Use nom to parse linesmain

8 files changed, 275 insertions, 63 deletions

diff --git a/Cargo.lock b/Cargo.lock
index bd06166..5670df0 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -251,6 +251,7 @@ dependencies = [
  "chrono",
  "clap",
  "crossterm",
+ "nom",
  "ratatui",
  "rusqlite",
  "rusqlite_migration",
@@ -378,6 +379,18 @@ dependencies = [
 ]
 
 [[package]]
+name = "memchr"
+version = "2.7.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "78ca9ab1a0babb1e7d5695e3530886289c18cf2f87ec19a575a0abdce112e3a3"
+
+[[package]]
+name = "minimal-lexical"
+version = "0.2.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "68354c5c6bd36d73ff3feceb05efa59b6acb7626617f4962be322a825e61f79a"
+
+[[package]]
 name = "mio"
 version = "0.8.9"
 source = "registry+https://github.com/rust-lang/crates.io-index"
@@ -390,6 +403,16 @@ dependencies = [
 ]
 
 [[package]]
+name = "nom"
+version = "7.1.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d273983c5a657a70a3e8f2a01329822f3b8c8172b73826411a55751e404a0a4a"
+dependencies = [
+ "memchr",
+ "minimal-lexical",
+]
+
+[[package]]
 name = "num-traits"
 version = "0.2.17"
 source = "registry+https://github.com/rust-lang/crates.io-index"
diff --git a/Cargo.toml b/Cargo.toml
index 1f22eb6..906a2ff 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -9,6 +9,7 @@ anyhow = "1.0"
 chrono = "0.4"
 clap = { version = "4.4", features = ["derive"] }
 crossterm = { version = "0.27" }
+nom = { version = "7.1", features = ["alloc"] }
 rusqlite = { version = "0.29", features = [ "chrono" ] }
 rusqlite_migration = "1.0"
 serde = { version = "1.0", features = ["derive"] }
diff --git a/README.md b/README.md
index 6941979..310c5de 100644
--- a/README.md
+++ b/README.md
@@ -20,13 +20,17 @@ Cards are created from a plain text `./deck` file:
 ```
 # This is a comment
 
-- good moorning : bonjour
-- alternative 1 | alternative 2 : choix 1 | choix 2
-- cat (indication) : chat [ʃa]
+good moorning : bonjour
+hi : salut # Another comment
+alternative 1 | alternative 2 : choix 1 | choix 2
+cat (indication) : chat [ʃa]
 ```
 
 # Backlog
 
+- migrate to FSRS https://github.com/open-spaced-repetition/rs-fsrs
 - Fix crashes on zoom / changing vertical size
-- Show a message when the deck can’t be read
-- Get a rough idea of card learning state
+- Get a rough idea of card learning state:
+    - number of cards
+    - amount of cards coming in in the following days
+    - show presentation page?
diff --git a/src/deck.rs b/src/deck.rs
index 4491d9b..3fff77f 100644
--- a/src/deck.rs
+++ b/src/deck.rs
@@ -1,4 +1,5 @@
-use crate::{model::Line, util::serialization};
+use crate::model::Line;
+use crate::parser;
 use anyhow::{Error, Result};
 use std::fmt;
 use std::fs::File;
@@ -7,13 +8,17 @@ use std::path::Path;
 
 #[derive(Debug, Clone)]
 struct ParseError {
-    line: usize,
-    message: String,
+    line: String,
+    line_number: usize,
 }
 
 impl fmt::Display for ParseError {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        write!(f, "Line {}: {}", self.line, self.message)
+        write!(
+            f,
+            "Error parsing line {}:\n\n{}",
+            self.line_number, self.line
+        )
     }
 }
 
