Update app.py

app.py

@@ -1,44 +1,62 @@
+# AUTOGENERATED! DO NOT EDIT! File to edit: app.ipynb.
+
+# %% auto 0
+__all__ = ['columns_to_click', 'title', 'description', 'dtypes', 'get_data']
+
+# %% app.ipynb 0
 import gradio as gr
 import pandas as pd
 
-# Assuming columns_to_click are the columns which contain URLs that you want to make clickable
-columns_to_click = ["Paper / Repo", "Playground"]
+
+# %% app.ipynb 1
+columns_to_click = ["Paper / Repo", "Selected \nplaygrounds"]
 
 def get_data():
-    # Load the CSV file into a DataFrame
     df = pd.read_csv(
         "https://docs.google.com/spreadsheets/d/e/2PACX-1vSC40sszorOjHfozmNqJT9lFiJhG94u3fbr3Ss_7fzcU3xqqJQuW1Ie_SNcWEB-uIsBi9NBUK7-ddet/pub?output=csv",
         skiprows=1,
     )
 
-    # Drop rows where the 'Model' column is NaN
-    df.dropna(subset=['Model'], inplace=True)
 
-    # Drop rows where the 'Parameters \n(B)' column is 'TBA'
-    df = df[df["Parameters \n(B)"] != "TBA"]
+    # %% app.ipynb 2
+    # Drop footers
+    df = df.copy()[~df["Model"].isna()]
+
+
+    # %% app.ipynb 3
+    # Drop TBA models
+    df = df.copy()[df["Parameters \n(B)"] != "TBA"]
+
 
-    # Apply make_clickable_cell to the specified columns
+    # %% app.ipynb 6
+    def make_clickable_cell(cell):
+        if pd.isnull(cell):
+            return ""
+        else:
+            return f'<a  target="_blank" href="{cell}">{cell}</a>'
+
+
+    # %% app.ipynb 7
     for col in columns_to_click:
         df[col] = df[col].apply(make_clickable_cell)
 
     return df
 
-def make_clickable_cell(cell):
-    if pd.isnull(cell) or not isinstance(cell, str):
-        return ""
-    else:
-        return f'<a target="_blank" href="{cell}">{cell}</a>'
 
-# Load the data to get the columns for setting up datatype
-dataframe = get_data()
-dtypes = ["str" if c not in columns_to_click else "html" for c in dataframe.columns]
+# %% app.ipynb 2
+title = """<h1 align="center">The Large Language Models Landscape</h1>"""
+description = """Large Language Models (LLMs) today come in a variety architectures and capabilities. This interactive landscape provides a visual overview of the most important LLMs, including their training data, size, release date, and whether they are openly accessible or not. It also includes notes on each model to provide additional context. This landscape is derived from data compiled by Dr. Alan D. Thompson at [lifearchitect.ai/models](https://lifearchitect.ai/models/).
+"""
+
+
+# %% app.ipynb 3
+dtypes = ["str" if c not in columns_to_click else "markdown" for c in get_data().columns]
+
 
-# Gradio app setup
+# %% app.ipynb 4
 with gr.Blocks() as demo:
-    # Markdown and DataFrame components
     gr.Markdown(title)
     gr.Markdown(description)
     gr.DataFrame(get_data, datatype=dtypes, every=60)
 
-# Launch the Gradio app
-demo.launch()
+demo.queue().launch()