fix(markdown): locate drop-cap paragraphs by indent run, not cap size

The previous rule only handled two-line caps: it took the line
immediately above the cap and required the cap's em to be under 2.5
line-steps. That fired on exactly one document in 386 — Shannon — and
rejected every other real initial, because production initials are
36-47pt over 11-14pt leading and cover three or four lines.

A drop cap pushes every line it covers to the right of the glyph, so
the paragraph's first line is the topmost line sharing that indent,
whatever the line count. Walk that run instead of deriving a count from
the font size, and evaluate the paragraph-start and continuation guards
against the run's top rather than the cap's immediate neighbour.

A cap in a different column has no such run, so it is left alone.

Across pdf-evals (192) and opendataloader-bench (200) this now fixes
three documents instead of one, each losing a truncated first word and
a junk heading built from the orphaned cap ("# Omany patients"). The
opendataloader score moves 0.87836 -> 0.87912 overall and 0.7878 ->
0.7906 MHS with no metric regressing; 275_41c50d3a's pdf-evals
composite moves 0.5772 -> 0.5862.

src/markdown/analysis.rs

@@ -171,6 +171,74 @@ pub(crate) fn is_toc_marker_heading(text: &str) -> bool {
 /// equation and absent from name-plus-number headings. A bare trailing colon
 /// is NOT a fragment signal either: real headings frequently end with colons
 /// ("Procedure:", "Steps for Using the Microscope:").
+/// True when the line opens with a section number ("3.", "2.1.4", "IV)").
+///
+/// Mirrors the acceptance of `heading::parse_numbering` rather than the
+/// stricter `convert::starts_with_section_number`, which deliberately
+/// requires two components because it bypasses isolation checks. Here a
+/// single "1." counts: numbering is independent evidence of a heading, and
+/// `heading.rs` applies its numbered-prefix allowance *after* consulting
+/// `is_heading_fragment`, so without this exemption a numbered
+/// sentence-case heading would be vetoed before that allowance can run.
+fn starts_with_numbering_prefix(t: &str) -> bool {
+    let Some(first) = t.split_whitespace().next() else {
+        return false;
+    };
+    let has_delimiter = first.ends_with(['.', ')', ':']);
+    let token = first.trim_end_matches(['.', ')', ':']);
+    if token.is_empty() {
+        return false;
+    }
+    let parts: Vec<&str> = token.split('.').collect();
+    let decimal = parts
+        .iter()
+        .all(|p| !p.is_empty() && p.len() <= 3 && p.chars().all(|c| c.is_ascii_digit()));
+    if decimal {
+        // "1." / "2.1." carry a delimiter; "2.3 Title" is written without
+        // one, so a multi-component number is accepted bare. A bare single
+        // number ("3 apples") is not — that is ordinary prose.
+        return has_delimiter || parts.len() >= 2;
+    }
+    // Roman numerals go through the heading parser's own grammar so the two
+    // agree: uppercase I/V/X/L/C only, at most 8 characters. A looser rule
+    // here would exempt markers the parser rejects — "iv)" or "d)" from an
+    // alphabetical list — letting an ordinary list item bypass the veto and
+    // reach heading promotion.
+    //
+    // A delimiter is also required: a bare leading "I" is the pronoun far
+    // more often than a section number.
+    has_delimiter && crate::markdown::heading::roman_value(token).is_some()
+}
+
+/// True when the line reads as a title rather than a sentence: every
+/// content word (ignoring minor words) starts uppercase. Used to spare real
+/// headings from the dangling-verb veto — "Bond Yields" is a section title,
+/// "the method yields" is a stranded clause, and only the casing tells them
+/// apart.
+fn looks_title_case(t: &str) -> bool {
+    const MINOR: &[&str] = &[
+        "a", "an", "the", "of", "and", "or", "for", "to", "in", "on", "at", "by", "with", "from",
+        "as", "is", "are", "that", "than", "into",
+    ];
+    let mut content = 0usize;
+    let mut capitalized = 0usize;
+    for w in t.split_whitespace() {
+        let cleaned: String = w.chars().filter(|c| c.is_alphabetic()).collect();
+        if cleaned.is_empty() {
+            continue;
+        }
+        if MINOR.contains(&cleaned.to_lowercase().as_str()) {
+            continue;
+        }
+        content += 1;
+        if cleaned.chars().next().is_some_and(char::is_uppercase) {
+            capitalized += 1;
+        }
+    }
+    // A single content word ("Yields") is a title by default.
+    content == 0 || capitalized == content
+}
+
 pub(crate) fn is_heading_fragment(text: &str) -> bool {
     let t = text.trim_end();
 
@@ -244,9 +312,133 @@ pub(crate) fn is_heading_fragment(text: &str) -> bool {
     if t.ends_with(':') && t.split_whitespace().any(is_equation_number) {
         return true;
     }
+
+    // Dangling clause: a stranded sentence lead-in ends on a relational
+    // verb with no terminal punctuation — "Note that the exact error equals"
+    // left ahead of its formula when a phantom table dissolved.
+    //
+    // Gated on the line reading as prose rather than a title. Case is the
+    // discriminator the trailing word alone cannot provide: a heading is
+    // title case ("Bond Yields", "The Method Yields") while a stranded
+    // lead-in is sentence case ("the method yields"). Without this gate the
+    // veto eats real headings — "Bond Yields", "Crop Yields" and any wrapped
+    // title-case heading the preprocessor failed to merge.
+    if !t.ends_with(['.', '!', '?', ':', ';', ')', ']'])
+        && !looks_title_case(t)
+        && !starts_with_numbering_prefix(t)
+    {
+        if let Some(last) = t.split_whitespace().next_back() {
+            let word: String = last
+                .trim_matches(|c: char| !c.is_alphanumeric())
+                .to_lowercase();
+            // Relational verbs only, and only those with no common noun
+            // sense. "yields" was dropped for exactly that reason: "Bond
+            // Yields" is a real section title. Function words, copulas and
+            // auxiliaries were measured and rejected outright — a heading
+            // that wraps across lines ends on those, and suppressing them
+            // destroyed real IRS Publication 17 headings.
+            const DANGLING_TAIL: &[&str] =
+                &["equals", "denotes", "implies", "satisfies", "signifies"];
+            if DANGLING_TAIL.contains(&word.as_str()) {
+                return true;
+            }
+        }
+    }
     false
 }
 
