MVP-ish parser. AST and error handling need to be improved further

Cargo.lock

@@ -17,3 +17,29 @@ version = "0.1.0"
 [[package]]
 name = "cogs_parser"
 version = "0.1.0"
+dependencies = [
+ "cogs_ast",
+ "nom",
+]
+
+[[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 = "nom"
+version = "7.1.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d273983c5a657a70a3e8f2a01329822f3b8c8172b73826411a55751e404a0a4a"
+dependencies = [
+ "memchr",
+ "minimal-lexical",
+]

crates/ast/src/lib.rs

@@ -1,24 +1,30 @@
+#[derive(Debug)]
 pub struct Component {
     pub elements: Vec<Element>,
 }
 
+#[derive(Debug, Clone)]
 pub enum Element {
     Html(HtmlTag),
     Block(CodeBlock),
+    Text(String)
 }
 
+#[derive(Debug, Clone)]
 pub struct HtmlTag {
     pub tag: String,
     pub attributes: Vec<Attribute>,
     pub content: Vec<Element>,
 }
 
+#[derive(Debug, Clone)]
 pub struct Attribute {
-    pub name: String,
+    pub name: Element,
-    pub value: String,
+    pub value: Option<Element>,
 }
 
+#[derive(Debug, Clone)]
 pub struct CodeBlock {
-    pub is_async: bool,
+    // pub is_async: bool,
-    pub content: String,
+    pub content: Vec<Element>,
 }

crates/parser/Cargo.toml

@@ -4,3 +4,5 @@ version = "0.1.0"
 edition = "2021"
 
 [dependencies]
+nom = "7.1.3"
+cogs_ast.path = "../ast"

crates/parser/src/lib.rs

@@ -1,14 +1,231 @@
-pub fn add(left: u64, right: u64) -> u64 {
+use nom::{
-    left + right
+    branch::alt, bytes::complete::{is_not, tag, take_while1}, character::complete::{char, multispace0, one_of, space0, space1}, combinator::{opt, peek}, error::{Error, ErrorKind}, multi::{many0, many1, separated_list0}, sequence::{delimited, pair, preceded, terminated, tuple}, Err, IResult, InputLength
+};
+
+use cogs_ast::{Attribute, CodeBlock, Component, Element, HtmlTag};
+
+pub fn parse_cog(input: &str) -> IResult<&str, Component> {
+    let (input, elements) = parse_consecutive_proper_elements(input)?;
+    Ok((input, Component { elements }))
 }
 
-#[cfg(test)]
+pub fn parse_consecutive_proper_elements(input: &str) -> IResult<&str, Vec<Element>> {
-mod tests {
+    let (input, _) = multispace0(input)?;
-    use super::*;
+    let res = many0(parse_proper_element)(input);
+    // dbg!(&res);
+    res
+}
 
