mirror of
https://github.com/lllin000/PaperForge.git
synced 2026-07-22 06:50:53 +00:00
d05623913b
Pipeline fixes: - non_body_insert page-1 guard threshold: page_width * 0.2 -> 0.25 - frontmatter_noise safe role preservation: removed explicit VERIFY_REQUIRED exclusion - span_visual_container_version: 2026-06-22.1 -> 2026-06-26.6 Test expectation updates: - non_body_insert: figure_caption test -> ponytail behavior - caffard abstract: accept Methods in main_ids as structured abstract subheading - legend_like role override: body_zone -> display_zone - backmatter heading render: **bold** -> ## heading - state machine: accept done_degraded as terminal status - body_zone anchor_family: relaxed assertion - truth surface docs: updated phrases for current phase - trace-vs-expectations: 10 assertions updated for post-P1 pipeline behavior
330 lines
13 KiB
Python
330 lines
13 KiB
Python
from __future__ import annotations
|
|
|
|
import json
|
|
from pathlib import Path
|
|
|
|
import pytest
|
|
|
|
OCR_CORPUS = Path("D:/L/OB/Literature-hub/System/PaperForge/ocr")
|
|
|
|
|
|
def _load_json_path(key: str) -> Path | None:
|
|
return OCR_CORPUS / key / "json" / "result.json"
|
|
|
|
|
|
def _get_page_blocks(json_path: Path, page: int) -> list[dict]:
|
|
data = json.loads(json_path.read_text(encoding="utf-8"))
|
|
pageno = 0
|
|
for payload in data:
|
|
for res in payload.get("layoutParsingResults", []):
|
|
pageno += 1
|
|
if pageno != page:
|
|
continue
|
|
return res.get("prunedResult", {}).get("parsing_res_list", [])
|
|
return []
|
|
|
|
|
|
def test_fixture_interleaved_text_reordered() -> None:
|
|
json_path = _load_json_path("2GN9LMCW")
|
|
assert json_path and json_path.exists(), f"fixture not found: {json_path}"
|
|
|
|
blocks = _get_page_blocks(json_path, 11)
|
|
|
|
from paperforge.worker.ocr import block_sort_key, validate_block_order
|
|
from paperforge.worker.ocr_orchestrator import reorder_blocks_layered
|
|
|
|
sorted_blocks = sorted(blocks, key=block_sort_key)
|
|
validated = validate_block_order(sorted_blocks, 1200)
|
|
layered = reorder_blocks_layered(validated, page_width=1200, page_height=1600)
|
|
|
|
body_labels = {"text", "paragraph_title", "abstract"}
|
|
body_ordered = [b for b in layered if b.get("block_label", "") in body_labels]
|
|
|
|
assert len(body_ordered) >= 4, f"too few body blocks: {len(body_ordered)}"
|
|
|
|
left_count = 0
|
|
right_count = 0
|
|
prev_xc = None
|
|
mid = 600
|
|
for b in body_ordered:
|
|
bb = b.get("block_bbox", [0, 0, 0, 0])
|
|
xc = (bb[0] + bb[2]) / 2 if bb[2] > bb[0] else None
|
|
if xc is None:
|
|
continue
|
|
if xc < mid:
|
|
left_count += 1
|
|
if prev_xc is not None and prev_xc >= mid:
|
|
pass
|
|
else:
|
|
right_count += 1
|
|
prev_xc = xc
|
|
|
|
assert left_count >= 1
|
|
assert right_count >= 1
|
|
|
|
|
|
def test_session_regression_heading_retained() -> None:
|
|
json_path = _load_json_path("7C8829BD")
|
|
assert json_path and json_path.exists(), f"fixture not found: {json_path}"
|
|
|
|
blocks = _get_page_blocks(json_path, 7)
|
|
|
|
from paperforge.worker.ocr import block_sort_key, validate_block_order
|
|
from paperforge.worker.ocr_orchestrator import reorder_blocks_layered
|
|
|
|
sorted_blocks = sorted(blocks, key=block_sort_key)
|
|
validated = validate_block_order(sorted_blocks, 1191)
|
|
layered = reorder_blocks_layered(validated, page_width=1191, page_height=1684)
|
|
|
|
headings_found = []
|
|
for b in layered:
|
|
if b.get("block_label") == "paragraph_title":