+#[cfg(test)]
+mod fragment_heading_tests {
+    use super::is_heading_fragment;
+
+    #[test]
+    fn dangling_tail_marks_stranded_clause() {
+        // opendataloader 01030000000144: left behind when a phantom table
+        // dissolved, ahead of its formula on the next line.
+        assert!(is_heading_fragment("Note that the exact error equals"));
+        assert!(is_heading_fragment("The remainder term satisfies"));
+        assert!(is_heading_fragment("we conclude that the sum equals"));
+    }
+
+    #[test]
+    fn real_headings_survive() {
+        assert!(!is_heading_fragment("Introduction"));
+        assert!(!is_heading_fragment("Error Analysis"));
+        assert!(!is_heading_fragment("Materials and Methods"));
+        assert!(!is_heading_fragment("Results"));
+        assert!(!is_heading_fragment("3.2 Richardson Extrapolation"));
+        assert!(!is_heading_fragment("Discussion and Conclusions"));
+        // Terminal punctuation means the clause is complete.
+        assert!(!is_heading_fragment("What is a Derivative?"));
+        assert!(!is_heading_fragment("Procedure:"));
+        assert!(!is_heading_fragment("Note that this is important."));
+    }
+
+    #[test]
+    fn title_case_headings_ending_in_a_verb_survive() {
+        // "yields" is also a plural noun; these are real section titles.
+        assert!(!is_heading_fragment("Bond Yields"));
+        assert!(!is_heading_fragment("Crop Yields"));
+        assert!(!is_heading_fragment("Dividend Yields"));
+        assert!(!is_heading_fragment("Yields"));
+        // A wrapped title-case heading whose first line ends on a listed
+        // verb must survive even if the preprocessor failed to merge it.
+        assert!(!is_heading_fragment("The Theorem Implies"));
+        assert!(!is_heading_fragment("What This Denotes"));
+    }
+
+    #[test]
+    fn numbered_sentence_case_headings_survive() {
+        // heading.rs consults is_heading_fragment BEFORE applying its
+        // numbered-prefix allowance, so the veto must not pre-empt it.
+        assert!(!is_heading_fragment("1. What the model implies"));
+        assert!(!is_heading_fragment("2.3 How the estimator satisfies"));
+        assert!(!is_heading_fragment("IV) What this denotes"));
+        // Without numbering the same wording is still a stranded clause.
+        assert!(is_heading_fragment("What the model implies"));
+        // A bare leading number or pronoun is prose, not numbering.
+        assert!(is_heading_fragment("3 apples and what that implies"));
+        assert!(is_heading_fragment("I think the model implies"));
+        // Markers heading::parse_numbering rejects must not be exempted
+        // either, or an ordinary list item bypasses the veto: lowercase
+        // roman, alphabetical markers, and over-long tokens.
+        assert!(is_heading_fragment("iv) the estimator satisfies"));
+        assert!(is_heading_fragment("d) the value implies"));
+        // Unsupported character (M is outside the parser's I/V/X/L/C set).
+        assert!(is_heading_fragment("MMMM. the value implies"));
+        // Over-long token: nine valid characters, so this exercises the
+        // 8-character bound rather than the character set.
+        assert!(is_heading_fragment("IIIIIIIII. the value implies"));
+        // Eight is still within the bound and stays exempt.
+        assert!(!is_heading_fragment("IIIIIIII. What this implies"));
+        // Uppercase roman within the parser's grammar is still exempt.
+        assert!(!is_heading_fragment("IV. What this denotes"));
+        assert!(!is_heading_fragment("XII) What this implies"));
+    }
+
+    #[test]
+    fn wrapped_headings_are_not_fragments() {
+        // A heading that wraps across lines ends on a function word. These
+        // are real headings from IRS Publication 17 and must survive.
+        assert!(!is_heading_fragment("Casualty and"));
+        assert!(!is_heading_fragment("Rule 10. You Must Be at"));
+        assert!(!is_heading_fragment("Higher Standard Deduction for"));
+        assert!(!is_heading_fragment("Qualifying Child of"));
+        assert!(!is_heading_fragment("When Can I Withdraw or"));
+        // Copulas and auxiliaries also end real wrapped headings.
+        assert!(!is_heading_fragment("Rule 15. Your AGI Must Be"));
+        assert!(!is_heading_fragment("What Medical Expenses Are"));
+        assert!(!is_heading_fragment("Rule 13. You Must Have"));
+        assert!(!is_heading_fragment("When Can a Roth IRA Be"));
+    }
+
+    #[test]
+    fn dangling_check_is_case_insensitive() {
+        // All-caps is not sentence case, so the veto must not fire there.
+        assert!(!is_heading_fragment("THE REMAINDER EQUALS"));
+    }
+}
+
 /// Compute the Y-gap threshold for paragraph break detection.
 ///
 /// Instead of using a fixed multiple of base_size (which fails for double-spaced

src/markdown/heading.rs

@@ -127,7 +127,9 @@ fn visual_style(line: &TextLine) -> Option<VisualStyle> {
     })
 }
 
-fn roman_value(token: &str) -> Option<u32> {
+/// Shared with `analysis::starts_with_numbering_prefix` so the veto
+/// exemption and the heading parser agree on what a roman numeral is.
+pub(super) fn roman_value(token: &str) -> Option<u32> {
     if token.is_empty() || token.len() > 8 {
         return None;
     }

src/markdown/preprocess.rs

@@ -150,6 +150,61 @@ pub(crate) fn merge_heading_lines(
 /// Merge drop caps with the appropriate line.
 /// A drop cap is a single large letter at the start of a paragraph.
 /// Due to PDF coordinate sorting, the drop cap may appear AFTER the line it belongs to.
+/// True when the text ends a sentence, as opposed to merely ending in a
+/// period. An abbreviation or list marker ("e.g.", "Fig.", "Mr.", "1.")
+/// closes with a period mid-sentence, so treating those as paragraph
+/// boundaries would let a drop cap be prepended to a continuation.
+fn ends_sentence(text: &str) -> bool {
+    let t = text.trim_end();
+    if t.ends_with(['!', '?']) {
+        return true;
+    }
+    let Some(stripped) = t.strip_suffix('.') else {
+        return false;
+    };
+    let last = stripped.split_whitespace().next_back().unwrap_or("");
+    if last.is_empty() {
+        return false;
+    }
+
+    // Abbreviations carry an internal period between very short segments
+    // ("e.g.", "i.e.", "U.S."). Domains and decimals have the same shape but
+    // longer or numeric segments ("example.com.", "3.14."), and those end
+    // sentences perfectly well, so require every segment to be short and
+    // alphabetic before reading the internal period as an abbreviation.
+    if last.contains('.')
+        && last
+            .split('.')
+            .filter(|seg| !seg.is_empty())
+            // Characters, not bytes: a two-letter non-ASCII abbreviation
+            // ("т.е.", "ú.d.") measures four or more bytes and would
+            // otherwise be read as a completed sentence.
+            .all(|seg| seg.chars().count() <= 2 && seg.chars().all(char::is_alphabetic))
+    {
+        return false;
+    }
+
+    // Enumerators stand alone on their line ("1.", "ii.", "IV."). A number
+    // or numeral in the tail of a sentence does not — "published in 2020.",
+    // "He scored 5." and "after World War II." all end sentences, and
+    // treating them as markers would block a legitimate drop-cap merge.
+    if stripped.split_whitespace().count() == 1 {
+        let is_numeric = last.chars().all(|c| c.is_ascii_digit());
+        let is_roman = last
+            .chars()
+            .all(|c| matches!(c.to_ascii_uppercase(), 'I' | 'V' | 'X' | 'L' | 'C'));
+        if is_numeric || is_roman {
+            return false;
+        }
+    }
+
+    const ABBREVIATIONS: &[&str] = &[
+        "Fig", "No", "Mr", "Mrs", "Ms", "Dr", "St", "vs", "etc", "al", "Ed", "Eq", "Ch", "pp",
+        "Vol", "cf", "Prof", "Inc", "Ltd", "Jr", "Sr",
+    ];
+    !ABBREVIATIONS.iter().any(|a| a.eq_ignore_ascii_case(last))
+}
+
 pub(crate) fn merge_drop_caps(lines: Vec<TextLine>, base_size: f32) -> Vec<TextLine> {
     let mut result: Vec<TextLine> = Vec::with_capacity(lines.len());
 
