mirror of
https://github.com/lllin000/PaperForge.git
synced 2026-07-22 06:50:53 +00:00
964e05b305
- A0: extract post-inventory seam into ocr_object_writeback module - A1: add ownership evidence contract + idempotency guard - A2: claim side-adjacent figure text (M84CTEM9 fix)
277 lines
9.9 KiB
Python
277 lines
9.9 KiB
Python
from __future__ import annotations
|
|
|
|
import pytest
|
|
|
|
from paperforge.worker.ocr_figures import (
|
|
_extract_figure_marker,
|
|
_extract_figure_number,
|
|
_format_figure_id,
|
|
build_figure_inventory_vnext,
|
|
)
|
|
|
|
|
|
class TestExtractFigureNumber:
|
|
"""Test _extract_figure_number after regex change to named groups."""
|
|
|
|
def test_main_after_regex_change(self):
|
|
assert _extract_figure_number("Figure 1. Caption") == 1
|
|
|
|
def test_supplementary_after_regex_change(self):
|
|
assert _extract_figure_number("Figure S1. Caption") == 1
|
|
|
|
def test_appendix_after_regex_change(self):
|
|
assert _extract_figure_number("Figure A1. Caption") == 1
|
|
|
|
|
|
class TestExtractFigureMarker:
|
|
"""Test _extract_figure_marker returns correct structured dict."""
|
|
|
|
def test_supplementary_keyword_has_no_prefix(self):
|
|
marker = _extract_figure_marker("Supplementary Figure 1. Caption")
|
|
# "Supplementary" keyword determines namespace, not the S prefix.
|
|
# Since there's no explicit prefix letter before the number,
|
|
# prefix is None.
|
|
assert marker["prefix"] is None
|
|
assert marker["namespace"] == "supplementary"
|
|
assert _format_figure_id("supplementary", 1) == "figure_s001"
|
|
|
|
def test_supplementary_keyword_overrides_appendix_prefix(self):
|
|
"""Supplementary keyword in text takes precedence over prefix letter A."""
|
|
marker = _extract_figure_marker("Supplementary Figure A1. Caption")
|
|
# prefix is "A", but "supplementary" keyword forces namespace to
|
|
# "supplementary", overriding the appendix inference from prefix.
|
|
assert marker["namespace"] == "supplementary"
|
|
|
|
def test_extended_data_keyword_overrides_appendix_prefix(self):
|
|
"""Extended Data keyword in text takes precedence over prefix letter A."""
|
|
marker = _extract_figure_marker("Extended Data Figure A1. Caption")
|
|
assert marker["namespace"] == "extended_data"
|
|
|
|
def test_lowercase_appendix_prefix_normalized(self):
|
|
"""Lowercase prefix letter is normalized to uppercase."""
|
|
marker = _extract_figure_marker("Figure a1")
|
|
assert marker["prefix"] == "A"
|
|
|
|
@pytest.mark.parametrize(
|
|
"text, expected_prefix",
|
|
[
|
|
("Figure S1", "S"),
|
|
("Figure A1", "A"),
|
|
("Figure A.1", "A"),
|
|
("Figure B2", "B"),
|
|
("Figure 1", None),
|
|
],
|
|
)
|
|
def test_prefix_extraction(self, text, expected_prefix):
|
|
marker = _extract_figure_marker(text)
|
|
assert marker["prefix"] == expected_prefix
|
|
|
|
@pytest.mark.parametrize(
|
|
"text, expected_namespace",
|
|
[
|
|
("Figure S1", "supplementary"),
|
|
("Figure A1", "appendix"),
|
|
("Figure 1", "main"),
|
|
],
|
|
)
|
|
def test_namespace_resolution(self, text, expected_namespace):
|
|
marker = _extract_figure_marker(text)
|
|
assert marker["namespace"] == expected_namespace
|
|
|
|
|
|
class TestFormatFigureId:
|
|
"""Test _format_figure_id with alpha_prefix support."""
