cassiopeia: implementing basic file parser

3 files changed, 165 insertions, 30 deletions

diff --git a/apps/cassiopeia/src/format/lexer.rs b/apps/cassiopeia/src/format/lexer.rs
index f062ca4238c1..bdb89f5180e5 100644
--- a/apps/cassiopeia/src/format/lexer.rs
+++ b/apps/cassiopeia/src/format/lexer.rs
@@ -1,6 +1,7 @@
 //! Cassiopeia file lexer
 
-use logos::Logos;
+use logos::{Lexer, Logos};
+use std::iter::Iterator;
 
 /// A basic line lexer type
 ///
@@ -8,8 +9,7 @@ use logos::Logos;
 /// does not attempt to parse the line specifics.  This is what the
 /// content lexer is for.
 #[derive(Logos, Debug, PartialEq)]
-enum Line {
-
+pub(crate) enum Token {
     #[token("HEADER")]
     Header,
 
@@ -21,7 +21,7 @@ enum Line {
 
     #[token("INVOICE")]
     Invoice,
-    
+
     #[regex(r"\w+=[^,$]+[,$]")]
     HeaderData,
 
@@ -32,85 +32,120 @@ enum Line {
 
     #[token(" ", logos::skip)]
     Space,
-    
+
+    #[regex(";;.*")]
+    Comment,
+
     #[error]
     Error,
 }
 
+/// A single token type on a line
+#[derive(Debug)]
+pub(crate) struct LineToken<'l> {
+    pub(crate) tt: Token,
+    pub(crate) slice: &'l str,
+}
 
-// pub fn test_this() {
-//     // let mut lex = Line::lexer("HEADER version=0.0.0,location=Berlin,");
-//     let mut lex = Line::lexer("START 2020-11-11 13:00:00+01:00");
+/// A lexer wrapped for a single line
+pub(crate) struct LineLexer<'l> {
+    lexer: Lexer<'l, Token>,
+}
+
+impl<'l> LineLexer<'l> {
+    pub(crate) fn get_all(self) -> Vec<LineToken<'l>> {
+        let mut acc = vec![];
+        for l in self {
+            acc.push(l);
+        }
+        acc
+    }
+}
 
-//     while let Some(t) = lex.next() {
-//         println!("{:?}: {}", t, lex.slice());
-//     }
-// }
+impl<'l> Iterator for LineLexer<'l> {
+    type Item = LineToken<'l>;
 
+    fn next(&mut self) -> Option<Self::Item> {
+        self.lexer.next().map(|tt| Self::Item {
+            tt,
+            slice: self.lexer.slice(),
+        })
+    }
+}
+
+/// Take a line of input and lex it into a stream of tokens
+pub(crate) fn lex<'l>(line: &'l mut String) -> LineLexer<'l> {
+    LineLexer {
+        lexer: Token::lexer(line),
+    }
+}
 
 #[test]
 fn basic_header() {
-    let mut lex = Line::lexer("HEADER version=0.0.0,location=Berlin Lichtenberg,");
+    let mut lex = Token::lexer("HEADER version=0.0.0,location=Berlin Lichtenberg,");
 
-    assert_eq!(lex.next(), Some(Line::Header));
+    assert_eq!(lex.next(), Some(Token::Header));
     assert_eq!(lex.span(), 0..6);
     assert_eq!(lex.slice(), "HEADER");
 
-    assert_eq!(lex.next(), Some(Line::HeaderData));
+    assert_eq!(lex.next(), Some(Token::HeaderData));
     assert_eq!(lex.span(), 7..21);
     assert_eq!(lex.slice(), "version=0.0.0,");
 
-    assert_eq!(lex.next(), Some(Line::HeaderData));
+    assert_eq!(lex.next(), Some(Token::HeaderData));
     assert_eq!(lex.span(), 21..49);
     assert_eq!(lex.slice(), "location=Berlin Lichtenberg,");
 
     assert_eq!(lex.next(), None);
 }
 
-
 #[test]
 fn basic_start() {
-    let mut lex = Line::lexer("START 2020-11-11 13:00:00+01:00");
+    let mut lex = Token::lexer("START 2020-11-11 13:00:00+01:00");
 
-    assert_eq!(lex.next(), Some(Line::Start));
+    assert_eq!(lex.next(), Some(Token::Start));
     assert_eq!(lex.span(), 0..5);
     assert_eq!(lex.slice(), "START");
 
-    assert_eq!(lex.next(), Some(Line::Date));
+    assert_eq!(lex.next(), Some(Token::Date));
     assert_eq!(lex.span(), 5..31);
     assert_eq!(lex.slice(), " 2020-11-11 13:00:00+01:00");
 
     assert_eq!(lex.next(), None);
 }
 
-
 #[test]
 fn basic_stop() {
-    let mut lex = Line::lexer("STOP 2020-11-11 13:00:00+01:00");
+    let mut lex = Token::lexer("STOP 2020-11-11 13:00:00+01:00");
 
-    assert_eq!(lex.next(), Some(Line::Stop));
+    assert_eq!(lex.next(), Some(Token::Stop));
     assert_eq!(lex.span(), 0..4);
     assert_eq!(lex.slice(), "STOP");
 
-    assert_eq!(lex.next(), Some(Line::Date));
+    assert_eq!(lex.next(), Some(Token::Date));
     assert_eq!(lex.span(), 4..30);
     assert_eq!(lex.slice(), " 2020-11-11 13:00:00+01:00");
 
     assert_eq!(lex.next(), None);
 }
 
-
 #[test]
 fn basic_invoice() {
-    let mut lex = Line::lexer("INVOICE 2020-11-11 13:00:00+01:00");
+    let mut lex = Token::lexer("INVOICE 2020-11-11 13:00:00+01:00");
 