@@ -169,6 +224,169 @@ pub(crate) fn merge_drop_caps(lines: Vec<TextLine>, base_size: f32) -> Vec<TextL
                 .map(|c| c.is_uppercase())
                 .unwrap_or(false);
 
+        // Embedded drop cap: a two-line cap's baseline aligns with the
+        // paragraph's SECOND line, so Y-grouping puts the glyph at the start
+        // of that line rather than on a line of its own. Left there it
+        // surfaces mid-sentence once the paragraph is joined — Shannon's
+        // "A Mathematical Theory of Communication" reads "...which exchange
+        // T bandwidth for signal-to-noise ratio...". Detect it, prepend the
+        // character to the paragraph's first line, and drop it from this one.
+        //
+        // The size gate is 1.8x rather than 2.5x because bitmap (Type3) caps
+        // report their glyph bbox rather than the em box, so a two-line cap
+        // can measure as little as ~1.9x the body size.
+        if line.items.len() > 1 {
+            let first = &line.items[0];
+            // The remainder must be a substantive body run: a lone label or
+            // math fragment beside a large glyph is not a drop-cap paragraph.
+            let rest_letters: usize = line.items[1..]
+                .iter()
+                .map(|i| i.text.chars().filter(|c| c.is_alphabetic()).count())
+                .sum();
+            let is_embedded_cap = first.font_size >= base_size * 1.8
+                && first.text.trim().chars().count() == 1
+                && first
+                    .text
+                    .trim()
+                    .chars()
+                    .next()
+                    .is_some_and(char::is_uppercase)
+                && line.items[1..]
+                    .iter()
+                    .all(|i| i.font_size < base_size * 1.5)
+                && line.items[1..].iter().any(|i| i.x > first.x)
+                && rest_letters >= 8;
+            if is_embedded_cap {
+                let drop_char = first.text.trim().chars().next().unwrap();
+                let cap_x = first.x;
+                let line_y = line.y;
+                // Text on the cap's own line, pushed right to clear the glyph.
+                let rest_x = line.items[1].x;
+
+                // Walk up the run of lines the cap has indented. A drop cap
+                // pushes every line it covers to the right of the glyph, so
+                // the paragraph's first line is the TOPMOST line sharing that
+                // indent — however many lines the cap spans. Using the indent
+                // rather than the cap's font size is what makes this work for
+                // three- and four-line initials as well as two-line ones;
+                // deriving a line count from the em size does not survive
+                // contact with real documents, where 36-47pt initials sit
+                // over 11-14pt leading.
+                //
+                // A cap in a different column has no such run (its neighbours
+                // sit at an unrelated x), so it is left alone — which is
+                // correct when the cap's own line already carries the rest of
+                // the word.
+                const INDENT_TOLERANCE: f32 = 2.0;
+                const MAX_CAP_LINES: usize = 8;
+                let max_step = base_size * 2.5;
+                let mut target_idx = result.len();
+                let mut expected_y = line_y;
+                while target_idx > 0 && result.len() - target_idx < MAX_CAP_LINES {
+                    let cand = &result[target_idx - 1];
+                    let step = cand.y - expected_y;
+                    let shares_indent = cand
+                        .items
+                        .first()
+                        .is_some_and(|i| (i.x - rest_x).abs() <= INDENT_TOLERANCE);
+                    if cand.page != line.page || step <= 0.0 || step > max_step || !shares_indent {
+                        break;
+                    }
+                    expected_y = cand.y;
+                    target_idx -= 1;
+                }
+
+                // The topmost line of the run is the paragraph's first line.
+                // The line above THAT tells us whether it starts a paragraph.
+                let before_target = target_idx
+                    .checked_sub(1)
+                    .and_then(|i| result.get(i))
+                    .filter(|l| l.page == line.page)
+                    .map(|l| (l.text().trim_end().to_string(), l.y));
+                // Leading within the run: the step from the target down to the
+                // next line of the paragraph, which is the cap's own line when
+                // the run is a single line.
+                let run_step = result
+                    .get(target_idx)
+                    .map(|t| {
+                        let below_y = result.get(target_idx + 1).map_or(line_y, |b| b.y);
+                        t.y - below_y
+                    })
+                    .unwrap_or(0.0);
+                let step_for_gap = if run_step > 0.0 {
+                    run_step
+                } else {
+                    base_size * 1.2
+                };
+
+                let target = (target_idx < result.len())
+                    .then(|| &mut result[target_idx])
+                    .filter(|prev| {
+                        let prev_text = prev.text();
+                        let prev_trimmed = prev_text.trim();
+                        // A hyphen on the line above means the target resumes
+                        // a split word, so it continues a paragraph rather
+                        // than starting one (polkuja_ylakoulu: "ylakou-" +
+                        // "lulaisten").
+                        //
+                        // Case cannot serve as a continuation signal here: the
+                        // target legitimately starts lowercase, because the
+                        // cap removes the word's first letter and leaves
+                        // "ver the course..." for "Over".
+                        let continues_previous = before_target
+                            .as_ref()
+                            .is_some_and(|(b, _)| b.ends_with('-'));
+                        // The target must START a paragraph: extra leading
+                        // above it, a completed sentence on the line above, or
+                        // nothing above it at all.
+                        let starts_paragraph = match before_target.as_ref() {
+                            None => true,
+                            Some((text, y)) => {
+                                y - prev.y > step_for_gap * 1.15 || ends_sentence(text)
+                            }
+                        };
+                        !continues_previous
+                            && starts_paragraph
+                            && prev.page == line.page
+                            && prev.y > line_y
+                            // Indented past the cap glyph, not merely to its
+                            // right by an arbitrary amount.
+                            && prev
+                                .items
+                                .first()
+                                .is_some_and(|i| i.x > cap_x && i.x - cap_x <= first.font_size * 2.0)
+                            // Body text, so headings, labels and table
+                            // fragments are never rewritten.
+                            && prev_trimmed
+                                .chars()
+                                .next()
+                                .is_some_and(char::is_alphabetic)
+                            && prev_trimmed.chars().filter(|c| c.is_alphabetic()).count() >= 8
+                    });
+                if let Some(prev_line) = target {
+                    if let Some(first_item) = prev_line.items.first_mut() {
+                        // A mid-word cap ("T" + "HE recent") joins directly.
+                        // Leading whitespace only marks a word boundary when
+                        // the cap is itself a single-letter word, since the
+                        // paragraph's indent can also arrive as whitespace.
+                        const SINGLE_LETTER_WORDS: &[char] = &['A', 'I', 'O', 'U', 'Y', 'E'];
+                        let had_leading_ws = first_item.text.starts_with(char::is_whitespace)
+                            && SINGLE_LETTER_WORDS.contains(&drop_char);
+                        let rest = first_item.text.trim_start().to_string();
+                        first_item.text = if had_leading_ws {
+                            format!("{} {}", drop_char, rest)
+                        } else {
+                            format!("{}{}", drop_char, rest)
+                        };
+                    }
+                    let mut line = line.clone();
+                    line.items.remove(0);
+                    result.push(line);
+                    continue;
+                }
+            }
+        }
+
         if is_drop_cap {
             let drop_char = trimmed.chars().next().unwrap();
 
