fix: TSR auto-fallback bugs found in review, v1.7.2 (#68)

Three fixes to extract_tables_with_structure_auto_mem (added in
1.7.1) caught by external review:

1. multi_row_in_cell over-triggered on legitimate multi-line cells.
   The previous threshold (item span > 1.3× either smallest cell or
   tallest item height) fires on any cell with 2+ y-separated text
   items — including rowspan>1 cells, wrapped descriptions, and
   superscript/subscript runs. Replaced with two gates:
   - skip cells whose declared rowspan > 1 (intentional multi-line)
   - require an actual whitespace gap (>~half a line height)
     between the bottom of one item and the top of the next, in
     PDF-native y-coordinates. Same-line items with tall glyphs or
     superscripts have negative or near-zero gap; truly separate
     visual rows have gap ≈ leading − line-height.
   FNBO regression test still passes; new test covers a rowspan=2
   cell with two visible text lines and verifies no fallback fires.

2. Heuristic returning empty silently replaced TSR markdown with
   "". The auto wrapper now keeps the TSR markdown when the
   heuristic markdown is empty/whitespace and tags fallback_reason
   with `_heuristic_empty` suffix (e.g.
   `multi_row_in_cell_heuristic_empty`). Worst case we ship the
   same wrong-but-non-empty TSR output we'd have shipped before
   1.7.1; we never replace useful output with literally nothing.

3. One bad input blanked the whole batch. Errors from
   detect_tsr_quality_issue or extract_tables_in_regions_mem now
   stay scoped to the single input — that input falls through to
   raw TSR markdown with a `_error` reason label so callers can
   metric on it. Other inputs in the batch are unaffected.

3 new integration tests:
- test_auto_does_not_fire_on_legit_rowspan_cell
- test_auto_keeps_tsr_markdown_when_heuristic_returns_empty
- test_auto_isolates_per_input_failures

All 6 auto tests + full 123-test suite pass. FNBO local replay
still triggers fallback (phantom_empty_row signal in this run) and
emits correct Shawnee/BVP/Sonoma rows with correct census tracts.

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

tests/integration_tests.rs

@@ -2190,6 +2190,219 @@ fn test_auto_returns_empty_inputs() {
     assert!(results.is_empty());
 }
 
