ocr-projects-archive / ocr-tech /tests /test_audit_limbu_book_line_split_suggestions.py
prosperious's picture
Add 1000 files (OCR_INDRA/.gitignore ...)
f9bc0c1 verified
Raw History Blame Contribute Delete
4.73 kB
from __future__ import annotations
import csv
import importlib.util
import json
import sys
from argparse import Namespace
from pathlib import Path
from PIL import Image, ImageDraw
from ocrtech.manifest import sha256_file
def _load_tool(name: str):
path = Path(__file__).resolve().parents[1] / "tools" / f"{name}.py"
spec = importlib.util.spec_from_file_location(name, path)
assert spec is not None
assert spec.loader is not None
module = importlib.util.module_from_spec(spec)
sys.modules[spec.name] = module
spec.loader.exec_module(module)
return module
def _write_packet(tmp_path: Path) -> Path:
page = Image.new("RGB", (220, 80), "white")
draw = ImageDraw.Draw(page)
draw.rectangle([20, 24, 70, 42], fill="black")
draw.rectangle([105, 24, 160, 42], fill="black")
source = tmp_path / "page.png"
page.save(source)
document_path = tmp_path / "document.json"
document_path.write_text(
json.dumps(
{
"schema_version": "1.0",
"source_path": str(source),
"pages": [{"page_index": 0, "text_lines": [], "blocks": []}],
"tables": [],
"figures": [],
"metadata": {},
}
)
+ "\n",
encoding="utf-8",
)
crop = tmp_path / "crop.png"
page.crop((10, 18, 180, 50)).save(crop)
packet = tmp_path / "packet.tsv"
with packet.open("w", encoding="utf-8", newline="") as handle:
fieldnames = [
"packet_row_id",
"review_item_type",
"book_page_id",
"book_line_id",
"book_document",
"bbox",
"text",
"crop_path",
"crop_sha256",
"review_status",
]
writer = csv.DictWriter(handle, fieldnames=fieldnames, delimiter="\t", lineterminator="\n")
writer.writeheader()
writer.writerow(
{
"packet_row_id": "existing-1",
"review_item_type": "existing_book_line",
"book_page_id": "p001",
"book_line_id": "line-1",
"book_document": str(document_path),
"bbox": "[10,18,170,32]",
"text": "wide noisy line",
"crop_path": str(crop),
"crop_sha256": sha256_file(crop),
"review_status": "pending",
}
)
return packet
def _write_suggestions(tmp_path: Path) -> Path:
prepare = _load_tool("prepare_limbu_book_line_split_suggestions")
report = prepare.prepare_suggestions(
Namespace(
review_packet=_write_packet(tmp_path),
out=tmp_path / "suggestions",
crop_padding=8,
col_min_ink=2,
max_col_gap=8,
min_segment_width=20,
min_segment_ink=12,
split_padding=1,
min_bbox_width=100.0,
min_aspect_ratio=2.0,
)
)
assert report["suggestion_count"] == 1
return Path(report["out"])
def test_audit_limbu_book_line_split_suggestions_passes_clean_artifact(tmp_path: Path) -> None:
audit_tool = _load_tool("audit_limbu_book_line_split_suggestions")
suggestions = _write_suggestions(tmp_path)
audit = audit_tool.audit_suggestions(suggestions)
assert audit["audit_passed"] is True
assert audit["operational_review_aid_safe"] is True
assert audit["suggestion_count"] == 1
assert audit["overlay_count"] == 1
assert audit["reason_counts"] == {"visual_column_projection_segments": 1}
assert (suggestions / "book-line-split-suggestions-audit.json").exists()
def test_audit_limbu_book_line_split_suggestions_rejects_tampered_crop_hash(tmp_path: Path) -> None:
audit_tool = _load_tool("audit_limbu_book_line_split_suggestions")
suggestions = _write_suggestions(tmp_path)
tsv = suggestions / "book-line-split-suggestions.tsv"
with tsv.open("r", encoding="utf-8", newline="") as handle:
reader = csv.DictReader(handle, delimiter="\t")
fieldnames = reader.fieldnames
rows = list(reader)
assert fieldnames is not None
rows[0]["crop_sha256"] = "0" * 64
with tsv.open("w", encoding="utf-8", newline="") as handle:
writer = csv.DictWriter(handle, fieldnames=fieldnames, delimiter="\t", lineterminator="\n")
writer.writeheader()
writer.writerows(rows)
audit = audit_tool.audit_suggestions(suggestions, tmp_path / "audit")
assert audit["audit_passed"] is False
assert any("suggestions TSV SHA-256 mismatch" in issue for issue in audit["issues"])
assert any("crop SHA-256 mismatch" in issue for issue in audit["issues"])