@@ -598,6 +816,314 @@ mod tests {
         }
     }
 
+    fn make_item_at(text: &str, font_size: f32, x: f32) -> TextItem {
+        let mut item = make_item(text, font_size, None);
+        item.x = x;
+        item.width = text.len() as f32 * font_size * 0.5;
+        item
+    }
+
+    #[test]
+    fn embedded_drop_cap_moves_to_paragraph_start() {
+        // A two-line cap baseline-aligns with the paragraph's SECOND line,
+        // so it lands as that line's first item (Shannon entropy.pdf p.1).
+        let first_line = TextLine {
+            items: vec![make_item_at(
+                "HE recent development which exchange",
+                10.0,
+                90.0,
+            )],
+            // 16pt baseline step under a 25pt cap: a genuine two-line cap.
+            y: 716.0,
+            page: 1,
+            adaptive_threshold: 0.10,
+        };
+        let second_line = TextLine {
+            items: vec![
+                make_item_at("T", 25.0, 72.0),
+                make_item_at("bandwidth for signal-to-noise ratio", 10.0, 90.0),
+            ],
+            y: 700.0,
+            page: 1,
+            adaptive_threshold: 0.10,
+        };
+        let result = merge_drop_caps(vec![first_line, second_line], 10.0);
+        assert_eq!(result.len(), 2);
+        assert!(
+            result[0].text().starts_with("THE recent"),
+            "cap should prepend to the paragraph start: {}",
+            result[0].text()
+        );
+        assert!(
+            result[1].text().starts_with("bandwidth"),
+            "cap must be removed from the second line: {}",
+            result[1].text()
+        );
+    }
+
+    #[test]
+    fn embedded_drop_cap_walks_a_multi_line_initial_to_the_paragraph_start() {
+        // A 47pt initial over 13pt leading covers four lines, so the
+        // paragraph's first line is three lines above the cap rather than
+        // immediately above it (polkuja_ylakoulu). The indented run, not the
+        // cap's em size, is what locates it.
+        let mut lines = vec![TextLine {
+            items: vec![make_item_at("Previous paragraph ends here.", 10.0, 72.0)],
+            y: 766.0,
+            page: 1,
+            adaptive_threshold: 0.10,
+        }];
+        for (i, text) in [
+            "rilaiset mediasisallot ovat tarkea osa",
+            "useimpien ylakoululaisten elamaa ja",
+            "muuta tekstia jatkuu tassa viela",
+        ]
+        .iter()
+        .enumerate()
+        {
+            lines.push(TextLine {
+                items: vec![make_item_at(text, 10.0, 90.0)],
+                y: 753.0 - 13.0 * i as f32,
+                page: 1,
+                adaptive_threshold: 0.10,
+            });
+        }
+        lines.push(TextLine {
+            items: vec![
+                make_item_at("E", 47.0, 72.0),
+                make_item_at("loppuosa tekstista tassa", 10.0, 90.0),
+            ],
+            y: 714.0,
+            page: 1,
+            adaptive_threshold: 0.10,
+        });
+
+        let result = merge_drop_caps(lines, 10.0);
+        assert!(
+            result[1].text().starts_with("Erilaiset"),
+            "cap belongs on the topmost line of the indented run: {}",
+            result[1].text()
+        );
+        assert!(
+            result[2].text().starts_with("useimpien"),
+            "intervening run lines must be untouched: {}",
+            result[2].text()
+        );
+        assert!(
+            result[4].text().starts_with("loppuosa"),
+            "cap must be removed from its own line: {}",
+            result[4].text()
+        );
+    }
+
+    #[test]
+    fn embedded_drop_cap_ignores_non_paragraph_neighbours() {
+        // Same geometry, but the preceding line is a short label rather than
+        // body text, so it must not be rewritten.
+        let label = TextLine {
+            items: vec![make_item_at("Fig. 2", 10.0, 90.0)],
+            y: 716.0,
+            page: 1,
+            adaptive_threshold: 0.10,
+        };
+        let second_line = TextLine {
+            items: vec![
+                make_item_at("T", 25.0, 72.0),
+                make_item_at("bandwidth for signal-to-noise ratio", 10.0, 90.0),
+            ],
+            y: 700.0,
+            page: 1,
+            adaptive_threshold: 0.10,
+        };
+        let result = merge_drop_caps(vec![label, second_line], 10.0);
+        assert_eq!(result[0].text().trim(), "Fig. 2");
+        assert!(
+            result[1].text().starts_with('T'),
+            "cap stays put: {}",
+            result[1].text()
+        );
+    }
+
+    #[test]
+    fn embedded_drop_cap_keeps_a_space_for_standalone_word_caps() {
+        // Leading whitespace on the paragraph's first item marks the cap as
+        // a word of its own rather than the first letter of one.
+        let mut lead = make_item_at("long time ago in a galaxy far away", 10.0, 90.0);
+        lead.text = " long time ago in a galaxy far away".to_string();
+        let first_line = TextLine {
+            items: vec![lead],
+            y: 716.0,
+            page: 1,
+            adaptive_threshold: 0.10,
+        };
+        let second_line = TextLine {
+            items: vec![
+                make_item_at("A", 25.0, 72.0),
+                make_item_at("continued here with more body text", 10.0, 90.0),
+            ],
+            y: 700.0,
+            page: 1,
+            adaptive_threshold: 0.10,
+        };
+        let result = merge_drop_caps(vec![first_line, second_line], 10.0);
+        assert!(
+            result[0].text().starts_with("A long time ago"),
+            "standalone-word cap keeps one space: {}",
+            result[0].text()
+        );
+    }
+
+    #[test]
+    fn embedded_drop_cap_skips_hyphenation_continuation_targets() {
+        // The line above the RUN ends on a hyphen, so the run's topmost line
