feat: add some graphs

bin/graph-inventory-not-in-buildstepoutputs.py

@@ -0,0 +1,178 @@
+#!/usr/bin/env python3
+import duckdb
+import matplotlib.pyplot as plt
+import matplotlib.dates as mdates
+import matplotlib.ticker as ticker
+import pandas as pd
+from pathlib import Path
+
+ORPHANS_FILE = "datasets/inventory-not-in-buildstepoutputs-2025-12-05-17:38:30Z.parquet"
+INVENTORY_FILE = "datasets/inventory-2026-01-06T01-13Z.parquet"
+OUTPUT_DIR = Path("graphs")
+
+# Pastel colors
+PASTEL_GREEN = "#77dd77"
+PASTEL_RED = "#ff6961"
+
+
+def plot_bar(ax, df, x_col, y_col, title, ylabel, color="steelblue", width=2.0, is_percentage=False):
+    ax.bar(df[x_col], df[y_col], width=width, color=color)
+    ax.set_title(title)
+    ax.set_xlabel("Date")
+    ax.set_ylabel(ylabel)
+    ax.grid(True, alpha=0.3)
+    ax.tick_params(axis="x", rotation=45)
+    if is_percentage:
+        ax.set_ylim(-5, 105)
+    else:
+        ax.set_ylim(-df[y_col].max() * 0.05, df[y_col].max() * 1.1)
+        ax.yaxis.set_major_formatter(ticker.FuncFormatter(lambda x, p: format(int(x), ",")))
+
+
+def main():
+    OUTPUT_DIR.mkdir(exist_ok=True)
+    con = duckdb.connect()
+
+    # === MONTHLY GRAPHS ===
+    fig, axes = plt.subplots(3, 1, figsize=(14, 18))
+
+    df = con.execute(
+        """
+        WITH orphans AS (
+            SELECT date_trunc('month', last_modified_date) as period, COUNT(*) as count
+            FROM read_parquet(?)
+            GROUP BY 1
+        ),
+        inventory AS (
+            SELECT date_trunc('month', last_modified_date) as period, COUNT(*) as count
+            FROM read_parquet(?)
+            WHERE key LIKE '%.narinfo'
+              AND last_modified_date <= TIMESTAMP '2025-12-05 17:38:30Z'
+            GROUP BY 1
+        )
+        SELECT i.period,
+               COALESCE(o.count, 0) as orphan_count,
+               i.count as total_count,
+               ROUND(COALESCE(o.count, 0) * 100.0 / i.count, 2) as percentage
+        FROM inventory i
+        LEFT JOIN orphans o ON i.period = o.period
+        ORDER BY i.period
+    """,
+        [ORPHANS_FILE, INVENTORY_FILE],
+    ).fetchdf()
+
+    plot_bar(axes[0], df, "period", "total_count", "Inventory Narinfo entries by Month", "Count", color=PASTEL_GREEN, width=25)
+    axes[0].xaxis.set_major_locator(mdates.YearLocator())
+    axes[0].xaxis.set_major_formatter(mdates.DateFormatter("%Y"))
+
+    plot_bar(axes[1], df, "period", "orphan_count", "Inventory Narinfo entries not in buildstepoutputs by Month", "Count", color=PASTEL_RED, width=25)
+    axes[1].xaxis.set_major_locator(mdates.YearLocator())
+    axes[1].xaxis.set_major_formatter(mdates.DateFormatter("%Y"))
+
+    plot_bar(axes[2], df, "period", "percentage", "Inventory Narinfo entries not in buildstepoutputs by Month (% of narinfos in inventory)", "Missing %", color=PASTEL_RED, width=25, is_percentage=True)
+    axes[2].xaxis.set_major_locator(mdates.YearLocator())
+    axes[2].xaxis.set_major_formatter(mdates.DateFormatter("%Y"))
+
+    plt.tight_layout()
+    plt.savefig(OUTPUT_DIR / "inventory-not-in-buildstepoutputs-monthly.png", dpi=150)
+    plt.close()
+
+    # === DAILY GRAPHS ===
+    fig, axes = plt.subplots(3, 1, figsize=(14, 18))
+
+    df = con.execute(
+        """
+        WITH orphans AS (
+            SELECT date_trunc('day', last_modified_date) as period, COUNT(*) as count
+            FROM read_parquet(?)
+            GROUP BY 1
+        ),
+        inventory AS (
+            SELECT date_trunc('day', last_modified_date) as period, COUNT(*) as count
+            FROM read_parquet(?)
+            WHERE key LIKE '%.narinfo'
+              AND last_modified_date <= TIMESTAMP '2025-12-05 17:38:30Z'
+            GROUP BY 1
+        )
+        SELECT i.period,