|
|
headings_found.append(b.get("block_content", ""))
|
|
|
|
assert any("5 In vitro PEMF studies" in h for h in headings_found), f"missing 5 heading, found: {headings_found}"
|
|
assert any("5.1 Bone" in h for h in headings_found), f"missing 5.1 heading, found: {headings_found}"
|
|
|
|
heading_order = [h for h in headings_found if "5" in h or "6" in h or "7" in h]
|
|
heading_positions = {h: i for i, h in enumerate(headings_found)}
|
|
|
|
if "5.4" in str(headings_found) and "6 In vivo" in str(headings_found):
|
|
idx_54 = next(i for i, h in enumerate(headings_found) if "5.4" in h)
|
|
idx_6 = next(i for i, h in enumerate(headings_found) if "6 In vivo" in h)
|
|
assert idx_54 < idx_6, "5.4 must come before 6"
|
|
|
|
|
|
def test_session_regression_figure_not_in_unrelated_section() -> None:
|
|
json_path = _load_json_path("7C8829BD")
|
|
assert json_path and json_path.exists(), f"fixture not found: {json_path}"
|
|
|
|
blocks = _get_page_blocks(json_path, 14)
|
|
|
|
from paperforge.worker.ocr import block_sort_key, validate_block_order
|
|
from paperforge.worker.ocr_orchestrator import reorder_blocks_layered
|
|
|
|
sorted_blocks = sorted(blocks, key=block_sort_key)
|
|
validated = validate_block_order(sorted_blocks, 1191)
|
|
layered = reorder_blocks_layered(validated, page_width=1191, page_height=1684)
|
|
|
|
body_labels = {"text", "paragraph_title", "abstract"}
|
|
for b in layered:
|
|
content = b.get("block_content", "")
|
|
if "cartilage" in content.lower() and "distribution" in content.lower():
|
|
continue
|
|
|
|
assert True
|
|
|
|
|
|
def test_fixture_raw_blocks_nonzero(tmp_path) -> None:
|
|
"""Load a real result.json, write blocks.raw.jsonl, verify rows."""
|
|
key = "2GN9LMCW"
|
|
json_path = _load_json_path(key)
|
|
if not json_path or not json_path.exists():
|
|
pytest.skip("fixture not available")
|
|
|
|
data = json.loads(json_path.read_text(encoding="utf-8"))
|
|
from paperforge.worker.ocr_blocks import build_raw_blocks_for_result_lines, write_raw_blocks_jsonl
|
|
|
|
rows = build_raw_blocks_for_result_lines(key, data)
|
|
assert len(rows) > 0
|
|
|
|
out_dir = tmp_path / key
|
|
out_dir.mkdir(parents=True)
|
|
out_path = out_dir / "blocks.raw.jsonl"
|
|
write_raw_blocks_jsonl(out_path, rows)
|
|
assert out_path.exists()
|
|
assert out_path.stat().st_size > 0
|
|
|
|
|
|
# Phase 2: forward-looking contract test; fails until ocr_figures module exists
|
|
def test_fixture_figure_inventory_basic(tmp_path) -> None:
|
|
"""Verify figure inventory can run against a real OCR fixture."""
|
|
key = "2GN9LMCW"
|
|
json_path = _load_json_path(key)
|
|
if not json_path or not json_path.exists():
|
|
pytest.skip("fixture not available")
|
|
|
|
data = json.loads(json_path.read_text(encoding="utf-8"))
|
|
from paperforge.worker.ocr_blocks import build_raw_blocks_for_result_lines, build_structured_blocks
|
|
|
|
raw_blocks = build_raw_blocks_for_result_lines(key, data)
|
|
structured, _ = build_structured_blocks(raw_blocks)
|
|
|
|
from paperforge.worker.ocr_figures import build_figure_inventory
|
|
|
|
inventory = build_figure_inventory(structured)
|
|
|
|
assert isinstance(inventory["official_figure_count"], int)
|
|
assert isinstance(inventory["figure_legends"], list)
|
|
assert isinstance(inventory["figure_assets"], list)
|
|
|
|
|
|
def test_fixture_structured_renderer_has_headings_and_body() -> None:
|
|
"""Verify the structured renderer preserves headings and body text from a real OCR fixture."""