+        // resumes a split word rather than starting a paragraph. It sits at
+        // the paragraph margin (x=72), outside the cap's indent, so it is not
+        // part of the run itself.
+        let split_word = TextLine {
+            items: vec![make_item_at(
+                "mediasisallot ovat osa useimpien ylakou-",
+                10.0,
+                72.0,
+            )],
+            y: 728.0,
+            page: 1,
+            adaptive_threshold: 0.10,
+        };
+        let run_top = TextLine {
+            items: vec![make_item_at(
+                "lulaisten elamaa ja muuta tekstia",
+                10.0,
+                90.0,
+            )],
+            y: 714.0,
+            page: 1,
+            adaptive_threshold: 0.10,
+        };
+        let cap_line = TextLine {
+            items: vec![
+                make_item_at("E", 25.0, 72.0),
+                make_item_at("jatkuu tassa lisaa leipatekstia", 10.0, 90.0),
+            ],
+            y: 700.0,
+            page: 1,
+            adaptive_threshold: 0.10,
+        };
+        let result = merge_drop_caps(vec![split_word, run_top, cap_line], 10.0);
+        assert!(
+            result[1].text().starts_with("lulaisten"),
+            "a run resuming a split word must not receive the cap: {}",
+            result[1].text()
+        );
+        assert!(
+            result[2].text().starts_with('E'),
+            "cap stays put when no valid target exists: {}",
+            result[2].text()
+        );
+    }
+
+    #[test]
+    fn embedded_drop_cap_indent_is_not_a_word_boundary() {
+        // The paragraph's first line is indented to clear the cap, and that
+        // indent can arrive as leading whitespace. A mid-word cap must still
+        // join directly — "T HE recent" would be the defect this fixes.
+        let mut lead = make_item_at("HE recent development and more body text", 10.0, 90.0);
+        lead.text = "  HE recent development and more body text".to_string();
+        let first_line = TextLine {
+            items: vec![lead],
+            y: 716.0,
+            page: 1,
+            adaptive_threshold: 0.10,
+        };
+        let cap_line = TextLine {
+            items: vec![
+                make_item_at("T", 25.0, 72.0),
+                make_item_at("bandwidth for signal-to-noise ratio", 10.0, 90.0),
+            ],
+            y: 700.0,
+            page: 1,
+            adaptive_threshold: 0.10,
+        };
+        let result = merge_drop_caps(vec![first_line, cap_line], 10.0);
+        assert!(
+            result[0].text().starts_with("THE recent"),
+            "indent must not be read as a word boundary: {}",
+            result[0].text()
+        );
+    }
+
+    #[test]
+    fn ends_sentence_rejects_abbreviations_and_markers() {
+        use super::ends_sentence;
+        assert!(ends_sentence("This completes the thought."));
+        assert!(ends_sentence("Is that so?"));
+        assert!(ends_sentence("Stop!"));
+        // Periods that do not end a sentence.
+        assert!(!ends_sentence("as shown in Fig."));
+        assert!(!ends_sentence("see e.g."));
+        // Non-ASCII abbreviations. The two-CHARACTER segment is the case
+        // that distinguishes a character count from a byte count: "пр" is
+        // 2 chars but 4 bytes, so a byte-based bound would reject it and
+        // read the line as a completed sentence.
+        assert!(!ends_sentence("и т.пр."));
+        assert!(!ends_sentence("см. т.е."));
+        assert!(!ends_sentence("napr. ú.d."));
+        assert!(!ends_sentence("reviewed by Dr."));
+        // Standalone enumerators, any case.
+        assert!(!ends_sentence("1."));
+        assert!(!ends_sentence("IV."));
+        assert!(!ends_sentence("ii."));
+        assert!(!ends_sentence("xii."));
+        // Numbers and numerals that genuinely end a sentence must count,
+        // or a legitimate drop-cap merge is blocked.
+        assert!(ends_sentence("The paper was published in 2020."));
+        assert!(ends_sentence("He scored 5."));
+        assert!(ends_sentence("after World War II."));
+        assert!(ends_sentence("the constant equals 3.14."));
+        assert!(ends_sentence("documented at example.com."));
+        assert!(!ends_sentence("a trailing clause with no period"));
+    }
+
+    #[test]
+    fn embedded_drop_cap_allows_first_paragraph_on_a_new_page() {
+        // The line two back is on the previous page, so its y is unrelated
+        // and must not be used as leading evidence.
+        let prev_page_tail = TextLine {
+            items: vec![make_item_at(
+                "tail of the previous page body text",
+                10.0,
+                90.0,
+            )],
+            y: 90.0,
+            page: 1,
+            adaptive_threshold: 0.10,
+        };
+        let first_line = TextLine {
+            items: vec![make_item_at(
+                "HE recent development which exchange",
+                10.0,
+                90.0,
+            )],
+            y: 716.0,
+            page: 2,
+            adaptive_threshold: 0.10,
+        };
+        let cap_line = TextLine {
+            items: vec![
+                make_item_at("T", 25.0, 72.0),
+                make_item_at("bandwidth for signal-to-noise ratio", 10.0, 90.0),
+            ],
+            y: 700.0,
+            page: 2,
+            adaptive_threshold: 0.10,
+        };
+        let result = merge_drop_caps(vec![prev_page_tail, first_line, cap_line], 10.0);
+        assert!(
+            result[1].text().starts_with("THE recent"),
+            "a page break must not suppress the merge: {}",
+            result[1].text()
+        );
+    }
+
     #[test]
     fn test_merge_struct_tree_headings() {
         // Two consecutive lines tagged as H2 via struct tree, same font size as body

