refactor: sort scripts/ by who runs it

Top level is now only what the runbook or a re-run touches: validate_persona_axes, bounded_thinking_judge, template_catalog, export_selections, parse_stage_a, run_axis, export_steering_selection. Corpus ingestion and publishing moved to scripts/corpus/, plotting and stats to scripts/report/. Moved files needed parents[1] -> parents[2]; the two corpus scripts that import template_catalog use the sys.path shim bounded_thinking_judge_liveproof already used. Also completes the export_steering_selection rename: an earlier git reset had dropped the staged deletion, leaving both filenames tracked. Co-Authored-By: Claudypoo <288921227+claudypoo@users.noreply.github.com>
2026-08-11 11:23:11 +08:00 · 2026-07-25 10:23:22 +08:00
parent 3593189142
commit b8e22f42fb
14 changed files with 10 additions and 182 deletions
@@ -0,0 +1,98 @@
+from __future__ import annotations
+
+import html
+from pathlib import Path
+import textwrap
+from typing import Any
+
+import plotly.graph_objects as go
+
+import docs_results
+
+MAIN_PNG = docs_results.ROOT / "docs/results/assets/on_off_axis.png"
+MAIN_SVG = docs_results.ROOT / "docs/results/assets/on_off_axis.svg"
+
+
+def _wrap_hover(text: str, width: int = 62) -> str:
+    text = docs_results.display_template_text(text)
+    escaped = html.escape(" ".join(text.split()))
+    return "<br>".join(
+        textwrap.wrap(escaped, width=width, break_long_words=True, break_on_hyphens=False))
+
+
+def main_plot_rows(path: Path = docs_results.NORMAL_TEMPLATE_PAIR_STATS) -> list[dict[str, Any]]:
+    return docs_results.mean_template_rows(docs_results.read_jsonl(path))
+
+
+def template_scatter(rows: list[dict[str, Any]] | None = None, width: int | None = None) -> go.Figure:
+    rows = main_plot_rows() if rows is None else rows
+    top_rank = {row["template"]: i for i, row in enumerate(rows[:10], start=1)}
+    text = [str(top_rank[row["template"]]) if row["template"] in top_rank else "" for row in rows]
+    hover = [
+        "<br>".join([
+            f"<b>{_wrap_hover(row['template'])}</b>",
+            f"rank: {i}",
+            f"score t: {row['score_t']:.2f}",
+            f"score mean: {row['score_mean']:.2f}",
+            f"axis delta: {row['axis_delta']:.2f}",
+            f"off-axis problem: {row['off_axis_problem']:.2f}",
+            f"judge std: {row['judge_std']:.2f}",
+            f"cells: {row['n_cells']}",
+        ])
+        for i, row in enumerate(rows, start=1)
+    ]
+    fig = go.Figure(
+        data=go.Scatter(
+            x=[row["on_axis"] for row in rows],
+            y=[row["off_axis"] for row in rows],
+            mode="markers+text",
+            text=text,
+            textposition="middle center",
+            textfont={"size": 9, "color": "white"},
+            customdata=hover,
+            hovertemplate="%{customdata}<extra></extra>",
+            marker={
+                "size": 10,
+                "color": [row["score_t"] for row in rows],
+                "colorscale": "Cividis",
+                "showscale": True,
+                "colorbar": {"title": "score t"},
+                "line": {"width": 0.5, "color": "white"},
+                "opacity": 0.9,
+            },
+        )
+    )
+    fig.update_layout(
+        autosize=True,
+        width=width,
+        height=620,
+        template="plotly_white",
+        margin={"l": 68, "r": 24, "t": 28, "b": 66},
+        xaxis={
+            "title": "on-axis movement, higher is better",
+            "range": [-0.02, 1.02],
+            "gridcolor": "rgba(0,0,0,0.08)",
+        },
+        yaxis={
+            "title": "off-axis confounding, lower is better",
+            "range": [-0.02, 1.02],
+            "gridcolor": "rgba(0,0,0,0.08)",
+        },
+        annotations=[{
+            "text": "normal pilot scenarios; one point per measured template",
+            "xref": "paper",
+            "yref": "paper",
+            "x": 1.0,
+            "y": -0.13,
+            "showarrow": False,
+            "font": {"size": 11, "color": "rgba(0,0,0,0.62)"},
+        }],
+    )
+    return fig
+
+
+def write_main_plot_assets() -> None:
+    fig = template_scatter(width=960)
+    MAIN_PNG.parent.mkdir(parents=True, exist_ok=True)
+    fig.write_image(MAIN_PNG, width=960, height=620, scale=2)
+    fig.write_image(MAIN_SVG, width=960, height=620)