From 262ddd3092ce42d2d17c02b20ac0f52462de109c Mon Sep 17 00:00:00 2001
From: James Bourbeau <jrbourbeau@gmail.com>
Date: Fri, 1 Mar 2024 15:28:32 -0600
Subject: [PATCH] Add basic streamlit app

---
 dashboard.py | 43 +++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 43 insertions(+)
 create mode 100644 dashboard.py

diff --git a/dashboard.py b/dashboard.py
new file mode 100644
index 0000000..503a143
--- /dev/null
+++ b/dashboard.py
@@ -0,0 +1,43 @@
+import dask_deltatable as ddt
+import streamlit as st
+
+from preprocess import OUTDIR
+
+
+def get_data(variable):
+    df = ddt.read_deltalake(OUTDIR / "dask" / variable).compute()
+    df = df.drop(columns="__index_level_0__")
+    df["repo"] = "https://github.com/" + df["repo"]
+    df["username"] = "https://github.com/" + df["username"]
+    # Streamlit doesn't like pyarrow strings
+    # (xref https://github.com/streamlit/streamlit/issues/6334)
+    str_cols = df.select_dtypes(include="string").columns
+    df = df.astype({c: object for c in str_cols})
+    return df
+
+
+st.markdown("""
+## Dask mentions on GitHub
+### Commits
+""")
+df = get_data("commits")
+st.dataframe(
+    df,
+    column_config={
+        "repo": st.column_config.LinkColumn("repo"),
+        "username": st.column_config.LinkColumn("user"),
+    },
+    hide_index=True,
+)
+st.markdown("""
+### Comments
+""")
+df = get_data("comments")
+st.dataframe(
+    df,
+    column_config={
+        "repo": st.column_config.LinkColumn("repo"),
+        "username": st.column_config.LinkColumn("user"),
+    },
+    hide_index=True,
+)