src/tables/detect_heuristic.rs

@@ -435,6 +435,72 @@ fn revised_table_cell_indices(
         .collect()
 }
 
+/// Index of candidate "body" items (larger-font attachment targets) sorted by
+/// Y, so script-attachment checks scan a narrow Y window instead of the whole
+/// page per candidate.
+struct ScriptBodyIndex<'a> {
+    /// (y, item), sorted ascending by y
+    by_y: Vec<(f32, &'a TextItem)>,
+    /// widest vertical attachment window any body item can produce
+    max_window: f32,
+}
+
+impl<'a> ScriptBodyIndex<'a> {
+    fn new(items: &'a [TextItem]) -> Self {
+        // Smallest table-candidate font is 6pt, so any possible attachment
+        // target is at least 6 x 1.2 pt.
+        let mut by_y: Vec<(f32, &TextItem)> = items
+            .iter()
+            .filter(|i| i.font_size >= 6.0 * 1.2)
+            .map(|i| (i.y, i))
+            .collect();
+        by_y.sort_by(|a, b| a.0.total_cmp(&b.0));
+        let max_window = by_y
+            .iter()
+            .map(|(_, i)| i.font_size * 0.8)
+            .fold(0.0f32, f32::max);
+        Self { by_y, max_window }
+    }
+
+    /// True when a small-font item is horizontally attached to a larger-font
+    /// item at a script baseline offset — a sub/superscript in running text
+    /// or math (equation subscripts, footnote markers). Script attachments
+    /// are not table cells; without this filter, display equations with
+    /// sub/superscripts form phantom small-font table regions (e.g. TeX
+    /// papers where log subscripts cluster with footnote lines into a fake
+    /// 3-column table). A genuine baseline offset is required so same-line
+    /// table neighbours (a small cell beside a larger label cell) are never
+    /// classified as scripts.
+    ///
+    /// `min_anchor_size` additionally constrains what counts as an
+    /// attachment target: the small-font pass accepts any sufficiently
+    /// larger item (0.0), while the body-font pass requires a heading-sized
+    /// anchor so a body-size table cell beside a slightly larger label with
+    /// baseline jitter is never treated as a script.
+    fn is_script_attachment(&self, small: &TextItem, min_anchor_size: f32) -> bool {
+        let attach_gap = small.font_size.max(4.0) * 0.6;
+        let lo = self
+            .by_y
+            .partition_point(|(y, _)| *y < small.y - self.max_window);
+        self.by_y[lo..]
+            .iter()
+            .take_while(|(y, _)| *y <= small.y + self.max_window)
+            .any(|(_, body)| {
+                let dy = (small.y - body.y).abs();
+                body.font_size >= small.font_size * 1.2
+                    && body.font_size >= min_anchor_size
+                    && dy > body.font_size * 0.05
+                    && dy <= body.font_size * 0.8
+                    && {
+                        let gap_after_body = small.x - (body.x + body.width);
+                        let gap_before_body = body.x - (small.x + small.width);
+                        (-attach_gap..=attach_gap).contains(&gap_after_body)
+                            || (-attach_gap..=attach_gap).contains(&gap_before_body)
+                    }
+            })
+    }
+}
+
 /// Detect tables in a set of text items from a single page
 pub fn detect_tables(items: &[TextItem], base_font_size: f32, skip_body_font: bool) -> Vec<Table> {
     detect_tables_with_page_width(items, base_font_size, skip_body_font, content_width(items))
@@ -483,6 +549,27 @@ pub(crate) fn detect_tables_with_page_width(
     // === Pass 1: Small-font tables (existing behavior) ===
     let table_font_threshold = base_font_size * 0.90;
 
+    // Mark sub/superscript attachments once per pass. They stay candidates —
+    // the masks only remove them from region qualification and column/row
+    // geometry.
+    //
+    // The two passes need different anchor thresholds. In the small-font pass
+    // any sufficiently larger neighbour is a plausible base for a script. In
+    // the body-font pass the candidates are themselves body-sized
+    // (0.85..1.05x), so a merely "slightly larger" neighbour is usually a bold
+    // label or an adjacent column header, not the base of a superscript —
+    // treating it as one would strip real cells out of the geometry and lose
+    // the table. Requiring a heading-sized anchor (>= 1.15x base) keeps the
+    // body pass to genuine scripts hanging off headings.
+    let script_index = ScriptBodyIndex::new(items);
+    let script_flags: Vec<bool> = items
+        .iter()
+        .map(|item| script_index.is_script_attachment(item, 0.0))
+        .collect();
+    let body_script_flags: Vec<bool> = items
+        .iter()
+        .map(|item| script_index.is_script_attachment(item, base_font_size * 1.15))
+        .collect();
     let table_candidates: Vec<(usize, &TextItem)> = items
         .iter()
         .enumerate()
@@ -494,7 +581,14 @@ pub(crate) fn detect_tables_with_page_width(
         .collect();
 
     if table_candidates.len() >= 6 {
-        let regions = find_table_regions(&table_candidates);
+        // Qualify regions from non-script items: a cluster of sub/superscripts
+        // must not, on its own, mark out a table region.
+        let region_evidence: Vec<(usize, &TextItem)> = table_candidates
+            .iter()
+            .filter(|(idx, _)| !script_flags[*idx])
+            .cloned()
+            .collect();
+        let regions = find_table_regions(&region_evidence);
 
         for (y_min, y_max) in regions {
             let region_items: Vec<(usize, &TextItem)> = table_candidates
@@ -508,7 +602,9 @@ pub(crate) fn detect_tables_with_page_width(
             }
 
             if let Some(mut table) =
-                detect_table_in_region(&region_items, TableDetectionMode::SmallFont)
+                detect_table_in_region(&region_items, TableDetectionMode::SmallFont, &|i| {
+                    script_flags[i]
+                })
             {
                 // Try to recover body-font header row above the small-font table
                 recover_header_row(&mut table, items, table_font_threshold);
@@ -553,8 +649,20 @@ pub(crate) fn detect_tables_with_page_width(
             body_font_low,
             body_font_high,
         );
+        // Scripts are NOT filtered out of the candidate set here, mirroring
+        // the small-font pass: they must stay eligible for cell assignment so
+        // a sub/superscript that belongs inside a table cell keeps its text.
+        // The heading-anchored `body_script_flags` mask removes them from
+        // geometry only.
         if body_candidates.len() >= 6 {
-            let regions = find_table_regions_strict(&body_candidates);
+            // Same reasoning as the small-font pass: scripts do not qualify
+            // regions, but remain available for cell assignment within one.
+            let region_evidence: Vec<(usize, &TextItem)> = body_candidates
+                .iter()
+                .filter(|(idx, _)| !body_script_flags[*idx])
+                .cloned()
+                .collect();
+            let regions = find_table_regions_strict(&region_evidence);
             log::debug!("body-font: {} strict regions found", regions.len());
 
             for (y_min, y_max, _x_min, _x_max) in &regions {
@@ -580,7 +688,9 @@ pub(crate) fn detect_tables_with_page_width(
                 }
 
                 if let Some(table) =
-                    detect_table_in_region(&region_items, TableDetectionMode::BodyFont)
+                    detect_table_in_region(&region_items, TableDetectionMode::BodyFont, &|i| {
+                        body_script_flags[i]
+                    })
                 {
                     tables.push(table);
                 }
@@ -808,10 +918,30 @@ fn find_table_regions_strict(items: &[(usize, &TextItem)]) -> Vec<(f32, f32, f32
     regions
 }
 
