Trim whitespace when parsing entries

[tablify] / src / lib.rs

diff --git a/src/lib.rs b/src/lib.rs
index 245a47ae469dae6a426f471b1297701432167494..5b06e705aecaaac8fd071f064dd63d731038687d 100644 (file)
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -39,7 +39,24 @@ const FOOTER: &str = "    </tbody>
 </html>";
 
 #[derive(Debug, PartialEq, Eq, Hash)]
 </html>";
 
 #[derive(Debug, PartialEq, Eq, Hash)]

-struct Entry(String);
+struct Entry {
+    col: String,
+    instance: Option<String>,
+}
+impl From<&str> for Entry {
+    fn from(value: &str) -> Entry {
+        match value.split_once(':') {
+            None => Entry {
+                col: String::from(value),
+                instance: None,
+            },
+            Some((col, instance)) => Entry {
+                col: String::from(col.trim()),
+                instance: Some(String::from(instance.trim())),
+            },
+        }
+    }
+}

 
 #[derive(Debug, PartialEq, Eq)]
 struct RowInput {
 
 #[derive(Debug, PartialEq, Eq)]
 struct RowInput {


@@ -81,7 +98,7 @@ impl<Input: Iterator<Item = Result<String, std::io::Error>>> Iterator for Reader
                             n + 1
                         ))))
                     }
                             n + 1
                         ))))
                     }

-                    Some(ref mut row) => row.entries.push(Entry(String::from(line.trim()))),
+                    Some(ref mut row) => row.entries.push(Entry::from(line.trim())),

                 },
                 Some((_, Ok(line))) => {
                     let prev = std::mem::take(&mut self.row);
                 },
                 Some((_, Ok(line))) => {
                     let prev = std::mem::take(&mut self.row);


@@ -105,9 +122,15 @@ fn read_rows(input: impl std::io::Read) -> impl Iterator<Item = Result<RowInput,
 fn column_counts(rows: &[RowInput]) -> Vec<(usize, String)> {
     let mut counts: Vec<_> = rows
         .iter()
 fn column_counts(rows: &[RowInput]) -> Vec<(usize, String)> {
     let mut counts: Vec<_> = rows
         .iter()

-        .flat_map(|r| r.entries.iter().collect::<HashSet<_>>().into_iter())
-        .fold(HashMap::new(), |mut cs, e| {
-            cs.entry(String::from(&e.0))
+        .flat_map(|r| {
+            r.entries
+                .iter()
+                .map(|e| &e.col)
+                .collect::<HashSet<_>>()
+                .into_iter()
+        })
+        .fold(HashMap::new(), |mut cs, col| {
+            cs.entry(String::from(col))

                 .and_modify(|n| *n += 1)
                 .or_insert(1);
             cs
                 .and_modify(|n| *n += 1)
                 .or_insert(1);
             cs


@@ -135,6 +158,31 @@ pub fn tablify(input: impl std::io::Read) -> Result<String, std::io::Error> {
 mod tests {
     use super::*;
 
 mod tests {
     use super::*;
 

+    #[test]
+    fn test_parse_entry() {
+        assert_eq!(
+            Entry::from("foo"),
+            Entry {
+                col: String::from("foo"),
+                instance: None
+            }
+        );
+        assert_eq!(
+            Entry::from("foo:bar"),
+            Entry {
+                col: String::from("foo"),
+                instance: Some(String::from("bar"))
+            }
+        );
+        assert_eq!(
+            Entry::from("foo: bar"),
+            Entry {
+                col: String::from("foo"),
+                instance: Some(String::from("bar"))
+            }
+        );
+    }
+

     #[test]
     fn test_read_rows() {
         assert_eq!(
     #[test]
     fn test_read_rows() {
         assert_eq!(


@@ -168,7 +216,7 @@ mod tests {
             read_rows(&b"foo\n bar\n"[..]).flatten().collect::<Vec<_>>(),
             vec![RowInput {
                 label: String::from("foo"),
             read_rows(&b"foo\n bar\n"[..]).flatten().collect::<Vec<_>>(),
             vec![RowInput {
                 label: String::from("foo"),

-                entries: vec![Entry(String::from("bar"))]
+                entries: vec![Entry::from("bar")]

             }]
         );
         assert_eq!(
             }]
         );
         assert_eq!(


@@ -177,7 +225,7 @@ mod tests {
                 .collect::<Vec<_>>(),
             vec![RowInput {
                 label: String::from("foo"),
                 .collect::<Vec<_>>(),
             vec![RowInput {
                 label: String::from("foo"),

-                entries: vec![Entry(String::from("bar")), Entry(String::from("baz"))]
+                entries: vec![Entry::from("bar"), Entry::from("baz")]

             }]
         );
         assert_eq!(
             }]
         );
         assert_eq!(


@@ -216,7 +264,7 @@ mod tests {
                 .collect::<Vec<_>>(),
             vec![RowInput {
                 label: String::from("foo"),
                 .collect::<Vec<_>>(),
             vec![RowInput {
                 label: String::from("foo"),

-                entries: vec![Entry(String::from("bar"))]
+                entries: vec![Entry::from("bar")]

             }]
         );
 
             }]
         );
 


@@ -255,5 +303,13 @@ mod tests {
             ),
             vec![(1, String::from("bar")), (2, String::from("baz"))]
         );
             ),
             vec![(1, String::from("bar")), (2, String::from("baz"))]
         );

+        assert_eq!(
+            column_counts(
+                &read_rows(&b"foo\n bar: 1\n bar: 2\n baz\n bar\nquux\n baz"[..])
+                    .collect::<Result<Vec<_>, _>>()
+                    .unwrap()
+            ),
+            vec![(1, String::from("bar")), (2, String::from("baz"))]
+        );

     }
 }
     }
 }