|
|
|
|
def test_appendix_letter_is_preserved(self):
|
|
"""Appendix prefix letter is embedded in the figure id."""
|
|
assert _format_figure_id("appendix", 2, alpha_prefix="B") == "figure_b002"
|
|
|
|
def test_figure_a1_and_figure_b1_do_not_collide(self):
|
|
"""Different appendix letters produce distinct ids for same number."""
|
|
id_a = _format_figure_id("appendix", 1, alpha_prefix="A")
|
|
id_b = _format_figure_id("appendix", 1, alpha_prefix="B")
|
|
assert id_a == "figure_a001"
|
|
assert id_b == "figure_b001"
|
|
assert id_a != id_b
|
|
|
|
def test_main_number_unchanged(self):
|
|
"""Main namespace unchanged by alpha_prefix addition."""
|
|
assert _format_figure_id("main", 1) == "figure_001"
|
|
|
|
def test_supplementary_unchanged(self):
|
|
"""Supplementary namespace unchanged by alpha_prefix addition."""
|
|
assert _format_figure_id("supplementary", 1) == "figure_s001"
|
|
|
|
|
|
class TestTableA1Leak:
|
|
"""Ensure Table A1 captions do not leak into the figure inventory."""
|
|
|
|
def test_table_a1_caption_not_consumed_by_figure_inventory(self):
|
|
blocks = [
|
|
{
|
|
"block_id": "t1",
|
|
"page": 1,
|
|
"role": "table_caption",
|
|
"text": "Table A1. Patient Demographics",
|
|
"bbox": [100, 100, 500, 120],
|
|
"zone": "display_zone",
|
|
"style_family": "legend_like",
|
|
},
|
|
{
|
|
"block_id": "a1",
|
|
"page": 1,
|
|
"role": "media_asset",
|
|
"text": "",
|
|
"bbox": [100, 130, 500, 600],
|
|
"zone": "display_zone",
|
|
},
|
|
]
|
|
result = build_figure_inventory_vnext(blocks, 1200)
|
|
matched = result.get("matched_figures", [])
|
|
# No figure should have been created with an appendix a-prefixed id
|
|
assert not any(m.get("figure_id", "") == "figure_a001" for m in matched)
|
|
# No matched figure text should contain "Table A1"
|
|
assert not any("Table A1" in (m.get("text") or "") for m in matched)
|
|
|
|
|
|
class TestTableLeakRegression:
|
|
"""Guard against table captions being parsed by figure regex helpers."""
|
|
|
|
def test_table_numeric_caption_is_not_a_figure_number(self):
|
|
assert _extract_figure_number("TABLE 2. Baseline characteristics") is None
|
|
|
|
def test_table_appendix_caption_marker_has_no_figure_number(self):
|
|
marker = _extract_figure_marker("TABLE A1. Patient Demographics")
|
|
assert marker["number"] is None
|
|
assert marker["namespace"] == "main"
|
|
|
|
|
|
class TestCrossPageIntBlockIds:
|
|
"""Real structured blocks use per-page int block IDs; settlement must preserve legend text."""
|
|
|
|
def test_cross_page_reserved_lookup_handles_int_block_ids(self):
|
|
blocks = [
|
|
{
|
|
"block_id": 1,
|
|
"page": 1,
|
|
"role": "figure_caption",
|
|
"text": "Figure 1. Cross-page caption.",
|
|
"bbox": [100, 100, 500, 150],
|
|
},
|
|
{
|
|
"block_id": 2,
|
|
"page": 2,
|
|
"role": "figure_asset",
|
|
"text": "",
|
|
"bbox": [100, 100, 500, 400],
|
|
},
|
|
]
|
|
|
|
result = build_figure_inventory_vnext(blocks, 1200)
|
|
matched = result.get("matched_figures", [])
|
|
assert len(matched) == 1
|
|
assert matched[0]["legend_block_id"] == "1"
|
|
assert matched[0]["text"] == "Figure 1. Cross-page caption."