-    assert_eq!(lex.next(), Some(Line::Invoice));
+    assert_eq!(lex.next(), Some(Token::Invoice));
     assert_eq!(lex.span(), 0..7);
     assert_eq!(lex.slice(), "INVOICE");
 
-    assert_eq!(lex.next(), Some(Line::Date));
+    assert_eq!(lex.next(), Some(Token::Date));
     assert_eq!(lex.span(), 7..33);
     assert_eq!(lex.slice(), " 2020-11-11 13:00:00+01:00");
 
     assert_eq!(lex.next(), None);
 }
+
+#[test]
+fn basic_comment() {
+    let mut lex = Token::lexer(";; This file is auto generated!");
+    assert_eq!(lex.next(), Some(Token::Comment));
+}
diff --git a/apps/cassiopeia/src/format/mod.rs b/apps/cassiopeia/src/format/mod.rs
index 766bb22be13e..beab2f7aac66 100644
--- a/apps/cassiopeia/src/format/mod.rs
+++ b/apps/cassiopeia/src/format/mod.rs
@@ -1,3 +1,25 @@
-//! cassiopeia file format handling
+//! cassiopeia file format
 
-pub(crate) mod lexer;
+mod lexer;
+mod parser;
+
+pub(crate) use lexer::{LineLexer, LineToken, Token};
+pub(crate) use parser::LineCfg;
+
+use crate::TimeFile;
+use std::{fs::File, io::Read};
+
+pub(crate) fn load_file(path: &str) {
+    let mut f = File::open(path).unwrap();
+    let mut content = String::new();
+    f.read_to_string(&mut content).unwrap();
+
+    let mut lines: Vec<String> = content.split("\n").map(|l| l.to_owned()).collect();
+
+    lines
+        .iter_mut()
+        .map(|line| lexer::lex(line))
+        .map(|lex| parser::parse(lex))
+        .filter(|line| line.valid())
+        .fold(TimeFile::default(), |file, line| file.append(line));
+}
diff --git a/apps/cassiopeia/src/format/parser.rs b/apps/cassiopeia/src/format/parser.rs
new file mode 100644
index 000000000000..cc4b1b7c77df
--- /dev/null
+++ b/apps/cassiopeia/src/format/parser.rs
@@ -0,0 +1,78 @@
+//! cassiopeia parser
+//!
+//! Takes a lexer's token stream as an input, and outputs a fully
+//! parsed time file.
+
+use crate::format::{LineLexer, LineToken, Token};
+use chrono::{DateTime, FixedOffset as Offset};
+use logos::Lexer;
+use std::collections::BTreeMap;
+use std::iter::Iterator;
+
+/// A type-parsed line in a time file
+#[derive(Debug)]
+pub enum LineCfg {
+    /// A header line with a set of keys and values
+    Header(BTreeMap<String, String>),
+    /// A session start line with a date and time
+    Start(Option<DateTime<Offset>>),
+    /// A session stop line with a date and time
+    Stop(Option<DateTime<Offset>>),
+    /// An invoice line with a date
+    Invoice(Option<DateTime<Offset>>),
+    /// An empty line
+    Empty,
+    /// A temporary value that is invalid
+    #[doc(hidden)]
+    Ignore,
+}
+
+impl LineCfg {
+    pub(crate) fn valid(&self) -> bool {
+        match self {
+            LineCfg::Ignore => false,
+            _ => true,
+        }
+    }
+}
+
+pub(crate) fn parse<'l>(lex: LineLexer<'l>) -> LineCfg {
+    use LineCfg::*;
+    use Token as T;
+
+    #[cfg_attr(rustfmt, rustfmt_skip)]
+    lex.get_all().into_iter().fold(Ignore, |cfg, tok| match (cfg, tok) {
+        // If the first token is a comment, we ignore it
+        (Ignore, LineToken { tt: T::Comment, .. }, ) => Ignore,
+        // If the first token is a keyword, we wait for more data
+        (Ignore, LineToken { tt: T::Header, .. }) => Header(Default::default()),
+        (Ignore, LineToken { tt: T::Start, .. }) => Start(None),
+        (Ignore, LineToken { tt: T::Stop, .. }) => Stop(None),
+        (Ignore, LineToken { tt: T::Invoice, .. }) => Invoice(None),
+
+        // If the first token _was_ a keyword, fill in the data
+        (Header(map), LineToken { tt: T::HeaderData, slice }) => Header(append_data(map, slice)),
+        (Start(_), LineToken { tt: T::Date, slice }) => Start(parse_date(slice)),
+        (Stop(_), LineToken { tt: T::Date, slice }) => Stop(parse_date(slice)),
+        (Invoice(_), LineToken { tt: T::Date, slice }) => Invoice(parse_date(slice)),
+
+        // Pass empty lines through,
+        (Empty, _) => Empty,
+
+        // Ignore everything else (which will be filtered)
+        _ => Ignore,
+    })
+}
+
+fn append_data(mut map: BTreeMap<String, String>, slice: &str) -> BTreeMap<String, String> {
+    let split = slice.split("=").collect::<Vec<_>>();
+    map.insert(split[0].into(), split[1].into());
+    map
+}
+
+fn parse_date(slice: &str) -> Option<DateTime<Offset>> {
+    Some(
+        DateTime::parse_from_str(slice, "%Y-%m-%d %H:%M:%S%:z")
+            .expect("Failed to parse date; invalid format!"),
+    )
+}