app.py

import plotly.express as px
import pandas as pd
import json
import plotly.graph_objects as go
from datasets import load_dataset
import streamlit as st

REPO_ID = "libeIO/Sciences-POC"
SAVE_PATH = "save"

with open('mapping_prompts.txt', 'r') as f:
    mapping = json.loads(f.read())

with open('mapping_noms.txt', 'r') as f:
    mapping_noms = json.loads(f.read())

if 'name' not in st.session_state.keys():
    st.session_state['name'] = 'Inconnus 1'
    
@st.cache_resource
def initialize(name):

    articles = pd.read_csv('extract_sciences_po.csv')

    with open(f"{SAVE_PATH}/{mapping[mapping_noms[name]]['save_path']}", 'r') as f : 
        out_dict = json.loads(f.read())
    
    df = pd.DataFrame.from_dict(out_dict)

    articles = pd.merge(df, articles, on='item_id', how='left')

    count_principale = df.groupby('categorie_principale').item_id.count()
    df['categorie_secondaire'] = df.apply(lambda x : x.categorie_secondaire.split(',')[0], axis=1)
    count_secondaire = df.groupby('categorie_secondaire').item_id.count()
    display_principale = count_principale.reset_index()
    display_principale.columns = ['Catégorie', 'Nombre d\'articles']
    display_secondaire = count_secondaire.reset_index()
    display_secondaire.columns = ['Catégorie', 'Nombre d\'articles']

    template ="ggplot2"

    fig = go.Figure()
        
    fig.update_layout(template=template, 
                      )

    fig.add_trace(go.Scatterpolar(
        r=display_principale['Nombre d\'articles'],
        theta=display_principale['Catégorie'],
        fill='toself',
        name='Catégorie Principale',
        marker = {'color' : 'red'},

    ))
    fig.add_trace(go.Scatterpolar(
        r=display_secondaire['Nombre d\'articles'],
        theta=display_secondaire['Catégorie'],
        fill='toself',
        name='Catégorie Secondaire',
        marker = {'color' : 'blue'},
        opacity=0.25,
    ))

    fig.update_layout(
    polar=dict(
        radialaxis=dict(
        visible=True,
        range=[0, max(max(display_principale['Nombre d\'articles']), max(display_secondaire['Nombre d\'articles']))]
        )),
    showlegend=True
    )
    fig.update_layout(legend=dict(
    yanchor="top",
    y=0.0001,
    xanchor="left",
    x=0.395
))
    
    path_prompt = mapping[mapping_noms[name]]['path_prompt']
    model = mapping[mapping_noms[name]]['client']
    with open(path_prompt, 'r') as f : 
        prompt = f.read()

    return fig, display_principale, articles, prompt, model


def display_article(article):

    url = article['url']

    colImage, colText = st.columns(2)
# try :   
    with colImage :
        st.image(article["image_url"]) # image URL
    with colText:
        if 'subhead' in article.index and article['subhead']!='nan':
            st.subheader(f":red[{article['subhead']}] [{article['titre'].rstrip('Libération').rstrip('-')[:-2]}]({url})") # Title 
        else : 
            # st.toast(article.index)
            titre_cleaned = article['titre'].removesuffix('Libération').rstrip('-').strip()
            st.subheader(f"[{titre_cleaned}]({url})") # Title 
        st.write(f"{article['description']}") # Header
        formatted_date = article["date_published"]
        if article.premium:
            st.markdown(
        f"""
        <span style='color:grey'>{formatted_date+" "} </span>  <span style='color:#eeb54e'> abonnés</span>
    """,
    unsafe_allow_html=True
    )
        else :
            st.markdown(
        f"""
        <span style='color:grey'>{formatted_date+" "} </span>
    """,
    unsafe_allow_html=True
    )
            
        st.badge(f"Catégories secondaires : {article['categorie_secondaire']}", icon=":material/info:", color="blue")
    # except :
    #     st.toast(f'Error displaying article {article.item_id}')
    #     return 
        

fig, display_principale, articles, prompt, model = initialize(st.session_state['name'])
# col1, col2, col3 = st.columns([0.5, 0.2, 0.3])

st.selectbox("Choisir groupe", [mapping[k]['auteurs'] for k in mapping.keys()], key='name')

if st.session_state['name']!=dimanov_et_al:
    with st.expander(f"Prompt for model : {model}") : 
        st.markdown(prompt)

    st.subheader('Répartition des articles par catégorie')
    # with col1:
    col1, col2 = st.columns([0.6, 0.4], vertical_alignment='center')

    with col1:
        st.plotly_chart(fig)

    with col2:
        st.dataframe(display_principale.set_index('Catégorie').sort_values(by='Nombre d\'articles', ascending=False))

    st.subheader('Exemples d\'articles')
    tabs = st.tabs(display_principale['Catégorie'].values.tolist())

    for i in range(len(tabs)):
        with tabs[i]:
            cat = display_principale['Catégorie'][i]
            for i, article in articles.loc[articles.categorie_principale==cat].sample(20, replace=True).drop_duplicates().iterrows():
                display_article(article)

else : 
    st.toast('Pas de prompt renseigné !')