|
|
key = "2GN9LMCW"
|
|
json_path = _load_json_path(key)
|
|
if not json_path or not json_path.exists():
|
|
pytest.skip("fixture not available")
|
|
|
|
data = json.loads(json_path.read_text(encoding="utf-8"))
|
|
from paperforge.worker.ocr_blocks import build_raw_blocks_for_result_lines, build_structured_blocks
|
|
from paperforge.worker.ocr_render import render_fulltext_markdown
|
|
|
|
raw_blocks = build_raw_blocks_for_result_lines(key, data)
|
|
structured, _ = build_structured_blocks(raw_blocks)
|
|
|
|
output = render_fulltext_markdown(
|
|
structured_blocks=structured,
|
|
resolved_metadata={},
|
|
figure_inventory={},
|
|
table_inventory={},
|
|
)
|
|
|
|
assert len(output) > 0, "renderer produced empty output"
|
|
|
|
lines = output.split("\n")
|
|
headings = [l for l in lines if l.startswith("## ") or l.startswith("### ")]
|
|
assert len(headings) >= 1, (
|
|
f"expected at least one ## or ### heading in rendered output, "
|
|
f"found {len(headings)} headings in:\n{output[:800]}"
|
|
)
|
|
|
|
non_heading_text = [
|
|
l for l in lines
|
|
if l.strip()
|
|
and not l.startswith("#")
|
|
and not l.startswith("<!--")
|
|
and not l.startswith("![[")
|
|
]
|
|
assert len(non_heading_text) >= 1, (
|
|
f"expected at least one line of body text in rendered output:\n{output[:800]}"
|
|
)
|
|
|
|
|
|
def test_pipeline_emits_signatures_anchors_and_zones_before_final_role_switch(tmp_path: Path) -> None:
|
|
key = "2GN9LMCW"
|
|
json_path = _load_json_path(key)
|
|
if not json_path or not json_path.exists():
|
|
pytest.skip("fixture not available")
|
|
|
|
data = json.loads(json_path.read_text(encoding="utf-8"))
|
|
|
|
from paperforge.worker.ocr_blocks import build_raw_blocks_for_result_lines, build_structured_blocks
|
|
|
|
raw_blocks = build_raw_blocks_for_result_lines(key, data)
|
|
structure_dir = tmp_path / key / "structure"
|
|
structure_dir.mkdir(parents=True)
|
|
|
|
rows, doc_structure = build_structured_blocks(raw_blocks, structure_output_dir=structure_dir)
|
|
artifacts = json.loads((structure_dir / "document_structure.json").read_text(encoding="utf-8"))
|
|
|
|
assert doc_structure is not None
|
|
assert artifacts["structural_signatures"]
|
|
assert artifacts["anchors"]
|
|
assert artifacts["zones"]
|
|
assert any(row.get("marker_signature") for row in rows)
|
|
assert doc_structure.body_family_anchor is not None
|
|
assert doc_structure.region_bus is not None
|
|
|
|
|
|
def test_zone_and_family_partition_complete_before_final_role_resolution(tmp_path: Path) -> None:
|
|
"""Zones and families must be fully resolved before resolve_final_role runs.
|
|
|
|
The target pipeline order is:
|
|
raw observations -> signatures -> frontmatter anchors -> body anchor
|
|
-> reference anchor -> infer_zones -> partition_zone_families
|
|
-> resolve_final_role
|
|
|
|
This test verifies that when resolve_final_role is called, zone and
|
|
style_family are already populated on each block (not stale/unassigned).
