feat(cv): qualify templates and reliable local import processing

2026-10-28 14:36:12 +01:00
parent 73ea888375
commit 3f4db8ca81
69 changed files with 2597 additions and 180 deletions
@@ -0,0 +1,51 @@
+"""Separate DOM Range/PDF character-cell metrics from visible raster ink alignment.
+
+This leaves the original 2mm fidelity oracle and all its failures unchanged.
+"""
+import argparse
+import json
+import math
+import re
+
+from pathlib import Path
+from PIL import Image
+
+
+def main():
+    parser = argparse.ArgumentParser()
+    parser.add_argument("directory", type=Path)
+    args = parser.parse_args()
+    source = json.loads((args.directory / "measurements.json").read_text(encoding="utf-8"))
+    oracle = json.loads((args.directory / "fidelity-results.json").read_text(encoding="utf-8"))
+    results = []
+    for fixture in source["fixtures"]:
+        directory = args.directory / fixture["fixture"]
+        marker = next(item for item in fixture["measured"]["tokens"] if item["token"] == "S01")
+        rect = next(item for item in fixture["measured"]["nameLines"] if item["text"] == "S01")
+        foreground = tuple(map(int, re.findall(r"\d+", marker["color"])[:3]))
+        bounds = (math.floor(rect["x"]), math.floor(rect["y"]), math.ceil(rect["x"] + rect["width"]), math.ceil(rect["y"] + rect["height"] + 12))
+        boxes = {}
+        for kind, filename in (("preview", "preview-1.png"), ("pdf", "pdf-1.png")):
+            with Image.open(directory / filename) as image:
+                crop = image.convert("RGB").crop(bounds)
+                pixels = [(x, y) for y in range(crop.height) for x in range(crop.width)
+                          if max(abs(a - b) for a, b in zip(crop.getpixel((x, y)), foreground)) <= 50]
+                if not pixels:
+                    raise ValueError(f"No foreground ink for {fixture['fixture']} {kind}")
+                boxes[kind] = [min(x for x, y in pixels) + bounds[0], min(y for x, y in pixels) + bounds[1],
+                               max(x for x, y in pixels) + bounds[0], max(y for x, y in pixels) + bounds[1]]
+                crop.resize((crop.width * 4, crop.height * 4)).save(directory / f"S01-{kind}-ink.png")
+        original = next(item for item in oracle["fixtures"] if item["fixture"] == fixture["fixture"])
+        comparison = next(item for item in original["comparisons"] if item["token"] == "S01")
+        results.append({"fixture": fixture["fixture"], "domRange": rect, "foregroundRgb": foreground,
+                        "originalOracleRangeVsPdfCharacterCellDeltaMm": comparison["deltaYmm"],
+                        "visibleInkBoundsPx": boxes, "visibleInkTopDeltaPx": abs(boxes["preview"][1] - boxes["pdf"][1]),
+                        "visibleInkTopDeltaMm": abs(boxes["preview"][1] - boxes["pdf"][1]) * 25.4 / 96})
+    report = {"qualification": "Separate diagnostic measurements only. Pixels within 50 per RGB channel of the recorded foreground identify visible ink in the S01 ROI. Original fidelity tolerances/failures are unchanged.",
+              "fixtures": results, "maxVisibleInkTopDeltaPx": max(item["visibleInkTopDeltaPx"] for item in results)}
+    (args.directory / "name-box-metrics.json").write_text(json.dumps(report, ensure_ascii=False, indent=2), encoding="utf-8")
+    print(f"S01 raster-ink diagnostic: {len(results)} documents, max top delta {report['maxVisibleInkTopDeltaPx']}px; original oracle unchanged")
+
+
+if __name__ == "__main__":
+    main()