-    #[test]
+fn parse_element(input: &str) -> IResult<&str, Element> {
-    fn it_works() {
+    let (input, _) = multispace0(input)?;
-        let result = add(2, 2);
+    alt((parse_html, parse_code_block))(input)
-        assert_eq!(result, 4);
 }
+
+fn parse_proper_element(input: &str) -> IResult<&str, Element> {
+    // dbg!(&input);
+    let res = alt((parse_element, parse_text))(input);
+    // dbg!(&res);
+    res
+}
+
+
+fn is_valid_tag_name_char(c: char) -> bool {
+    c.is_alphanumeric() || c == '-'
+}
+
+fn is_valid_attr_char(c: char) -> bool {
+    c.is_alphanumeric() || c == '-' || c == '_'
+}
+
+fn parse_tag_name(input: &str) -> IResult<&str, &str> {
+    take_while1(is_valid_tag_name_char)(input)
+}
+
+fn parse_attribute(input: &str) -> IResult<&str, Attribute> {
+    // Shouldn't be needed with take_while1, re-add if fail
+    /*
+    if input.starts_with('>') || input.is_empty() {
+        // Return a recoverable error so `separated_list0` stops parsing cleanly
+        return Err(Err::Error(Error::new(input, nom::error::ErrorKind::Eof)));
+    }
+    */
+
+    let (input, key) = take_while1(is_valid_attr_char)(input)?;
+    let (input, value) = opt(preceded(
+        tuple((tag("="), space0)),
+        delimited(tag("\""), is_not("\""), tag("\""))
+    ))(input)?;
+
+    let mut resulting_value = None;
+
+    if value.is_some_and(|x| !x.is_empty()) {
+        resulting_value = Some(Element::Text(value.unwrap().to_string()));
+    }
+
+    Ok((
+        input,
+        Attribute {
+            name: Element::Text(key.to_string()),
+            value: resulting_value,
+        },
+    ))
+}
+
+
+fn parse_attributes(input: &str) -> IResult<&str, Vec<Attribute>> {
+    let (input, attrs) = separated_list0(
+        pair(alt((char(','), char(' '))), space0),
+        parse_attribute,
+    )(input)?;
+
+    let (input, _) = space0(input)?; // dump any trailing spaces
+
+    // dbg!(&attrs);
+
+    Ok((
+        input,
+        attrs
+    ))
+}
+
+fn parse_inside_html_opening_tag(input: &str) -> IResult<&str, HtmlTag> {
+    let (input, tag) = parse_tag_name(input)?;
+    // dbg!(&tag);
+    let (input, _) = space0(input)?;
+    let (input, attributes) = parse_attributes(input)?;
+    // dbg!(&attributes);
+
+    Ok((
+        input,
+        HtmlTag{
+            tag: tag.to_string(),
+            attributes,
+            content: Vec::new()
+        }
+    ))    
+}
+
+fn parse_html_opening_tag(input: &str) -> IResult<&str, HtmlTag> {
+    let (input, tag) = delimited(char('<'), parse_inside_html_opening_tag, char('>'))(input)?;
+
+    // dbg!(&tag);
+
+    Ok((
+        input,
+        tag
+    ))
+}
+
+fn parse_html_closing_tag(input: &str) -> IResult<&str, &str> {
+    let (input, tag) = delimited(tag("</"), parse_tag_name, char('>'))(input)?;
+
+    Ok((
+        input,
+        tag
+    ))
+}
+
+
+fn parse_text(input: &str) -> IResult<&str, Element> {
+    // dbg!(&input);
+    let mut index = 0;
+    while index < input.len() {
+        let current_slice = &input[index..];
+
+        
+        if peek(parse_element)(current_slice).is_ok() {
+            // dbg!(&current_slice);
+            break;
+        }
+        
+        if peek(parse_html_opening_tag)(current_slice).is_ok() ||
+           peek(parse_html_closing_tag)(current_slice).is_ok() ||
+           peek::<_, _, Error<&str>, _>(char('}'))(current_slice).is_ok(){
+                // dbg!(&current_slice);
+                break; // Stop if any of these parsers match
+        }
+
+        index += 1; // Increment to check the next character
+    }
+
+    if input[0..index].is_empty() {
+        // dbg!(&input[0..index]);
+        return Err(Err::Error(Error::new(input, ErrorKind::Eof)));
+    }
+
+    // dbg!(&input[0..index]);
+
+    Ok((&input[index..], Element::Text(input[0..index].to_string())))
+}
+
+fn parse_html_contents(input: &str) -> IResult<&str, Vec<Element>> {
+    let (input, out) = parse_consecutive_proper_elements(input)?;
+
+    // dbg!(&out);
+
+    Ok((
+        input,
+        out
+    ))
+}
+
+fn parse_html(input: &str) -> IResult<&str, Element> {
+    let (input, _) = multispace0(input)?; // remove spaces when debugging is complete
+    let (input, mut htag) = parse_html_opening_tag(input)?;
+    let (input, content) = parse_html_contents(input)?; // parse_consecutive_elements(input)?;
+    htag.content = content;
+
+    let (input, close_name) = parse_html_closing_tag(input)?;
+    if htag.tag != close_name {
+        return Err(Err::Failure(Error::new(input, ErrorKind::Fail))); // Is there a way to give a custom error message?
+    }
+
+
+    Ok((
+        input,
+        Element::Html(htag)
+    ))
+}
+
+/*
+fn parse_code_until_interrupted(input: &str) -> IResult<&str, Element> {
+    let (input, code) = pair(is_not("{};"), one_of("{};"))(input)?;
+
+    Ok((
+        input,
+        Element::Text(format!("{}{}", code.0, code.1))
+    ))
+}
+*/
+
+fn parse_inside_code_block(input: &str) -> IResult<&str, Vec<Element>> {
+    println!("Attempting inside code block {input}");
+    let (input, elems) = parse_consecutive_proper_elements(input)?;
+
+    dbg!(&elems);
+    Ok((
+        input,
+        elems
+    ))
+
+}
+
+
+fn parse_code_block(input: &str) -> IResult<&str, Element> {
+    if input.chars().nth(0) == Some('{') {
+        println!("Attempting code block on {input}");
+    }
+    let (input, content) = delimited(
+        char('{'),
+        parse_inside_code_block, // get here eventually, currently code blocks do not work at all.
+        char('}'),
+    )(input)?;
+
+    Ok((
+        input,
+        Element::Block(CodeBlock {
+            // is_async: is_async.unwrap_or(false),
+            content: content
+        }),
+    ))
 }

crates/parser/src/main.rs

@@ -0,0 +1,13 @@
+use cogs_parser::parse_cog;
+
+fn main() {
+    println!("{:#?}", parse_cog("
+<body>
+    <h1>Yo.</h1>
+    <a src=\"https://www.youtube.com/watch?v=dQw4w9WgXcQ\">Click this</a>
+    {
+        println!(\"test\");
+        <p>More Html</p>
+    }
+</body>"));
+}