-/// Detect a table within a specific region
-fn detect_table_in_region(items: &[(usize, &TextItem)], mode: TableDetectionMode) -> Option<Table> {
-    // Find column boundaries
-    let columns = find_column_boundaries(items, mode);
+/// Detect a table within a specific region.
+///
+/// `is_script` marks items that are sub/superscript attachments. Those are
+/// excluded from the *geometry* — they must not be able to create a column,
+/// which is how equation subscript clusters used to fabricate phantom grids —
+/// but they remain eligible for cell assignment, so legitimate cell content
+/// (exponents in an engineering-notation table, footnote markers) stays in
+/// the cell it belongs to instead of leaking out into the reading order.
+fn detect_table_in_region(
+    items: &[(usize, &TextItem)],
+    mode: TableDetectionMode,
+    is_script: &dyn Fn(usize) -> bool,
+) -> Option<Table> {
+    // Column geometry from non-script items only.
+    let geometry_items: Vec<(usize, &TextItem)> = items
+        .iter()
+        .filter(|(idx, _)| !is_script(*idx))
+        .cloned()
+        .collect();
+    // A region that is *entirely* scripts has no table structure at all.
+    if geometry_items.is_empty() {
+        return None;
+    }
+    let columns = find_column_boundaries(&geometry_items, mode);
     let min_cols = 2;
     if columns.len() < min_cols || columns.len() > 25 {
         log::debug!(
@@ -822,8 +952,8 @@ fn detect_table_in_region(items: &[(usize, &TextItem)], mode: TableDetectionMode
         return None;
     }
 
-    // Find row boundaries
-    let rows = find_row_boundaries(items);
+    // Find row boundaries (geometry items only, same reasoning)
+    let rows = find_row_boundaries(&geometry_items);
     let min_rows = 2;
     if rows.len() < min_rows {
         log::debug!(
@@ -842,6 +972,11 @@ fn detect_table_in_region(items: &[(usize, &TextItem)], mode: TableDetectionMode
     );
 
     // Verify this looks like a table: multiple items should align to columns
+    // Validate against ALL items, including scripts. Columns are derived from
+    // non-script geometry so scripts cannot *create* a column, but excluding
+    // them from validation too would let a region manufacture alignment: drop
+    // the awkward items and whatever remains looks like a tidy grid. Block
+    // diagrams did exactly that. Everything in the region must fit.
     let col_alignment = check_column_alignment(items, &columns, mode);
     let min_alignment = match mode {
         TableDetectionMode::SmallFont => 0.5,
@@ -912,6 +1047,29 @@ fn detect_table_in_region(items: &[(usize, &TextItem)], mode: TableDetectionMode
         cells.push(row_cells);
     }
 
+    // Validation 0 (small-font pass only): reject tiny all-numeric
+    // fragments. A <=2-row grid whose every cell is a bare 1-2 digit number
+    // carries no tabular information — in practice these are
+    // exponent/subscript clusters from display math that happen to align.
+    // Body-font tables are not subject to this veto: their cells cannot be
+    // script glyphs.
+    if matches!(mode, TableDetectionMode::SmallFont) {
+        let nonempty_cells: Vec<&String> =
+            cells.iter().flatten().filter(|c| !c.is_empty()).collect();
+        if rows.len() <= 2
+            && !nonempty_cells.is_empty()
+            && nonempty_cells
+                .iter()
+                .all(|c| c.len() <= 2 && c.chars().all(|ch| ch.is_ascii_digit()))
+        {
+            log::debug!(
+                "  validation 0 fail: tiny all-numeric fragment ({} cells)",
+                nonempty_cells.len()
+            );
+            return None;
+        }
+    }
+
     // Validation 1: some rows should have content in first column.
     // Use a lower threshold (25%) for tables with wrapped cells where
     // continuation lines leave the first column empty.
@@ -1977,6 +2135,146 @@ fn try_add_label_column(
 
 #[cfg(test)]
 mod tests {
+
+    fn make_item(text: &str, x: f32, y: f32, font_size: f32, width: f32) -> TextItem {
+        TextItem {
+            text: text.to_string(),
+            x,
+            y,
+            width,
+            height: font_size,
+            font: "TestFont".to_string(),
+            font_size,
+            page: 1,
+            is_bold: false,
+            is_italic: false,
+            is_underline: false,
+            is_strikeout: false,
+            item_type: ItemType::Text,
+            mcid: None,
+        }
+    }
+
+    #[test]
+    fn script_attachment_detects_subscript_after_body_text() {
+        let body = make_item("log", 100.0, 500.0, 10.0, 15.0);
+        let sub = make_item("10", 115.5, 497.0, 7.0, 7.0);
+        let items = vec![body, sub.clone()];
+        assert!(ScriptBodyIndex::new(&items).is_script_attachment(&sub, 0.0));
+    }
+
+    #[test]
+    fn script_attachment_detects_superscript_footnote_marker() {
+        let body = make_item("Hartley", 200.0, 500.0, 10.0, 35.0);
+        let sup = make_item("2", 235.8, 504.0, 6.6, 3.5);
+        let items = vec![body, sup.clone()];
+        assert!(ScriptBodyIndex::new(&items).is_script_attachment(&sup, 0.0));
+    }
+
+    #[test]
+    fn script_attachment_ignores_small_cell_far_from_body_text() {
+        let body = make_item("Revenue", 100.0, 500.0, 10.0, 40.0);
+        let cell = make_item("1,234", 180.0, 500.0, 7.0, 20.0);
+        let items = vec![body, cell.clone()];
+        assert!(!ScriptBodyIndex::new(&items).is_script_attachment(&cell, 0.0));
+    }
+
+    #[test]
+    fn body_pass_anchor_spares_cells_beside_slightly_larger_labels() {
+        // A body-font table cell (10pt) sitting beside a slightly larger,
+        // NON-heading label (12.5pt) with a little baseline jitter. The
+        // small-font pass treats any larger neighbour as a possible script
+        // base, but the body pass must not: at body sizes a slightly larger
+        // neighbour is a bold label or column header, and flagging the cell
+        // would strip it out of the table geometry and lose the table.
+        // Cell at the low end of the body band (0.85x base) beside a 10.5pt
+        // label. 10.5 clears the inherent 1.2x-of-cell rule (10.2) but falls
+        // below the body pass's heading anchor (11.5), which is exactly the
+        // band where the two masks must disagree.
+        let label = make_item("Revenue", 100.0, 500.0, 10.5, 40.0);
+        let cell = make_item("1,234", 141.0, 496.5, 8.5, 22.0);
+        let items = vec![label, cell.clone()];
+        let index = ScriptBodyIndex::new(&items);
+        let base = 10.0;
+        assert!(
+            index.is_script_attachment(&cell, 0.0),
+            "small-font pass anchor should still see this as an attachment"
+        );
+        assert!(
+            !index.is_script_attachment(&cell, base * 1.15),
+            "body pass must not treat a cell beside a slightly larger label \
+             as a script — that removes real cells from the geometry"
+        );
+        // A genuine heading-sized anchor still qualifies in the body pass.
+        let heading = make_item("Section", 100.0, 500.0, 20.0, 60.0);
+        let sup = make_item("3", 161.0, 508.0, 10.0, 5.0);
+        let h_items = vec![heading, sup.clone()];
+        assert!(
+            ScriptBodyIndex::new(&h_items).is_script_attachment(&sup, base * 1.15),
+            "script hanging off a heading must still be excluded in the body pass"
+        );
+    }
+
+    #[test]
+    fn script_attachment_ignores_same_baseline_neighbor_cell() {
+        // A small cell beside a larger label on the SAME baseline is a table
+        // layout, not a subscript — a genuine baseline offset is required.
+        let label = make_item("Total", 100.0, 500.0, 10.0, 25.0);
+        let cell = make_item("42", 127.0, 500.0, 7.5, 9.0);
+        let items = vec![label, cell.clone()];
+        assert!(!ScriptBodyIndex::new(&items).is_script_attachment(&cell, 0.0));
+    }
+
+    #[test]
+    fn script_attachment_ignores_neighbor_on_different_line() {
+        let body = make_item("Header", 100.0, 500.0, 10.0, 30.0);
+        let cell = make_item("42", 131.0, 486.0, 7.0, 10.0);
+        let items = vec![body, cell.clone()];
+        assert!(!ScriptBodyIndex::new(&items).is_script_attachment(&cell, 0.0));
+    }
+
+    /// Equation-subscript + footnote layout from Shannon entropy.pdf page 1,
+    /// with real coordinates. Without the larger-font anchors the small items
+    /// alone DO form a phantom table — proving the layout reaches detection —
+    /// and adding the anchors must suppress it.
+    fn shannon_page1_small_items() -> Vec<TextItem> {
+        vec![
+            make_item("2", 267.4, 133.9, 7.4, 3.7),
+            make_item("10", 306.2, 133.9, 7.4, 7.4),
+            make_item("10", 342.7, 133.9, 7.4, 7.4),
+            make_item("10", 325.0, 118.9, 7.4, 7.4),
+            make_item("Bell System Technical Journal,", 295.7, 101.9, 8.0, 95.0),
+            make_item(
+                "April 1924, p. 324; Certain Topics in",
+                396.7,
+                101.9,
+                8.0,
+                130.0,
+            ),
+            make_item("v. 47, April 1928, p. 617.", 250.9, 92.5, 8.0, 90.0),
+            make_item("Bell System Technical Journal,", 264.2, 82.6, 8.0, 95.0),
+            make_item("July 1928, p. 535.", 364.3, 82.6, 8.0, 65.0),
+        ]
+    }
+
+    #[test]
+    fn equation_scripts_do_not_form_phantom_table() {
+        let bare = shannon_page1_small_items();
+        assert!(
+            !detect_tables(&bare, 10.0, false).is_empty(),
+            "test layout must form a phantom table when the filter cannot fire"
+        );
+        let mut items = shannon_page1_small_items();
+        items.push(make_item("log", 253.0, 137.0, 10.0, 13.5));
+        items.push(make_item("log", 291.5, 137.0, 10.0, 13.5));
+        items.push(make_item("log", 328.0, 137.0, 10.0, 13.5));
+        items.push(make_item("log", 310.3, 122.0, 10.0, 13.5));
+        let tables = detect_tables(&items, 10.0, false);
+        assert!(
+            tables.is_empty(),
+            "equation scripts + footnotes must not become a table: {tables:?}"
+        );
+    }
     use super::*;
     use crate::types::ItemType;
 

