Merge pull request #764 from not-my-profile/refactor-generate

Refactor generate and add tests

crates/typos-dict/tests/verify.rs

@@ -9,38 +9,76 @@ type Dict = BTreeMap<UniCase<String>, BTreeSet<String>>;
 #[test]
 fn verify() {
     let asset_path = "assets/words.csv";
-    let data = std::fs::read(asset_path).unwrap();
+    let typos_dict = parse_dict(asset_path);
+    let new_dict = process(typos_dict);
 
     let mut content = vec![];
-    generate(&mut content, &data);
+
+    let mut wtr = csv::WriterBuilder::new()
+        .flexible(true)
+        .from_writer(&mut content);
+    for (typo, corrections) in new_dict {
+        let mut row = vec![typo.as_str().to_owned()];
+        row.extend(corrections);
+        wtr.write_record(&row).unwrap();
+    }
+    wtr.flush().unwrap();
+    drop(wtr);
 
     let content = String::from_utf8(content).unwrap();
     snapbox::assert_eq_path(asset_path, content);
 }
 
-fn generate<W: std::io::Write>(file: &mut W, dict: &[u8]) {
-    let mut rows = Dict::new();
-    csv::ReaderBuilder::new()
+fn parse_dict(path: &str) -> Vec<(String, Vec<String>)> {
+    let data = std::fs::read(path).unwrap();
+
+    let mut reader = csv::ReaderBuilder::new()
         .has_headers(false)
         .flexible(true)
-        .from_reader(dict)
-        .records()
-        .map(Result::unwrap)
-        .for_each(|r| {
-            let mut i = r.iter();
-            let mut typo = i.next().expect("typo").to_owned();
-            typo.make_ascii_lowercase();
-            let typo = UniCase::new(typo);
-            rows.entry(typo)
-                .or_insert_with(BTreeSet::new)
-                .extend(i.map(|c| {
-                    let mut c = c.to_owned();
-                    c.make_ascii_lowercase();
-                    c
-                }));
-        });
+        .from_reader(&*data);
 
-    let rows: Dict = rows
+    reader
+        .records()
+        .into_iter()
+        .map(Result::unwrap)
+        .map(|record| {
+            let mut iter = record.into_iter();
+            let typo = iter.next().expect("typo");
+            (
+                typo.to_owned(),
+                iter.map(ToOwned::to_owned).collect::<Vec<_>>(),
+            )
+        })
+        .collect()
+}
+
+fn dict_from_iter<S: Into<String>>(
+    iter: impl IntoIterator<Item = (S, impl IntoIterator<Item = S>)>,
+) -> Dict {
+    let mut dict = Dict::new();
+
+    for (typo, corrections) in iter {
+        let typo = UniCase::new(typo.into().to_ascii_lowercase());
+
+        // duplicate entries are merged
+        dict.entry(typo)
+            .or_insert_with(BTreeSet::new)
+            .extend(corrections.into_iter().map(|c| {
+                let mut c = c.into();
+                c.make_ascii_lowercase();
+                c
+            }));
+    }
+
+    dict
+}
+
+fn process<S: Into<String>>(
+    iter: impl IntoIterator<Item = (S, impl IntoIterator<Item = S>)>,
+) -> Dict {
+    let dict = dict_from_iter(iter);
+
+    let rows: Dict = dict
         .into_iter()
         .filter(|(t, _)| is_word(t))
         .filter_map(|(t, c)| {
@@ -95,8 +133,7 @@ fn generate<W: std::io::Write>(file: &mut W, dict: &[u8]) {
         .iter()
         .flat_map(|(bad, good)| good.iter().map(|good| (good.to_owned(), bad.to_owned())))
         .collect();
-    let rows: Vec<_> = dict
-        .into_iter()
+    dict.into_iter()
         .filter(|(typo, _)| {
             if let Some(correction) = corrections.get(typo.as_str()) {
                 eprintln!("{typo} <-> {correction} cycle detected");
@@ -105,15 +142,45 @@ fn generate<W: std::io::Write>(file: &mut W, dict: &[u8]) {
                 true
             }
         })
-        .collect();
+        .collect()
+}
 
-    let mut wtr = csv::WriterBuilder::new().flexible(true).from_writer(file);
-    for (typo, corrections) in rows {
-        let mut row = vec![typo.as_str().to_owned()];
-        row.extend(corrections);
-        wtr.write_record(&row).unwrap();
-    }
-    wtr.flush().unwrap();
+#[test]
+fn test_merge_duplicates() {
+    assert_eq!(
+        process([("foo", ["bar"]), ("foo", ["baz"])]),
+        dict_from_iter([("foo", ["bar", "baz"])])
+    );
+}
+
+#[test]
+fn test_duplicate_correction_removal() {
+    let dict = process([("foo", ["bar", "bar"])]);
+    assert_eq!(dict, dict_from_iter([("foo", ["bar"])]));
+}
+
+#[test]
+fn test_cycle_removal() {
+    assert!(process([("foo", ["bar"]), ("bar", ["foo"])]).is_empty());
+}
+
+#[test]
+fn test_varcon_removal() {
+    assert!(process([("colour", ["color"])]).is_empty());
+}
+
+#[test]
+fn test_varcon_best_match() {
+    assert_eq!(
+        process([(
+            "neighourhood", // note the missing 'b'
+            ["neighborhood"],
+        )]),
+        dict_from_iter([(
+            "neighourhood",
+            ["neighbourhood"] // note that 'bor' has become 'bour' to match the typo
+        )])
+    );
 }
 
 fn is_word(word: &str) -> bool {