+#[test]
+fn test_auto_does_not_fire_on_legit_rowspan_cell() {
+    use pdf_inspector::{extract_tables_with_structure_auto_mem, TsrTableInput};
+
+    let buf = synthetic_dense_table_pdf();
+    // 2 columns, 3 rows in the visible PDF. SLANet emits a 2-row table
+    // where the LEFT cell of row 1 is a rowspan=2 cell that legitimately
+    // covers Oak Street + Boardwalk on two visual lines. The right
+    // column has two normal rows. multi_row_in_cell must NOT fire on
+    // the rowspan=2 cell.
+    let tokens: Vec<String> = [
+        "<table>",
+        "<thead>",
+        "<tr>",
+        "<th></th>",
+        "<th></th>",
+        "</tr>",
+        "</thead>",
+        "<tbody>",
+        "<tr>",
+        // First data cell explicitly declares rowspan=2.
+        "<td",
+        " rowspan=\"2\"",
+        ">",
+        "</td>",
+        "<td></td>",
+        "</tr>",
+        "<tr>",
+        "<td></td>",
+        "</tr>",
+        "</tbody>",
+        "</table>",
+    ]
+    .into_iter()
+    .map(String::from)
+    .collect();
+    // Header row, then a tall left cell covering both data lines, plus
+    // two narrow right cells (one per line).
+    let cell_bboxes = vec![
+        poly(10.0, 88.0, 100.0, 105.0),
+        poly(90.0, 88.0, 180.0, 105.0),
+        poly(10.0, 105.0, 100.0, 145.0), // rowspan=2 — covers both lines
+        poly(90.0, 105.0, 180.0, 122.0), // row 1 only
+        poly(90.0, 122.0, 180.0, 145.0), // row 2 only
+    ];
+
+    let results = extract_tables_with_structure_auto_mem(
+        &buf,
+        &[TsrTableInput {
+            page: 0,
+            crop_pdf_pt_bbox: [0.0, 0.0, 200.0, 800.0],
+            render_dpi: 72.0,
+            structure_tokens: tokens,
+            cell_bboxes,
+        }],
+    )
+    .unwrap();
+    assert_eq!(results.len(), 1);
+    assert!(
+        results[0].fallback_reason.is_none(),
+        "rowspan=2 cell containing 2 visual lines should not trip multi_row_in_cell, got reason={:?}",
+        results[0].fallback_reason,
+    );
+}
+
+#[test]
+fn test_auto_keeps_tsr_markdown_when_heuristic_returns_empty() {
+    use pdf_inspector::{extract_tables_with_structure_auto_mem, TsrTableInput};
+
+    let buf = synthetic_dense_table_pdf();
+    // Same shape as the multi_row_in_cell regression — a tall data cell
+    // that catches Oak Street + Boardwalk. But the crop bbox we pass
+    // points at a strip of the page that has NO text items, so the
+    // heuristic's region will be empty when it tries to extract there.
+    // The auto wrapper must keep the TSR markdown rather than ship "".
+    let tokens: Vec<String> = [
+        "<table>",
+        "<thead>",
+        "<tr>",
+        "<th></th>",
+        "<th></th>",
+        "</tr>",
+        "</thead>",
+        "<tbody>",
+        "<tr>",
+        "<td></td>",
+        "<td></td>",
+        "</tr>",
+        "</tbody>",
+        "</table>",
+    ]
+    .into_iter()
+    .map(String::from)
+    .collect();
+    // Cell bboxes overlap the actual PDF text (so multi_row_in_cell
+    // fires) — but the crop_pdf_pt_bbox we hand to the heuristic is a
+    // wholly-empty region of the page. The heuristic should return "".
+    let cell_bboxes = vec![
+        poly(10.0, 88.0, 100.0, 105.0),
+        poly(90.0, 88.0, 180.0, 105.0),
+        poly(10.0, 105.0, 100.0, 145.0),
+        poly(90.0, 105.0, 180.0, 145.0),
+    ];
+
+    let results = extract_tables_with_structure_auto_mem(
+        &buf,
+        &[TsrTableInput {
+            page: 0,
+            // Crop is at the BOTTOM of the page where there's no text.
+            crop_pdf_pt_bbox: [0.0, 0.0, 200.0, 50.0],
+            render_dpi: 72.0,
+            structure_tokens: tokens,
+            cell_bboxes,
+        }],
+    )
+    .unwrap();
+    assert_eq!(results.len(), 1);
+    let r = &results[0];
+    assert_eq!(
+        r.fallback_reason.as_deref(),
+        Some("multi_row_in_cell_heuristic_empty"),
+        "expected _heuristic_empty suffix, got {:?}",
+        r.fallback_reason,
+    );
+    // TSR markdown should be preserved — non-empty, contains the cell
+    // text we know was assigned by the TSR path.
+    assert!(
+        !r.markdown.trim().is_empty(),
+        "expected TSR markdown to be preserved, got empty",
+    );
+    assert!(
+        r.markdown.contains("Oak Street") || r.markdown.contains("Boardwalk"),
+        "expected TSR markdown to contain at least one row, got: {}",
+        r.markdown,
+    );
+}
+
+#[test]
+fn test_auto_isolates_per_input_failures() {
+    use pdf_inspector::{extract_tables_with_structure_auto_mem, TsrTableInput};
+
+    let buf = synthetic_dense_table_pdf();
+    let good_tokens: Vec<String> = [
+        "<table>",
+        "<thead>",
+        "<tr>",
+        "<th></th>",
+        "<th></th>",
+        "</tr>",
+        "</thead>",
+        "<tbody>",
+        "<tr>",
+        "<td></td>",
+        "<td></td>",
+        "</tr>",
+        "<tr>",
+        "<td></td>",
+        "<td></td>",
+        "</tr>",
+        "</tbody>",
+        "</table>",
+    ]
+    .into_iter()
+    .map(String::from)
+    .collect();
+    // A clean input that should pass through with no fallback.
+    let good_input = TsrTableInput {
+        page: 0,
+        crop_pdf_pt_bbox: [0.0, 0.0, 200.0, 800.0],
+        render_dpi: 72.0,
+        structure_tokens: good_tokens,
+        cell_bboxes: vec![
+            poly(10.0, 72.0, 100.0, 112.0),
+            poly(90.0, 72.0, 180.0, 112.0),
+            poly(10.0, 88.8, 100.0, 128.8),
+            poly(90.0, 88.8, 180.0, 128.8),
+            poly(10.0, 105.6, 100.0, 145.6),
+            poly(90.0, 105.6, 180.0, 145.6),
+        ],
+    };
+    // A bad input that targets a non-existent page. The detection
+    // helper short-circuits on missing pages with Ok(None), so this
+    // shouldn't itself crash, but pairing it with a flagged input
+    // exercises the per-input control flow regardless. The point of
+    // this test is that one input's outcome doesn't poison the other.
+    let bad_input = TsrTableInput {
+        page: 9999,
+        crop_pdf_pt_bbox: [0.0, 0.0, 100.0, 100.0],
+        render_dpi: 72.0,
+        structure_tokens: vec![
+            "<table>".into(),
+            "<tr>".into(),
+            "<td></td>".into(),
+            "</tr>".into(),
+            "</table>".into(),
+        ],
+        cell_bboxes: vec![poly(0.0, 0.0, 50.0, 50.0)],
+    };
+
+    let results = extract_tables_with_structure_auto_mem(&buf, &[good_input, bad_input]).unwrap();
+    assert_eq!(results.len(), 2);
+    // Good input still produces non-empty TSR markdown with no fallback.
+    assert!(
+        results[0].fallback_reason.is_none(),
+        "good input should pass through, got reason={:?}",
+        results[0].fallback_reason,
+    );
+    assert!(results[0].markdown.contains("Oak Street"));
+    // Bad input collapses to empty markdown but doesn't take the
+    // batch down with it.
+    assert_eq!(results[1].markdown, "");
+}
+
 // =========================================================================
 // extract_pages_markdown_mem tests
 // =========================================================================