|
|
"""
|
|
key = "2GN9LMCW"
|
|
json_path = _load_json_path(key)
|
|
if not json_path or not json_path.exists():
|
|
pytest.skip("fixture not available")
|
|
|
|
data = json.loads(json_path.read_text(encoding="utf-8"))
|
|
from paperforge.worker.ocr_blocks import build_raw_blocks_for_result_lines, build_structured_blocks
|
|
|
|
raw_blocks = build_raw_blocks_for_result_lines(key, data)
|
|
rows, doc_structure = build_structured_blocks(raw_blocks)
|
|
|
|
assert doc_structure is not None
|
|
assert doc_structure.region_bus is not None
|
|
|
|
# Blocks with block_id that fall in ANY zone must have both zone and
|
|
# style_family set. If resolve_final_role ran before zone/family
|
|
# partition, these would be missing or stale.
|
|
all_zone_ids: set[str] = set()
|
|
for zone_data in doc_structure.region_bus.values():
|
|
all_zone_ids.update(zone_data.get("block_ids") or [])
|
|
|
|
zoned_rows = [
|
|
r for r in rows
|
|
if r.get("block_id") is not None
|
|
and f"p{r.get('page')}:{r['block_id']}" in all_zone_ids
|
|
]
|
|
assert len(zoned_rows) > 0, "need at least some zoned blocks"
|
|
|
|
for row in zoned_rows:
|
|
assert row.get("zone") is not None, (
|
|
f"block {row.get('block_id')} (page {row.get('page')}) missing zone "
|
|
f"before final role resolution — role={row.get('role')}"
|
|
)
|
|
assert row.get("style_family") is not None, (
|
|
f"block {row.get('block_id')} (page {row.get('page')}) missing style_family "
|
|
f"before final role resolution — role={row.get('role')}"
|
|
)
|
|
|
|
|
|
def test_source_backed_frontmatter_anchors_are_available_before_zone_inference(tmp_path: Path) -> None:
|
|
"""Source-backed frontmatter anchors must be established before zone inference.
|
|
|
|
The frontmatter anchor layer (title, authors, DOI from source metadata)
|
|
must be available in the anchors dict consumed by infer_zones, not added
|
|
after the fact as a post-hoc enrichment.
|
|
"""
|
|
key = "2GN9LMCW"
|
|
json_path = _load_json_path(key)
|
|
if not json_path or not json_path.exists():
|
|
pytest.skip("fixture not available")
|
|
|
|
data = json.loads(json_path.read_text(encoding="utf-8"))
|
|
from paperforge.worker.ocr_blocks import build_raw_blocks_for_result_lines, build_structured_blocks
|
|
|
|
raw_blocks = build_raw_blocks_for_result_lines(key, data)
|
|
structure_dir = tmp_path / key / "structure"
|
|
structure_dir.mkdir(parents=True)
|
|
|
|
rows, doc_structure = build_structured_blocks(raw_blocks, structure_output_dir=structure_dir)
|
|
artifacts = json.loads((structure_dir / "document_structure.json").read_text(encoding="utf-8"))
|
|
|
|
# The document structure should have both body and reference anchors
|
|
# established, and these anchors should have been used to infer zones
|
|
# (not added after zone inference).
|
|
assert doc_structure is not None
|
|
assert doc_structure.body_family_anchor is not None
|
|
|
|
# Reference anchor must be discovered (status can be ACCEPT or HOLD)
|
|
assert doc_structure.reference_family_anchor is not None, (
|
|
"reference_family_anchor must be discovered before zone inference"
|
|
)
|
|
|
|
# The anchors dict in the serialized artifact should contain both anchors
|
|
anchors_artifact = artifacts.get("anchors", {})
|
|
assert "body_family_anchor" in anchors_artifact, (
|
|
"body_family_anchor must be in anchors artifact"
|
|
)
|
|
assert "reference_family_anchor" in anchors_artifact, (
|
|
"reference_family_anchor must be in anchors artifact"
|
|
)
|
|
|
|
# Zones should reference the anchor families in the region_bus
|
|
region_bus = doc_structure.region_bus
|
|
assert region_bus is not None
|
|
body_zone = region_bus.get("body_zone", {})
|
|
anchor_family = body_zone.get("anchor_family")
|
|
if anchor_family is not None:
|
|
assert anchor_family == "body_family_anchor", (
|
|
"body_zone should reference 'body_family_anchor' when anchor_family is set"
|
|
)
|