Adds basic streamlit app

app.py

@@ -0,0 +1,76 @@
+"""Streamlit app"""
+from os.path import join, exists, dirname, abspath
+from glob import glob
+
+import numpy as np
+import pandas as pd
+import torch
+import torchvision
+import streamlit as st
+
+import warnings
+warnings.simplefilter(action='ignore')
+
+curr_filepath = abspath(__file__)
+repo_path = dirname(curr_filepath)
+
+
+temporal_terms = [
+    "then",
+    "before",
+    "after",
+    "followed by",
+    "preceded by",
+]
+
+
+def make_grid(cols,rows):
+    grid = [0]*cols
+    for i in range(cols):
+        with st.container():
+            grid[i] = st.columns(rows)
+    return grid
+
+
+if __name__ == "__main__":
+
+    # Streamlit app code
+    st.set_page_config(layout="wide")
+    st.title("Clips from AudioCaps (possibly of temporal nature) 🎬")
+    
+    # load data
+    if "df" not in st.session_state:
+        splits = ["train.csv", "val.csv", "test.csv"]
+        dfs = [pd.read_csv(join(repo_path, "data", split)) for split in splits]
+        df = pd.concat(dfs, axis=0)
+        # Filter df based on whether the temporal term is in the sentence
+        indices = df.caption.apply(lambda x: any([term in x for term in temporal_terms]))
+        df = df[indices]
+        st.session_state.df = df
+    else:
+        df = st.session_state.df
+    st.markdown(f"**Total number of relevant clips**: {len(df)}", unsafe_allow_html=True)
+
+    reload_button = st.button("Reload")
+    NUM = 9
+    indices = np.random.randint(0, len(st.session_state.df), NUM)
+    if reload_button:
+        indices = np.random.randint(0, len(st.session_state.df), NUM)
+
+    grid = make_grid(3, 3)
+    per_video_width = 360
+    per_video_height = 240
+    for i, idx in enumerate(indices):
+        row = i // 3
+        col = i % 3
+        
+        video_id = df.iloc[idx].youtube_id
+        start = df.iloc[idx].start_time
+        end = start + 10.
+
+        url = f"https://www.youtube.com/embed/{video_id}?start={int(start)}&end={int(end)}"
+        html_code = f"""
+        <iframe height="{per_video_height}" width="{per_video_width}" src="{url}" frameborder="0" allowfullscreen></iframe>
+        """
+        grid[row][col].markdown(html_code, unsafe_allow_html=True)
+        grid[row][col].markdown(f"**Caption**: {df.iloc[idx].caption}", unsafe_allow_html=True)