+               COALESCE(o.count, 0) as orphan_count,
+               i.count as total_count,
+               ROUND(COALESCE(o.count, 0) * 100.0 / i.count, 2) as percentage
+        FROM inventory i
+        LEFT JOIN orphans o ON i.period = o.period
+        ORDER BY i.period
+    """,
+        [ORPHANS_FILE, INVENTORY_FILE],
+    ).fetchdf()
+
+    plot_bar(axes[0], df, "period", "total_count", "Inventory Narinfo entries by Day", "Count", color=PASTEL_GREEN)
+    axes[0].xaxis.set_major_locator(mdates.YearLocator())
+    axes[0].xaxis.set_major_formatter(mdates.DateFormatter("%Y"))
+
+    plot_bar(axes[1], df, "period", "orphan_count", "Inventory Narinfo entries not in buildstepoutputs by Day (Count)", "Count", color=PASTEL_RED)
+    axes[1].xaxis.set_major_locator(mdates.YearLocator())
+    axes[1].xaxis.set_major_formatter(mdates.DateFormatter("%Y"))
+
+    plot_bar(axes[2], df, "period", "percentage", "Inventory Narinfo entries not in buildstepoutputs by Day (% of narinfos in inventory)", "Missing %", color=PASTEL_RED, is_percentage=True)
+    axes[2].xaxis.set_major_locator(mdates.YearLocator())
+    axes[2].xaxis.set_major_formatter(mdates.DateFormatter("%Y"))
+
+    plt.tight_layout()
+    plt.savefig(OUTPUT_DIR / "inventory-not-in-buildstepoutputs-daily.png", dpi=150)
+    plt.close()
+
+    # === YEARLY COMPARISON GRAPH ===
+    fig, ax1 = plt.subplots(figsize=(14, 8))
+
+    df = con.execute(
+        """
+        WITH orphans AS (
+            SELECT date_trunc('year', last_modified_date) as period, COUNT(*) as count
+            FROM read_parquet(?)
+            GROUP BY 1
+        ),
+        inventory AS (
+            SELECT date_trunc('year', last_modified_date) as period, COUNT(*) as count
+            FROM read_parquet(?)
+            WHERE key LIKE '%.narinfo'
+              AND last_modified_date <= TIMESTAMP '2025-12-05 17:38:30Z'
+            GROUP BY 1
+        )
+        SELECT i.period,
+               COALESCE(o.count, 0) as orphan_count,
+               i.count as total_count,
+               ROUND(COALESCE(o.count, 0) * 100.0 / i.count, 2) as percentage
+        FROM inventory i
+        LEFT JOIN orphans o ON i.period = o.period
+        ORDER BY i.period
+    """,
+        [ORPHANS_FILE, INVENTORY_FILE],
+    ).fetchdf()
+
+    # Primary axis: total narinfos (green bars)
+    bar_width = 100
+    ax1.bar(df["period"] - pd.Timedelta(days=60), df["total_count"], width=bar_width, color=PASTEL_GREEN, label="Narinfos in Inventory")
+    ax1.set_xlabel("Year")
+    ax1.set_ylabel("Total Narinfos")
+    ax1.tick_params(axis="x", rotation=45)
+    ax1.set_ylim(-df["total_count"].max() * 0.05, df["total_count"].max() * 1.1)
+    ax1.yaxis.set_major_formatter(ticker.FuncFormatter(lambda x, p: format(int(x), ",")))
+    ax1.xaxis.set_major_locator(mdates.YearLocator())
+    ax1.xaxis.set_major_formatter(mdates.DateFormatter("%Y"))
+    ax1.grid(True, alpha=0.3)
+
+    # Secondary axis: percentage missing (red bars)
+    ax2 = ax1.twinx()
+    ax2.bar(df["period"] + pd.Timedelta(days=60), df["percentage"], width=bar_width, color=PASTEL_RED, label="Narinfos not in buildstepoutputs (%)")
+    ax2.set_ylabel("Missing %")
+    ax2.set_ylim(-5, 105)
+
+    fig.suptitle("Yearly Comparison: Total Narinfos vs Missing Entries %")
+    fig.legend(loc="upper right", bbox_to_anchor=(0.9, 0.9))
+    plt.tight_layout()
+    plt.savefig(OUTPUT_DIR / "inventory-not-in-buildstepoutputs-yearly-comparison.png", dpi=150)
+    plt.close()
+
+
+if __name__ == "__main__":
+    main()

shell.nix

@@ -9,5 +9,10 @@ pkgs.mkShell {
   packages = [
     pkgs.git-lfs
     pkgs.git-xet
+    (pkgs.python313.withPackages (ps: [
+      ps.duckdb
+      ps.matplotlib
+      ps.pandas
+    ]))
   ];
 }