tests/snapshots/shannon-entropy-p1-2.md

@@ -1,16 +1,12 @@
 Reprinted with corrections from *The Bell System Technical Journal,* Vol. 27, pp. 379–423, 623–656, July, October, 1948.
 
-## A Mathematical Theory of Communication
+# A Mathematical Theory of Communication
 
-### By C. E. SHANNON
+## By C. E. SHANNON
 
 INTRODUCTION
 
-HE recent development of various methods of modulation such as PCM and PPM which exchange
-
-# Tbandwidth for signal-to-noise ratio has intensified the interest in a general theory of communication. A
-
-basis for such a theory is contained in the important papers of Nyquist¹ and Hartley² on this subject. In the present paper we will extend the theory to include a number of new factors, in particular the effect of noise in the channel, and the savings possible due to the statistical structure of the original message and due to the nature of the final destination of the information. The fundamental problem of communication is that of reproducing at one point either exactly or ap- proximately a message selected at another point. Frequently the messages have *meaning*; that is they refer to or are correlated according to some system with certain physical or conceptual entities. These semantic aspects of communication are irrelevant to the engineering problem. The significant aspect is that the actual message is one *selected from a set* of possible messages. The system must be designed to operate for each possible selection, not just the one which will actually be chosen since this is unknown at the time of design. If the number of messages in the set is finite then this number or any monotonic function of this number can be regarded as a measure of the information produced when one message is chosen from the set, all choices being equally likely. As was pointed out by Hartley the most natural choice is the logarithmic function. Although this definition must be generalized considerably when we consider the influence of the statistics of the message and when we have a continuous range of messages, we will in all cases use an essentially logarithmic measure. The logarithmic measure is more convenient for various reasons:
+THE recent development of various methods of modulation such as PCM and PPM which exchange bandwidth for signal-to-noise ratio has intensified the interest in a general theory of communication. A basis for such a theory is contained in the important papers of Nyquist¹ and Hartley² on this subject. In the present paper we will extend the theory to include a number of new factors, in particular the effect of noise in the channel, and the savings possible due to the statistical structure of the original message and due to the nature of the final destination of the information. The fundamental problem of communication is that of reproducing at one point either exactly or ap- proximately a message selected at another point. Frequently the messages have *meaning*; that is they refer to or are correlated according to some system with certain physical or conceptual entities. These semantic aspects of communication are irrelevant to the engineering problem. The significant aspect is that the actual message is one *selected from a set* of possible messages. The system must be designed to operate for each possible selection, not just the one which will actually be chosen since this is unknown at the time of design. If the number of messages in the set is finite then this number or any monotonic function of this number can be regarded as a measure of the information produced when one message is chosen from the set, all choices being equally likely. As was pointed out by Hartley the most natural choice is the logarithmic function. Although this definition must be generalized considerably when we consider the influence of the statistics of the message and when we have a continuous range of messages, we will in all cases use an essentially logarithmic measure. The logarithmic measure is more convenient for various reasons:
 
 1. It is practically more useful. Parameters of engineering importance such as time, bandwidth, number of relays, etc., tend to vary linearly with the logarithm of the number of possibilities. For example, adding one relay to a group doubles the number of possible states of the relays. It adds 1 to the base 2 logarithm of this number. Doubling the time roughly squares the number of possible messages, or doubles the logarithm, etc.
 2. It is nearer to our intuitive feeling as to the proper measure. This is closely related to (1) since we in- tuitively measures entities by linear comparison with common standards. One feels, for example, that two punched cards should have twice the capacity of one for information storage, and two identical channels twice the capacity of one for transmitting information.