Improve charting for admin

tabby/admin/pages/Metrics.py

@@ -16,7 +16,7 @@ def query_data():
     df = conn.sql(
         """
     SELECT
-        date_trunc('hour', to_timestamp(CAST(created AS int64))) AS Date,
+        date_trunc('day', to_timestamp(CAST(created AS int64))) AS Date,
         SUM(IF(view, 1, 0)) as "Views",
         SUM(IF("select", 1, 0)) as "Acceptances"
     FROM completion_events
@@ -30,12 +30,36 @@ def query_data():
 
 df = query_data()
 
-if len(df) > 0:
+
+def plot_summary():
+    sum_views = int(sum(df.Views))
+    sum_acceptances = int(sum(df.Acceptances))
+    ratio = (sum_acceptances / sum_views) * 100
+
+    col1, col2, col3 = st.columns(3)
+
+    with col1:
+        st.metric("Views", sum_views)
+    with col2:
+        st.metric("Acceptances", sum_acceptances)
+    with col3:
+        st.metric("Accept Ratio", f"{round(ratio)} %")
+
+
+plot_summary()
+st.write("---")
+
+
+def plot_charts():
     st.markdown("### Completion Events")
     st.line_chart(df, x="Date")
 
     st.markdown("### Acceptance Rate")
     df["Acceptance Rate"] = df["Acceptances"] / df["Views"]
     st.line_chart(df, x="Date", y="Acceptance Rate")
+
+
+if len(df) > 0:
+    plot_charts()
 else:
     st.markdown("No data available")

tabby/admin/pages/Projects.py

@@ -1,6 +1,7 @@
 import os
 from pathlib import Path
 
+import pandas as pd
 import streamlit as st
 import toml
 from datasets import load_from_disk
@@ -12,27 +13,42 @@ dataset_dir = os.environ.get("DATASET_DIR", None)
 git_repositories_dir = os.environ.get("GIT_REPOSITORIES_DIR", None)
 config_file = os.environ.get("CONFIG_FILE", None)
 config = toml.load(config_file)
+projects = config.get("projects", {})
+
+
+def count_by_language(dataset):
+    key = "language"
+    df = (
+        pd.DataFrame(dataset[key], columns=[key])
+        .groupby([key])
+        .size()
+        .to_frame("# Files")
+    )
+    return df
 
 
 def dataset_info():
-    st.subheader("Dataset")
     if not Path(dataset_dir).is_dir():
-        st.write("*Not populated*")
+        st.write("*n/a*")
         return
 
-    info = load_from_disk(dataset_dir)
+    dataset = load_from_disk(dataset_dir)
-    st.write("Source files: ", len(info))
+
+    col1, col2 = st.columns(2)
+    with col1:
+        st.metric("Total files", len(dataset))
+
+    with col2:
+        st.bar_chart(count_by_language(dataset))
 
 
 def project_list():
-    data = config.get("projects", {})
+    if len(projects) <= 0:
-
-    if len(data) <= 0:
         st.write("Your project list is empty")
         st.write(f"Edit `{config_file}` to add projects")
         return
 
-    for k, v in data.items():
+    for k, v in projects.items():
         st.subheader(k)
         st.write(f'Git: {v["git_url"]}')