From fd15d074c99202819ea78015e4f92a18dadb73f5 Mon Sep 17 00:00:00 2001
From: Research Assistant <research@example.com>
Date: Fri, 5 Jun 2026 10:22:18 +0800
Subject: [PATCH] test: preserve OCR render regressions through structured
 renderer

---
 tests/test_ocr_integration_fixtures.py |  42 ++++++++
 tests/test_ocr_rendering.py            | 136 +++++++++++++++++++++++++
 2 files changed, 178 insertions(+)

diff --git a/tests/test_ocr_integration_fixtures.py b/tests/test_ocr_integration_fixtures.py
index 26b63b11..6d5abcab 100644
--- a/tests/test_ocr_integration_fixtures.py
+++ b/tests/test_ocr_integration_fixtures.py
@@ -157,3 +157,45 @@ def test_fixture_figure_inventory_basic(tmp_path) -> None:
     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]}"
+    )
diff --git a/tests/test_ocr_rendering.py b/tests/test_ocr_rendering.py
index 92eaa2bf..86ebd64f 100644
--- a/tests/test_ocr_rendering.py
+++ b/tests/test_ocr_rendering.py
@@ -703,3 +703,139 @@ def test_embedded_figure_text_keeps_narrow_in_figure_note() -> None:
     ]
 
     assert is_embedded_figure_text_block(blocks[2], blocks, page_width=1200, page_height=1600) is True
+
+
+# --- Structured renderer regression tests ---
+
+
+def test_structured_renderer_abstract_before_introduction() -> None:
+    from paperforge.worker.ocr_render import render_fulltext_markdown
+
+    structured_blocks = [
+        {
+            "role": "section_heading",
+            "text": "1 Introduction",
+            "render_default": True,
+            "page": 1,
+        },
+        {
+            "role": "abstract_heading",
+            "text": "",
+            "render_default": True,
+            "page": 1,
+        },
+        {
+            "role": "abstract_body",
+            "text": "This is the abstract body text.",
+            "render_default": True,
+            "page": 1,
+        },
+        {
+            "role": "body_paragraph",
+            "text": "Some body paragraph after the introduction heading.",
+            "render_default": True,
+            "page": 1,
+        },
+    ]
+
+    output = render_fulltext_markdown(
+        structured_blocks=structured_blocks,
+        resolved_metadata={},
+        figure_inventory={},
+        table_inventory={},
+    )
+
+    abstract_idx = output.index("## Abstract")
+    abstract_text_idx = output.index("This is the abstract body text.")
+    intro_idx = output.index("## 1 Introduction")
+
+    assert abstract_idx < abstract_text_idx < intro_idx
+
+
+def test_structured_renderer_figure_and_table_links_rendered() -> None:
+    from paperforge.worker.ocr_render import render_fulltext_markdown
+
+    structured_blocks = [
+        {
+            "role": "figure_caption",
+            "text": "Figure 1. A test figure caption.",
+            "render_default": True,
+            "page": 1,
+        },
+        {
+            "role": "table_caption",
+            "text": "Table 1. A test table caption.",
+            "render_default": True,
+            "page": 1,
+        },
+        {
+            "role": "body_paragraph",
+            "text": "Body text between figures and tables.",
+            "render_default": True,
+            "page": 1,
+        },
+    ]
+
+    figure_inventory = {
+        "matched_figures": [
+            {"figure_id": "fig_001"},
+        ]
+    }
+    table_inventory = {
+        "tables": [
+            {"table_id": "tbl_001", "has_asset": True},
+        ]
+    }
+
+    output = render_fulltext_markdown(
+        structured_blocks=structured_blocks,
+        resolved_metadata={},
+        figure_inventory=figure_inventory,
+        table_inventory=table_inventory,
+    )
+
+    assert "![[render/figures/fig_001.md]]" in output
+    assert "![[render/tables/tbl_001.md]]" in output
+
+
+def test_structured_renderer_respects_render_default_false() -> None:
+    from paperforge.worker.ocr_render import render_fulltext_markdown
+
+    structured_blocks = [
+        {
+            "role": "body_paragraph",
+            "text": "VISIBLE_BODY",
+            "render_default": True,
+            "page": 1,
+        },
+        {
+            "role": "body_paragraph",
+            "text": "HIDDEN_BODY",
+            "render_default": False,
+            "page": 1,
+        },
+        {
+            "role": "section_heading",
+            "text": "Hidden Heading",
+            "render_default": False,
+            "page": 2,
+        },
+        {
+            "role": "body_paragraph",
+            "text": "ALSO_VISIBLE",
+            "render_default": True,
+            "page": 2,
+        },
+    ]
+
+    output = render_fulltext_markdown(
+        structured_blocks=structured_blocks,
+        resolved_metadata={},
+        figure_inventory={},
+        table_inventory={},
+    )
+
+    assert "VISIBLE_BODY" in output
+    assert "ALSO_VISIBLE" in output
+    assert "HIDDEN_BODY" not in output
+    assert "Hidden Heading" not in output