Irregular verbs

harper-core/src/linting/irregular_verbs.json

@@ -0,0 +1,127 @@
+[
+	"// comments can appear in the line before an entry",
+	"// or in place of an entry",
+	["arise", "arose", "arisen"],
+	["awake", "awoke", "awoken"],
+	"// be/am/are/is -- was/were -- been",
+	["become", "became", "become"],
+	["begin", "began", "begun"],
+	["bend", "bent", "bent"],
+	["bet", "bet", "bet"],
+	["bid", "bade", "bidden"],
+	["bind", "bound", "bound"],
+	["bite", "bit", "bitten"],
+	["bleed", "bled", "bled"],
+	["blow", "blew", "blown"],
+	["break", "broke", "broken"],
+	["breed", "bred", "bred"],
+	["bring", "brought", "brought"],
+	["build", "built", "built"],
+	["burst", "burst", "burst"],
+	["buy", "bought", "bought"],
+	["catch", "caught", "caught"],
+	["choose", "chose", "chosen"],
+	["come", "came", "come"],
+	["cost", "cost", "cost"],
+	["cut", "cut", "cut"],
+	["dive", "dove", "dove"],
+	["do", "did", "done"],
+	["drink", "drank", "drunk"],
+	["drive", "drove", "driven"],
+	["eat", "ate", "eaten"],
+	["fall", "fell", "fallen"],
+	["feed", "fed", "fed"],
+	["feel", "felt", "felt"],
+	["fight", "fought", "fought"],
+	["find", "found", "found"],
+	["fly", "flew", "flown"],
+	["forget", "forgot", "forgotten"],
+	["forgo", "forwent", "forgone"],
+	["freeze", "froze", "frozen"],
+	"// get -- got -- gotten",
+	["get", "got", "got"],
+	["give", "gave", "given"],
+	["go", "went", "gone"],
+	["grow", "grew", "grown"],
+	["have", "had", "had"],
+	["hear", "heard", "heard"],
+	["hit", "hit", "hit"],
+	["hold", "held", "held"],
+	["hurt", "hurt", "hurt"],
+	["input", "input", "input"],
+	["keep", "kept", "kept"],
+	["know", "knew", "known"],
+	["lay", "laid", "lain"],
+	["lead", "led", "led"],
+	["light", "lit", "lit"],
+	["lose", "lost", "lost"],
+	["make", "made", "made"],
+	["mistake", "mistook", "mistaken"],
+	["output", "output", "output"],
+	["overtake", "overtook", "overtaken"],
+	["overthrow", "overthrew", "overthrown"],
+	["overwrite", "overwrote", "overwritten"],
+	["partake", "partook", "partaken"],
+	["pay", "paid", "paid"],
+	["put", "put", "put"],
+	["read", "read", "read"],
+	["redo", "redid", "redone"],
+	["remake", "remade", "remade"],
+	["reread", "reread", "reread"],
+	["reset", "reset", "reset"],
+	["ride", "rode", "ridden"],
+	["ring", "rang", "rung"],
+	["rise", "rose", "risen"],
+	["run", "ran", "run"],
+	["see", "saw", "seen"],
+	["sell", "sold", "sold"],
+	["send", "sent", "sent"],
+	["set", "set", "set"],
+	["shake", "shook", "shaken"],
+	["shed", "shed", "shed"],
+	["shine", "shone", "shone"],
+	["shoe", "shod", "shod"],
+	["shoot", "shot", "shot"],
+	["show", "showed", "shown"],
+	["shrink", "shrank", "shrunk"],
+	["shut", "shut", "shut"],
+	["sing", "sang", "sung"],
+	"// sink -- sank -- sunken??",
+	["sink", "sank", "sunk"],
+	["sit", "sat", "sat"],
+	["slay", "slew", "slain"],
+	["sleep", "slept", "slept"],
+	["slide", "slid", "slid"],
+	["slit", "slit", "slit"],
+	"// sneak -- sneaked/snuck -- sneaked/snuck",
+	["speak", "spoke", "spoken"],
+	["spin", "spun", "spun"],
+	["spit", "spat", "spat"],
+	["split", "split", "split"],
+	["spread", "spread", "spread"],
+	["spring", "sprang", "sprung"],
+	["stand", "stood", "stood"],
+	["steal", "stole", "stolen"],
+	["stick", "stuck", "stuck"],
+	["sting", "stung", "stung"],
+	["stink", "stank", "stunk"],
+	["stride", "strode", "stridden"],
+	["strike", "struck", "stricken"],
+	["string", "strung", "strung"],
+	["sew", "sewed", "sewn"],
+	["swear", "swore", "sworn"],
+	["swim", "swam", "swum"],
+	["swing", "swung", "swung"],
+	["take", "took", "taken"],
+	["teach", "taught", "taught"],
+	["tear", "tore", "torn"],
+	["think", "thought", "thought"],
+	["throw", "threw", "thrown"],
+	["tread", "trod", "trodden"],
+	["undo", "undid", "undone"],
+	["wake", "woke", "woken"],
+	["wear", "wore", "worn"],
+	["weave", "wove", "woven"],
+	["wind", "wound", "wound"],
+	["write", "wrote", "written"]
+]