@@ -40,39 +45,12 @@ pub fn read_file(path: &str) -> Result<Vec<Line>> {
 }
 
 fn read_line(index: usize, line: &str) -> Result<Option<Line>> {
-    let line = line.trim();
-
-    if line.starts_with('#') || line.is_empty() {
-        Ok(None)
-    } else if !line.starts_with('-') {
-        Err(Error::from(ParseError {
-            line: index + 1,
-            message: "an entry should starts with “-”.".to_string(),
-        }))
-    } else {
-        let without_minus = line.split('-').skip(1).collect::<Vec<&str>>().join("-");
-        let without_comment = without_minus.split('#').collect::<Vec<&str>>()[0].trim();
-        let translation = without_comment.split(':').collect::<Vec<&str>>();
-        if translation.len() != 2 {
-            Err(Error::from(ParseError {
-                line: index + 1,
-                message: "an entry should contain two parts separated by “:”.".to_string(),
-            }))
-        } else {
-            let t1 = translation[0].trim();
-            let t2 = translation[1].trim();
-            if t1.is_empty() || t2.is_empty() {
-                Err(Error::from(ParseError {
-                    line: index + 1,
-                    message: "an entry should contain two parts separated by “:”.".to_string(),
-                }))
-            } else {
-                Ok(Some(Line {
-                    part_1: serialization::line_to_words(t1),
-                    part_2: serialization::line_to_words(t2),
-                }))
-            }
-        }
+    match parser::parse_line(line) {
+        Ok(("", line)) => Ok(line),
+        _ => Err(Error::from(ParseError {
+            line_number: index + 1,
+            line: line.to_string(),
+        })),
     }
 }
 
@@ -98,12 +76,11 @@ pub mod tests {
 
     #[test]
     fn errors() {
-        is_error("A : a");
-        is_error("- A");
-        is_error("- A -> a");
-        is_error("- A : B : C");
-        is_error("- : ");
-        is_error("- A : a\n-")
+        is_error("A");
+        is_error("A -> a");
+        is_error("A : B : C");
+        is_error(":");
+        is_error("A : a\n-")
     }
 
     #[test]
@@ -117,19 +94,19 @@ pub mod tests {
 
     #[test]
     fn card() {
-        check("- A : a", &[(&["A"], &["a"])]);
+        check("A : a", &[(&["A"], &["a"])]);
     }
 
     #[test]
     fn cards() {
-        check("- A : a\n- B : b", &[(&["A"], &["a"]), (&["B"], &["b"])]);
+        check("A : a\nB : b", &[(&["A"], &["a"]), (&["B"], &["b"])]);
     }
 
     #[test]
     fn alternatives() {
-        check("- A : a1 | a2", &[(&["A"], &["a1", "a2"])]);
-        check("- A1 | A2 : a", &[(&["A1", "A2"], &["a"])]);
-        check("- A1 | A2 : a1 | a2", &[(&["A1", "A2"], &["a1", "a2"])]);
+        check("A : a1 | a2", &[(&["A"], &["a1", "a2"])]);
+        check("A1 | A2 : a", &[(&["A1", "A2"], &["a"])]);
+        check("A1 | A2 : a1 | a2", &[(&["A1", "A2"], &["a1", "a2"])]);
     }
 
     fn is_error(content: &str) {
diff --git a/src/main.rs b/src/main.rs
index ebc2d7e..3454897 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -2,6 +2,7 @@ mod db;
 mod deck;
 mod gui;
 mod model;
+mod parser;
 mod space_repetition;
 mod sync;
 mod util;
@@ -47,13 +48,10 @@ fn main() -> Result<()> {
                     gui::message::show(&mut term, &deck_name, &format!("{msg}"), true)?
                 }
             }
-        },
-        Err(msg) => {
-            gui::message::show(&mut term, &deck_name, &format!("{msg}"), true)?
         }
+        Err(msg) => gui::message::show(&mut term, &deck_name, &format!("{msg}"), true)?,
     }
 
