app.py

import os
import pandas as pd
import plotly.express as px
import streamlit as st

DATA_FILE = "data/anthology-2020-23_specter2_base.json"


def load_df(data_file: os.PathLike):
    df = pd.read_json(data_file, orient="records")
    df["x"] = df["point2d"].apply(lambda x: x[0])
    df["y"] = df["point2d"].apply(lambda x: x[1])
    if "publication_type" in df.columns:
        df["type"] = df["publication_type"]
        df = df.drop(columns=["point2d", "publication_type"])
    else:
        df = df.drop(columns=["point2d"])
    return df


@st.cache_data
def load_dataframe():
    return load_df(DATA_FILE)

DF = load_dataframe()

with st.sidebar:
    venues = st.multiselect(
        "Venues",
        ["ACL", "EMNLP", "NAACL", "TACL"],
        ["ACL", "EMNLP", "NAACL", "TACL"],
    )

    start_year, end_year = st.select_slider("Publication year", options=("2020", "2021", "2022", "2023"), value=("2020", "2023"))
    author_names = st.text_input('Author names (separated by comma)')

    start_year = int(start_year)
    end_year = int(end_year)
    df = DF[(DF["year"] >= start_year) & (DF["year"] <= end_year)]
    if len(venues) < 4:
        selected_venues = [v.lower() for v in venues]
        df = df[df["source"].isin(selected_venues)]

    if author_names:
        authors = [a.strip().lower() for a in author_names.split(",")]
        author_mask = df.authors.apply(lambda x: all(a in x for a in authors))
        df = df[author_mask]

    st.write(f"Number of points: {df.shape[0]}")



fig = px.scatter(
    df,
    x="x",
    y="y",
    color="cluster",
    width=1200,
    height=750,
    hover_data=["title", "authors", "year", "source", "type"],
    color_continuous_scale="fall",
)
fig.update_layout(
    # margin=dict(l=10, r=10, t=10, b=10),
    showlegend=False,
    font=dict(
        family="Times New Roman",
        size=30,
    ),
)
fig.update_xaxes(title="")
fig.update_yaxes(title="")

st.plotly_chart(fig, use_container_width=True)