|
|
|
|
|
|
class TestPageLocalBlockIds:
|
|
"""Structured block ids repeat per page; figure materialization must use page + block_id."""
|
|
|
|
def test_same_page_match_uses_legend_from_same_page(self):
|
|
blocks = [
|
|
{
|
|
"block_id": 4,
|
|
"page": 4,
|
|
"role": "figure_caption",
|
|
"text": "TABLE 2\nWrong text from another page",
|
|
"bbox": [100, 100, 500, 150],
|
|
},
|
|
{
|
|
"block_id": 4,
|
|
"page": 5,
|
|
"role": "figure_caption",
|
|
"text": "Figure 1. Real same-page caption.",
|
|
"bbox": [100, 100, 500, 150],
|
|
},
|
|
{
|
|
"block_id": "asset1",
|
|
"page": 5,
|
|
"role": "figure_asset",
|
|
"text": "",
|
|
"bbox": [100, 180, 500, 500],
|
|
},
|
|
]
|
|
|
|
result = build_figure_inventory_vnext(blocks, 1200)
|
|
matched = result.get("matched_figures", [])
|
|
assert len(matched) == 1
|
|
assert matched[0]["page"] == 5
|
|
assert matched[0]["text"] == "Figure 1. Real same-page caption."
|
|
|
|
|
|
class TestAppendixSamePageMatching:
|
|
"""Appendix captions should score as formal figure legends when near media."""
|
|
|
|
def test_appendix_same_page_candidate_matches_media(self):
|
|
blocks = [
|
|
{
|
|
"block_id": "c1",
|
|
"page": 1,
|
|
"role": "figure_caption_candidate",
|
|
"text": "Figure A1. Appendix caption.",
|
|
"bbox": [100, 100, 500, 130],
|
|
},
|
|
{
|
|
"block_id": "a1",
|
|
"page": 1,
|
|
"role": "media_asset",
|
|
"text": "",
|
|
"bbox": [100, 160, 500, 500],
|
|
},
|
|
]
|
|
|
|
result = build_figure_inventory_vnext(blocks, 1200)
|
|
matched = result.get("matched_figures", [])
|
|
assert len(matched) == 1
|
|
assert matched[0]["figure_id"] == "figure_a001"
|
|
|
|
class TestSideAdjacentFigureTextLeak:
|
|
"""M84CTEM9 class: body blocks side-adjacent to a figure should not leak into body."""
|
|
|
|
def test_side_adjacent_text_does_not_leak_into_body(self):
|
|
from paperforge.worker.ocr_figures import build_figure_inventory_vnext
|
|
from paperforge.worker.ocr_tables import build_table_inventory_vnext
|
|
from paperforge.worker.ocr_object_writeback import apply_object_writebacks
|
|
|
|
blocks = [
|
|
{
|
|
"block_id": "note1",
|
|
"page": 8,
|
|
"role": "body_paragraph",
|
|
"raw_label": "text",
|
|
"bbox": [191, 1022, 514, 1083],
|
|
"text": "Age in years - 67:52.8",
|
|
},
|
|
{
|
|
"block_id": "asset1",
|
|
"page": 8,
|
|
"role": "media_asset",
|
|
"raw_label": "image",
|
|
"bbox": [534, 976, 978, 1334],
|
|
"text": "",
|
|
},
|
|
{
|
|
"block_id": "cap1",
|
|
"page": 8,
|
|
"role": "figure_caption",
|
|
"raw_label": "text",
|
|
"bbox": [534, 1340, 978, 1420],
|
|
"text": "Figure A1. Odds ratios with CIs for the probability of a rotator cuff tear.",
|
|
},
|
|
]
|
|
|
|
fig_inv = build_figure_inventory_vnext(blocks, 1200)
|
|
tab_inv = build_table_inventory_vnext(blocks)
|
|
apply_object_writebacks(structured_blocks=blocks, figure_inventory=fig_inv, table_inventory=tab_inv)
|
|
|
|
assert blocks[0]["role"] == "figure_inner_text"
|