-
     gui::restore_terminal(&mut term)
 }
 
diff --git a/src/parser.rs b/src/parser.rs
new file mode 100644
index 0000000..62df908
--- /dev/null
+++ b/src/parser.rs
@@ -0,0 +1,206 @@
+use crate::model::Line;
+use nom::{
+    bytes::complete::{is_not, take, take_until},
+    character::complete::{char, space0},
+    combinator::{opt, peek},
+    error::{Error, ErrorKind},
+    multi::separated_list1,
+    sequence::delimited,
+    IResult,
+};
+
+// Temporary API
+// Only expose parse_next ATM, so that tracking line number is done outside of the parser.
+
+pub fn parse_line(s: &str) -> IResult<&str, Option<Line>> {
+    let (s, _) = space0(s)?;
+    let (s, line) = opt(parse_parts)(s)?;
+    let (s, _) = parse_end_of_line(s)?;
+    Ok((s, line))
+}
+
+fn parse_parts(s: &str) -> IResult<&str, Line> {
+    let (s, part_1) = parse_options(s)?;
+    let (s, _) = sep(':')(s)?;
+    let (s, part_2) = parse_options(s)?;
+    Ok((s, Line { part_1, part_2 }))
+}
+
+// Rest
+
+fn parse_options(s: &str) -> IResult<&str, Vec<String>> {
+    Ok(separated_list1(sep('|'), parse_term)(s)?)
+}
+
+pub fn parse_term(s: &str) -> IResult<&str, String> {
+    let mut term = String::from("");
+    let mut s = s;
+
+    loop {
+        match peek(take::<usize, &str, ()>(1 as usize))(s) {
+            Ok((_, c)) => {
+                if c == "[" {
+                    let (s1, cs) = take_until("]")(s)?;
+                    s = s1;
+                    term.push_str(cs);
+                } else if c == "(" {
+                    let (s1, cs) = take_until(")")(s)?;
+                    s = s1;
+                    term.push_str(cs);
+                } else if c == ":" || c == "|" || c == "#" || c == "\n" || c == "\r" {
+                    break;
+                } else {
+                    let (s1, cs) = is_not("[(:|#\n\r")(s)?;
+                    s = s1;
+                    term.push_str(cs);
+                }
+            }
+            _ => break,
+        }
+    }
+
+    if term.is_empty() {
+        Err(nom::Err::Error(Error {
+            input: s,
+            code: ErrorKind::Fail,
+        }))
+    } else {
+        Ok((s, term.trim().to_string()))
+    }
+}
+
+fn parse_end_of_line(s: &str) -> IResult<&str, ()> {
+    let (s, _) = space0(s)?;
+    let (s, _) = opt(parse_comment)(s)?;
+    Ok((s, ()))
+}
+
+fn parse_comment(s: &str) -> IResult<&str, ()> {
+    let (s, _) = char('#')(s)?;
+    let (s, _) = is_not("\n\r")(s)?;
+    Ok((s, ()))
+}
+
+// Helpers
+
+pub fn sep(c: char) -> impl FnMut(&str) -> IResult<&str, ()> {
+    move |s: &str| {
+        let (s, _) = delimited(space0, char(c), space0)(s)?;
+        Ok((s, ()))
+    }
+}
+
+// Tests
+
+#[cfg(test)]
+mod tests {
+
+    use super::*;
+    use nom::{character::complete::newline, multi::many0};
+
+    fn parse(s: &str) -> IResult<&str, Vec<Line>> {
+        let (s, lines) = many0(parse_next)(s)?;
+        let (s, _) = many0(parse_empty_line)(s)?;
+        Ok((s, lines))
+    }
+
+    fn parse_next(s: &str) -> IResult<&str, Line> {
+        let (s, _) = many0(parse_empty_line)(s)?;
+        let (s, _) = space0(s)?;
+        let (s, part_1) = parse_options(s)?;
+        let (s, _) = sep(':')(s)?;
+        let (s, part_2) = parse_options(s)?;
+        let (s, _) = parse_end_of_line(s)?;
+        let (s, _) = opt(newline)(s)?;
+        Ok((s, Line { part_1, part_2 }))
+    }
+
+    fn parse_empty_line(s: &str) -> IResult<&str, ()> {
+        let (s, _) = parse_end_of_line(s)?;
+        let (s, _) = newline(s)?;
+        Ok((s, ()))
+    }
+
+    #[test]
+    fn simple() {
+        assert_eq!(parse("foo : bar"), lines(vec!((vec!("foo"), vec!("bar")))))
+    }
+
+    #[test]
+    fn spaces() {
+        assert_eq!(
+            parse("     foo     :      bar      "),
+            lines(vec!((vec!("foo"), vec!("bar"))))
+        )
+    }
+
+    #[test]
+    fn comments() {
+        assert_eq!(
+            parse("foo : bar # This is a comment"),
+            lines(vec!((vec!("foo"), vec!("bar"))))
+        )
+    }
+
+    #[test]
+    fn options() {
+        assert_eq!(
+            parse("foo | bar | baz : A | B | C"),
+            lines(vec!((vec!("foo", "bar", "baz"), vec!("A", "B", "C"))))
+        )
+    }
+
+    #[test]
+    fn term_with_spaces() {
+        assert_eq!(
+            parse("foo bar : baz baz"),
+            lines(vec!((vec!("foo bar"), vec!("baz baz"))))
+        )
+    }
+
+    #[test]
+    fn paren() {
+        assert_eq!(
+            parse("foo (|:[) : bar [::|)]"),
+            lines(vec!((vec!("foo (|:[)"), vec!("bar [::|)]"))))
+        )
+    }
+
+    #[test]
+    fn empty_lines() {
+        assert_eq!(parse("\n \n   \n# Hello\n     # Test\n\n\n"), lines(vec!()))
+    }
+
+    #[test]
+    fn multi_lines() {
+        assert_eq!(
+            parse("foo : FOO\nbar : BAR\nbaz : BAZ"),
+            lines(vec!(
+                (vec!("foo"), vec!("FOO")),
+                (vec!("bar"), vec!("BAR")),
+                (vec!("baz"), vec!("BAZ"))
+            ))
+        )
+    }
+
+    // Helpers
+
+    static EMPTY_STRING: &str = "";
+
+    fn lines(lines: Vec<(Vec<&str>, Vec<&str>)>) -> IResult<&'static str, Vec<Line>> {
+        Ok((
+            EMPTY_STRING,
+            lines
+                .into_iter()
+                .map(|line| to_line(line.0, line.1))
+                .collect(),
+        ))
+    }
+
+    fn to_line(part_1: Vec<&str>, part_2: Vec<&str>) -> Line {
+        Line {
+            part_1: part_1.into_iter().map(|s| s.to_string()).collect(),
+            part_2: part_2.into_iter().map(|s| s.to_string()).collect(),
+        }
+    }
+}
diff --git a/src/sync.rs b/src/sync.rs
index 6e3d84b..974e838 100644
--- a/src/sync.rs
+++ b/src/sync.rs
@@ -113,7 +113,7 @@ mod tests {
 
