|
import streamlit as st |
|
import pandas as pd |
|
|
|
import plotly.graph_objects as go |
|
import plotly.express as px |
|
|
|
from ast import literal_eval |
|
from lxml import etree as ET |
|
|
|
|
|
def prepare_data(): |
|
data = pd.read_csv( |
|
"data/2002829_mapped_roles.csv", index_col=0, |
|
converters={"frame": literal_eval, "changed_roles": literal_eval, "unchanged_roles": literal_eval, "roles": literal_eval} |
|
) |
|
|
|
frame_to_scenario, frame_to_super_scenario = load_kicktionary_info() |
|
|
|
|
|
data_ = ( |
|
data |
|
.assign(sentence_idx=data["frame"].apply(lambda frame: frame[0])) |
|
.assign(frame_idx=data["frame"].apply(lambda frame: frame[1])) |
|
.assign(frame_name=data["frame"].apply(lambda frame: frame[2])) |
|
.assign(frame_scenario=data["frame"].apply(lambda frame: frame_to_scenario[frame[2]])) |
|
.assign(frame_super_scenario=data["frame"].apply(lambda frame: frame_to_super_scenario[frame[2]])) |
|
.assign(frame_target=data["frame"].apply(lambda frame: frame[3])) |
|
.drop(columns=["frame"]) |
|
)[["sentence_idx", "frame_idx", "frame_name", "frame_scenario", "frame_super_scenario", "frame_target", "changed_roles", "roles"]] |
|
|
|
|
|
max_sent = max(data_["sentence_idx"]) |
|
max_frame_per_sent = data_.groupby("sentence_idx").agg({"frame_idx": max}).reset_index() |
|
sent_to_max_frame = dict(zip(max_frame_per_sent["sentence_idx"], max_frame_per_sent["frame_idx"])) |
|
data_with_time = data_.assign( |
|
time_point= (data_ |
|
.apply(lambda row: (row["sentence_idx"] + row["frame_idx"] / (sent_to_max_frame[row["sentence_idx"]])) / (max_sent + 1), axis=1) |
|
) |
|
) |
|
data_with_first_roles = data_with_time.assign( |
|
first_role = data_with_time["changed_roles"].apply(lambda roles: roles[0] if len(roles) > 0 else None) |
|
) |
|
return data_with_first_roles |
|
|
|
|
|
def load_kicktionary_info(): |
|
|
|
kicktionary = ET.parse("kicktionary_lu_info.xml") |
|
frame_to_scenario = { |
|
lu.attrib["frame"]: lu.attrib["scenario"] |
|
for lu in kicktionary.xpath(".//LEXICAL-UNIT") if lu.attrib["frame"] |
|
} |
|
frame_to_super_scenario = { |
|
lu.attrib["frame"]: lu.attrib["super-scenario"] |
|
for lu in kicktionary.xpath(".//LEXICAL-UNIT") if lu.attrib["frame"] |
|
} |
|
|
|
return frame_to_scenario, frame_to_super_scenario |
|
|
|
|
|
def explore_timeline(): |
|
|
|
with st.container(): |
|
|
|
st.title("Football Perspective Chains") |
|
|
|
frame_label_map = { |
|
"frame_name": "frames", |
|
"frame_scenario": "scenarios (groups of related frames)", |
|
"frame_super_scenario": "super scenarios (groups of related scenarios)" |
|
} |
|
frame_column = st.selectbox( |
|
label="Display frames as: ", |
|
options=("frame_name", "frame_scenario", "frame_super_scenario"), |
|
format_func=lambda label: frame_label_map[label] |
|
) |
|
|
|
st.header("Timeline") |
|
data = prepare_data() |
|
time_scatter = ( |
|
data |
|
.dropna(axis=0, subset=["first_role"]) |
|
.plot.scatter( |
|
x="first_role", y="time_point", backend="plotly", color=frame_column |
|
) |
|
) |
|
time_scatter.update_traces(marker_size=20) |
|
time_scatter.update_layout(height=1000) |
|
st.plotly_chart(time_scatter) |
|
|
|
st.header("Overall focus") |
|
focus_bar = data.dropna(axis=0, subset=["first_role"])["first_role"].value_counts().plot.bar(y="first_role", backend="plotly") |
|
st.plotly_chart(focus_bar) |
|
|
|
st.header("Focus by frame") |
|
|
|
for team in ["Man. United", "Rangers"]: |
|
st.subheader(team) |
|
frame_bar = data.dropna(axis=0, subset=["first_role"])["frame_scenario"].value_counts().plot.bar(y=frame_column, backend="plotly") |
|
st.plotly_chart(frame_bar) |
|
|
|
|
|
|
|
|
|
if __name__ == "__main__": |
|
explore_timeline() |
|
|
