From 76cf5e496ce1e9964af442ce5a4f1a16219d7765 Mon Sep 17 00:00:00 2001
From: Alex Bethel <abethel@cs.nmt.edu>
Date: Mon, 29 Nov 2021 22:16:57 -0700
Subject: [PATCH] Start work on parser

---
 src/lib.rs | 117 ++++++++++++++++++++++++++++++++++++++++++++++++-----
 1 file changed, 107 insertions(+), 10 deletions(-)

diff --git a/src/lib.rs b/src/lib.rs
index fc1415d..40e6420 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -81,20 +81,117 @@ pub struct List {}
 
 pub fn parse(source: &str) -> Document {
     let lines = source.lines().collect::<Vec<_>>();
-    let mut document = vec![];
-    let mut paragraph = vec![];
+    let mut document = Vec::new();
 
-    for line in lines {
-        if line.starts_with('#') {
-            todo!()
-        } else if line.is_empty() {
-            document.push(paragraph.join(" "));
-        } else {
-            paragraph.push(line);
+    let mut to_parse = &lines[..];
+    while let Some((node, remaining)) = parse_node(to_parse) {
+        document.push(node);
+        to_parse = &to_parse[remaining..];
+    }
+
+    Document(document)
+}
+
+/// Parse a single node from the list of lines, if one exists. Return
+/// the node parsed, and the number of lines of input consumed.
+fn parse_node(mut lines: &[&str]) -> Option<(Node, usize)> {
+    if lines.is_empty() {
+        return None;
+    }
+
+    while lines[0].is_empty() {
+        lines = &lines[1..];
+        if lines.is_empty() {
+            return None;
         }
     }
 
-    todo!()
+    None.or_else(|| parse_normal_header(lines))
+        .or_else(|| parse_underline_header(lines))
+        .or_else(|| parse_hrule(lines))
+        .or_else(|| parse_paragraph(lines))
+}
+
+/// Parse a header as a number of "#"-signs followed by the header
+/// text.
+fn parse_normal_header(lines: &[&str]) -> Option<(Node, usize)> {
+    if lines[0].starts_with('#') {
+        let depth = lines[0].chars().take_while(|&c| c == '#').count();
+        let text = lines[0][depth..].trim().to_string();
+        Some((
+            Node::Header(Header {
+                label: text,
+                level: depth as u8,
+            }),
+            1,
+        ))
+    } else {
+        None
+    }
+}
+
+/// Parse a header as a single line of text followed by an underline,
+/// i.e. a line of "-" or "=" characters.
+fn parse_underline_header(lines: &[&str]) -> Option<(Node, usize)> {
+    if lines.len() < 2 {
+        return None;
+    }
+
+    let text = lines[0];
+    let underline = lines[1];
+    if !text.is_empty() && !underline.is_empty() {
+        if underline.chars().all(|c| c == '-') {
+            Some((
+                Node::Header(Header {
+                    label: text.to_string(),
+                    level: 2,
+                }),
+                2,
+            ))
+        } else if underline.chars().all(|c| c == '=') {
+            Some((
+                Node::Header(Header {
+                    label: text.to_string(),
+                    level: 1,
+                }),
+                2,
+            ))
+        } else {
+            None
+        }
+    } else {
+        None
+    }
+}
+
+/// Parse a horizontal rule.
+fn parse_hrule(lines: &[&str]) -> Option<(Node, usize)> {
+    let line = lines[0];
+    if !line.is_empty() && (line.chars().all(|c| c == '-') || line.chars().all(|c| c == '=')) {
+        Some((Node::HRule, 1))
+    } else {
+        None
+    }
+}
+
+/// Parse a paragraph of text.
+fn parse_paragraph(lines: &[&str]) -> Option<(Node, usize)> {
+    let len = lines.iter().take_while(|line| !line.is_empty()).count();
+    let text = lines[0..len].join(" ");
+
+    // TODO: Parse style components from `text` here, and turn it into
+    // a Vec<StyledText>.
+
+    Some((
+        Node::Paragraph(vec![StyledText {
+            italic: false,
+            bold: false,
+            underline: false,
+            color: None,
+            text,
+        }]),
+        len,
+    ))
 }
 
 #[cfg(test)]