     #[test]
     fn test_added() {
-        let diff = deck_diff("- A : a", "- A : a\n- B : b");
+        let diff = deck_diff("A : a", "A : a\nB : b");
 
         has_questions(diff.new, vec![("B", vec!["b"]), ("b", vec!["B"])]);
         assert!(diff.deleted.is_empty());
@@ -122,7 +122,7 @@ mod tests {
 
     #[test]
     fn test_updated() {
-        let diff = deck_diff("- A : a1", "- A : a2");
+        let diff = deck_diff("A : a1", "A : a2");
 
         has_questions(diff.new, vec![("A", vec!["a2"]), ("a2", vec!["A"])]);
         has_questions(diff.deleted, vec![("A", vec!["a1"]), ("a1", vec!["A"])]);
@@ -131,7 +131,7 @@ mod tests {
 
     #[test]
     fn test_deleted() {
-        let diff = deck_diff("- A : a", "");
+        let diff = deck_diff("A : a", "");
 
         assert!(diff.new.is_empty());
         has_questions(diff.deleted, vec![("A", vec!["a"]), ("a", vec!["A"])]);
@@ -146,7 +146,7 @@ mod tests {
             deleted: Some(0),
         }];
 
-        let diff = super::diff(db_entries, deck::tests::read_string("- A : a").unwrap());
+        let diff = super::diff(db_entries, deck::tests::read_string("A : a").unwrap());
 
         has_questions(diff.new, vec![("a", vec!["A"])]);
         assert!(diff.deleted.is_empty());
@@ -154,7 +154,7 @@ mod tests {
     }
     #[test]
     fn regroup_same_question() {
-        let diff = deck_diff("", "- A : a\n- A | B : b");
+        let diff = deck_diff("", "A : a\nA | B : b");
 
         has_questions(
             diff.new,
diff --git a/src/util/serialization.rs b/src/util/serialization.rs
index 189a41a..a8b5438 100644
--- a/src/util/serialization.rs
+++ b/src/util/serialization.rs
@@ -23,6 +23,9 @@ mod tests {
     #[test]
     fn test_words_to_line() {
         assert_eq!(words_to_line(&["a".to_string()]), "a");
-        assert_eq!(words_to_line(&["a".to_string(), "b".to_string(), "c".to_string()]), "a | b | c");
+        assert_eq!(
+            words_to_line(&["a".to_string(), "b".to_string(), "c".to_string()]),
+            "a | b | c"
+        );
     }
 }