citefind: a colon in running prose is not an imprint

Thirdly, not all fields of science and technology were seen as equally
    relevant to the nation's war effort. Bush himself was an electrical
    engineer… And finally, growing out of the OSRD's experience, Bush
    proposed — most famously in his 1945 report, Science: The Endless
    Frontier — that Washington continue to play a leading role…

Six sentences of body text, stored as a citation whose title was its first
clause. The check that refuses a note for running on made an exception for
anything carrying an imprint, and `his 1945 report, Science: The Endless
Frontier` is shaped exactly like `Place: Publisher`. One colon anywhere in a
paragraph bought the whole paragraph a pass.

A page reference or an identifier still buys the exemption — a note that
cites a work and then discusses it is a citation. An imprint no longer does.
And an entry past a hundred and twenty words is prose whatever its
punctuation: the longest real entry in the corpus sample is under sixty.

crates/citefind/src/parse.rs

@@ -245,6 +245,14 @@ static CITATION_ANCHOR: LazyLock<Regex> = LazyLock::new(|| {
     Regex::new(r"(?i)\bpp?\.\s*\d|\bvol\.\s*\d|\bno\.\s*\d").expect("a valid pattern")
 });
 
+/// How many words a reference runs to before it is prose.
+///
+/// The longest real entry in the corpus sample — a chapter in an edited
+/// volume, with two titles, an editor, an imprint and a page range — is
+/// under sixty. A hundred and twenty is generous and still nowhere near a
+/// paragraph of body text.
+const MAX_WORDS: usize = 120;
+
 /// How many sentences a reference runs to before it is prose.
 ///
 /// A full entry in the heaviest style — names, title, an edited volume, an
@@ -379,13 +387,18 @@ pub fn one(text: &str, marker: Option<&str>) -> Option<Reference> {
         out.authors = short_cite_authors(text);
         return Some(out);
     }
-    // A note that runs on for six sentences with no imprint, no page
-    // reference and no identifier is a remark about the argument. The LLM
-    // path this replaces was told to skip those; a parser has to decide.
-    if sentences(text).len() > MAX_SENTENCES
-        && !IMPRINT.is_match(text)
-        && !CITATION_ANCHOR.is_match(text)
-        && ids(text).is_empty()
+    // A note that runs on past a few sentences, or past a couple of hundred
+    // words, is a remark about the argument. The LLM path this replaces was
+    // told to skip those; a parser has to decide.
+    //
+    // A page reference or an identifier buys an exemption, because a note
+    // that cites a work and then discusses it is still a citation. An
+    // imprint does not: a colon in running prose — `his 1945 report,
+    // Science: The Endless Frontier` — is shaped exactly like `Place:
+    // Publisher`, and used to let whole paragraphs of body text through.
+    let exempt = CITATION_ANCHOR.is_match(text) || !ids(text).is_empty();
+    if !exempt
+        && (sentences(text).len() > MAX_SENTENCES || text.split_whitespace().count() > MAX_WORDS)
     {
         return None;
     }
@@ -1779,6 +1792,36 @@ mod tests {
         assert_eq!(found.pages.as_deref(), Some("365"));
     }
 
+    #[test]
+    fn a_colon_in_running_prose_is_not_an_imprint() {
+        // Six sentences of body text, with `his 1945 report, Science: The
+        // Endless Frontier` in the middle of it — shaped exactly like
+        // `Place: Publisher`, which used to buy the whole paragraph an
+        // exemption from the prose test.
+        let prose = "Thirdly, not all fields of science were seen as equally \
+            relevant to the war effort. Bush himself was an electrical \
+            engineer; his chief lieutenants were drawn from the physical \
+            sciences. Weapons such as radar drew heavily upon such \
+            disciplines. OSRD was busily creating an electronic environment \
+            for war. At any rate, physics emerged from the war as the most \
+            prestigious discipline. And finally, Bush proposed - most \
+            famously in his 1945 report, Science: The Endless Frontier - \
+            that Washington continue to play a leading role.";
+        assert!(one(prose, None).is_none(), "{prose}");
+
+        // A long entry that cites something is still a citation: the page
+        // range is what buys the exemption, not the colon.
+        let chapter = read(
+            "Heder, Stephen P. 'The Kampuchean-Vietnamese Conflict.' In David \
+             W.P. Elliott, ed. *The Third Indochina Conflict.* Boulder: \
+             Westview Press. 1981. pp. 21-67.",
+        );
+        assert_eq!(
+            chapter.title.as_deref(),
+            Some("The Third Indochina Conflict")
+        );
+    }
+
     #[test]
     fn a_remark_about_the_argument_is_not_a_reference() {
         let found = one(