harper-core/src/linting/irregular_verbs.rs

@@ -0,0 +1,120 @@
+use lazy_static::lazy_static;
+use serde::Deserialize;
+use std::sync::Arc;
+
+type Verb = (String, String, String);
+
+#[derive(Debug, Deserialize)]
+pub struct IrregularVerbs {
+    verbs: Vec<Verb>,
+}
+
+/// The uncached function that is used to produce the original copy of the
+/// irregular verb table.
+fn uncached_inner_new() -> Arc<IrregularVerbs> {
+    IrregularVerbs::from_json_file(include_str!("irregular_verbs.json"))
+        .map(Arc::new)
+        .unwrap_or_else(|e| panic!("Failed to load irregular verb table: {}", e))
+}
+
+lazy_static! {
+    static ref VERBS: Arc<IrregularVerbs> = uncached_inner_new();
+}
+
+impl IrregularVerbs {
+    pub fn new() -> Self {
+        Self { verbs: vec![] }
+    }
+
+    pub fn from_json_file(json: &str) -> Result<Self, serde_json::Error> {
+        // Deserialize into Vec<serde_json::Value> to handle mixed types
+        let values: Vec<serde_json::Value> =
+            serde_json::from_str(json).expect("Failed to parse irregular verbs JSON");
+
+        let mut verbs = Vec::new();
+
+        for value in values {
+            match value {
+                serde_json::Value::Array(arr) if arr.len() == 3 => {
+                    // Handle array of 3 strings
+                    if let (Some(lemma), Some(preterite), Some(past_participle)) =
+                        (arr[0].as_str(), arr[1].as_str(), arr[2].as_str())
+                    {
+                        verbs.push((
+                            lemma.to_string(),
+                            preterite.to_string(),
+                            past_participle.to_string(),
+                        ));
+                    }
+                }
+                // Strings are used for comments to guide contributors editing the file
+                serde_json::Value::String(_) => {}
+                _ => {}
+            }
+        }
+
+        Ok(Self { verbs })
+    }
+
+    pub fn get() -> Arc<Self> {
+        (*VERBS).clone()
+    }
+
+    pub fn get_past_participle_for_preterite(&self, preterite: &str) -> Option<&str> {
+        self.verbs
+            .iter()
+            .find(|(_, pt, _)| pt.eq_ignore_ascii_case(preterite))
+            .map(|(_, _, pp)| pp.as_str())
+    }
+}
+
+impl Default for IrregularVerbs {
+    fn default() -> Self {
+        Self::new()
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn can_find_irregular_past_participle_for_preterite_lowercase() {
+        assert_eq!(
+            IrregularVerbs::get().get_past_participle_for_preterite("arose"),
+            Some("arisen")
+        );
+    }
+
+    #[test]
+    fn can_find_irregular_past_participle_for_preterite_uppercase() {
+        assert_eq!(
+            IrregularVerbs::get().get_past_participle_for_preterite("WENT"),
+            Some("gone")
+        );
+    }
+
+    #[test]
+    fn can_find_irregular_past_participle_same_as_past_tense() {
+        assert_eq!(
+            IrregularVerbs::get().get_past_participle_for_preterite("taught"),
+            Some("taught")
+        );
+    }
+
+    #[test]
+    fn cant_find_regular_past_participle() {
+        assert_eq!(
+            IrregularVerbs::get().get_past_participle_for_preterite("walked"),
+            None
+        );
+    }
+
+    #[test]
+    fn cant_find_non_verb() {
+        assert_eq!(
+            IrregularVerbs::get().get_past_participle_for_preterite("the"),
+            None
+        );
+    }
+}

harper-core/src/linting/mod.rs

@@ -78,6 +78,7 @@ mod inflected_verb_after_to;
 mod initialism_linter;
 mod initialisms;
 mod interested_in;
+mod irregular_verbs;
 mod it_is;
 mod it_looks_like_that;
 mod it_would_be;
@@ -201,6 +202,7 @@ mod would_never_have;
 
 pub use expr_linter::ExprLinter;
 pub use initialism_linter::InitialismLinter;
+pub use irregular_verbs::IrregularVerbs;
 pub use lint::Lint;
 pub use lint_group::{LintGroup, LintGroupConfig};
 pub use lint_kind::LintKind;