year', 'Citing Works Count': 'Cited by',
'Publication Type': 'Document Type', 'Source Title': 'Source title'}, inplace=True)
return papers
@st.cache_data(ttl=3600)
def conv_txt(extype):
col_dict = {'TI': 'Title',
'SO': 'Source title',
'DT': 'Document Type',
'DE': 'Author Keywords',
'ID': 'Keywords Plus',
'AB': 'Abstract',
'TC': 'Cited by',
'PY': 'Year',}
papers = pd.read_csv(uploaded_file, sep='\t', lineterminator='\r')
papers.rename(columns=col_dict, inplace=True)
return papers
#===Read data===
uploaded_file = st.file_uploader('', type=['csv', 'txt'], on_change=reset_all)
if uploaded_file is not None:
try:
extype = get_ext(uploaded_file)
if extype.endswith('.csv'):
papers = upload(extype)
elif extype.endswith('.txt'):
papers = conv_txt(extype)
@st.cache_data(ttl=3600)
def get_minmax(extype):
extype = extype
MIN = int(papers['Year'].min())
MAX = int(papers['Year'].max())
GAP = MAX - MIN
return papers, MIN, MAX, GAP
tab1, tab2 = st.tabs(["πŸ“ˆ Generate visualization", "πŸ““ Recommended Reading"])
with tab1:
#===sunburst===
try:
papers, MIN, MAX, GAP = get_minmax(extype)
except KeyError:
st.error('Error: Please check again your columns.')
sys.exit(1)
if (GAP != 0):
YEAR = st.slider('Year', min_value=MIN, max_value=MAX, value=(MIN, MAX), on_change=reset_all)
else:
st.write('You only have data in ', (MAX))
YEAR = (MIN, MAX)
@st.cache_data(ttl=3600)
def listyear(extype):
global papers
years = list(range(YEAR[0],YEAR[1]+1))
papers = papers.loc[papers['Year'].isin(years)]
return years, papers
@st.cache_data(ttl=3600)
def vis_sunbrust(extype):
papers['Cited by'] = papers['Cited by'].fillna(0)
vis = pd.DataFrame()
vis[['doctype','source','citby','year']] = papers[['Document Type','Source title','Cited by','Year']]
viz=vis.groupby(['doctype', 'source', 'year'])['citby'].agg(['sum','count']).reset_index()
viz.rename(columns={'sum': 'cited by', 'count': 'total docs'}, inplace=True)
fig = px.sunburst(viz, path=['doctype', 'source', 'year'], values='total docs',
color='cited by',
color_continuous_scale='RdBu',
color_continuous_midpoint=np.average(viz['cited by'], weights=viz['total docs']))
fig.update_layout(height=800, width=1200)
return fig
years, papers = listyear(extype)
if {'Document Type','Source title','Cited by','Year'}.issubset(papers.columns):
fig = vis_sunbrust(extype)
st.plotly_chart(fig, height=800, width=1200) #use_container_width=True)
else:
st.error('We require these columns: Document Type, Source title, Cited by, Year', icon="🚨")
with tab2:
st.markdown('**numpy.average β€” NumPy v1.24 Manual. (n.d.). Numpy.Average β€” NumPy v1.24 Manual.** https://numpy.org/doc/stable/reference/generated/numpy.average.html')
st.markdown('**Sunburst. (n.d.). Sunburst Charts in Python.** https://plotly.com/python/sunburst-charts/')
except:
st.error("Please ensure that your file is correct. Please contact us if you find that this is an error.", icon="🚨")
st.stop()