Initial commit: git-issue-commit CLI tool

src/parser/markdown.rs

@@ -0,0 +1,142 @@
+use regex::Regex;
+
+#[derive(Debug, Clone, Default)]
+pub struct MarkdownParser;
+
+impl MarkdownParser {
+    pub fn new() -> Self {
+        MarkdownParser
+    }
+
+    pub fn extract_title(&self, content: &str) -> Option<String> {
+        let lines: Vec<&str> = content.lines().collect();
+        extract_title(&lines)
+    }
+
+    pub fn extract_code_blocks(&self, content: &str) -> Vec<CodeBlock> {
+        extract_code_blocks(content)
+    }
+
+    pub fn extract_key_points(&self, content: &str) -> Vec<String> {
+        let re = Regex::new(r"^[\s]*[-*+]\s+(.+)$").expect("Invalid regex pattern");
+        re.captures_iter(content)
+            .filter_map(|cap| cap.get(1).map(|m| m.as_str().to_string()))
+            .filter(|s| !s.starts_with('['))
+            .collect()
+    }
+
+    pub fn extract_breaking_change_section(&self, content: &str) -> Option<String> {
+        let re = Regex::new(r"(?i)BREAKING\s*CHANGE[^\n]*\n*([^\n#][^\n]*)").expect("Invalid regex");
+        re.captures(content)
+            .and_then(|cap| cap.get(1))
+            .map(|m| m.as_str().trim().to_string())
+    }
+
+    pub fn clean_text(&self, content: &str) -> String {
+        content
+            .lines()
+            .map(|line| line.trim().trim_start_matches('#').trim())
+            .filter(|line| !line.is_empty())
+            .collect::<Vec<&str>>()
+            .join(" ")
+            .trim()
+            .to_string()
+    }
+}
+
+pub fn parse_markdown(content: &str) -> MarkdownParseResult {
+    let lines: Vec<&str> = content.lines().collect();
+
+    let title = extract_title(&lines);
+    let body = extract_body(&lines);
+    let code_blocks = extract_code_blocks(content);
+    let links = extract_links(content);
+    let lists = extract_lists(content);
+
+    MarkdownParseResult {
+        title,
+        body,
+        code_blocks,
+        links,
+        lists,
+    }
+}
+
+fn extract_title(lines: &[&str]) -> Option<String> {
+    for line in lines {
+        if line.starts_with("# ") {
+            return Some(line[2..].trim().to_string());
+        }
+    }
+    None
+}
+
+fn extract_body(lines: &[&str]) -> String {
+    let mut body_lines = Vec::new();
+    let mut in_code_block = false;
+
+    for line in lines {
+        if line.starts_with("```") {
+            in_code_block = !in_code_block;
+            continue;
+        }
+        if !in_code_block && !line.starts_with('#') {
+            body_lines.push(*line);
+        }
+    }
+
+    body_lines.join("\n")
+}
+
+fn extract_code_blocks(content: &str) -> Vec<CodeBlock> {
+    let re = Regex::new(r"```(\w*)\n([\s\S]*?)```").expect("Invalid regex pattern for code blocks");
+    re.captures_iter(content)
+        .map(|cap| CodeBlock {
+            language: cap.get(1).map(|m| m.as_str().to_string()).unwrap_or_default(),
+            code: cap.get(2).map(|m| m.as_str().to_string()).unwrap_or_default(),
+        })
+        .collect()
+}
+
+fn extract_links(content: &str) -> Vec<Link> {
+    let re = Regex::new(r"\[([^\]]+)\]\(([^)]+)\)").expect("Invalid regex pattern for links");
+    re.captures_iter(content)
+        .map(|cap| Link {
+            text: cap.get(1).map(|m| m.as_str().to_string()).unwrap_or_default(),
+            url: cap.get(2).map(|m| m.as_str().to_string()).unwrap_or_default(),
+        })
+        .collect()
+}
+
+fn extract_lists(content: &str) -> Vec<ListItem> {
+    let re = Regex::new(r"^[\s]*[-*+]\s+(\[.*?\]\s+)?(.+)$").expect("Invalid regex pattern for lists");
+    re.captures_iter(content)
+        .map(|cap| ListItem {
+            text: cap.get(2).map(|m| m.as_str().to_string()).unwrap_or_default(),
+            checked: cap.get(1).map(|m| !m.as_str().contains(' ')).unwrap_or(false),
+        })
+        .collect()
+}
+
+pub struct MarkdownParseResult {
+    pub title: Option<String>,
+    pub body: String,
+    pub code_blocks: Vec<CodeBlock>,
+    pub links: Vec<Link>,
+    pub lists: Vec<ListItem>,
+}
+
+pub struct CodeBlock {
+    pub language: String,
+    pub code: String,
+}
+
+pub struct Link {
+    pub text: String,
+    pub url: String,
+}
+
+pub struct ListItem {
+    pub text: String,
+    pub checked: bool,
+}