tags working now. rust parser works pretty good. next is cache

This commit is contained in:
2025-06-25 17:30:45 +02:00
parent b0b6625810
commit 0878b7dcec
3 changed files with 75 additions and 14 deletions

View File

@@ -17,3 +17,4 @@ regex = "1.10"
clap = { version = "4.4", features = ["derive"] } clap = { version = "4.4", features = ["derive"] }
serde_json = "1.0" serde_json = "1.0"
html-escape = "0.2.13" html-escape = "0.2.13"
once_cell = "1.18"

View File

@@ -15,6 +15,7 @@ use std::time::Duration;
use syntect::highlighting::{ThemeSet, Style}; use syntect::highlighting::{ThemeSet, Style};
use syntect::parsing::SyntaxSet; use syntect::parsing::SyntaxSet;
use syntect::html::{highlighted_html_for_string, IncludeBackground}; use syntect::html::{highlighted_html_for_string, IncludeBackground};
use once_cell::sync::Lazy;
#[derive(Debug, Deserialize, Clone, serde::Serialize)] #[derive(Debug, Deserialize, Clone, serde::Serialize)]
pub struct PostFrontmatter { pub struct PostFrontmatter {
@@ -77,6 +78,39 @@ fn process_anchor_links(content: &str) -> String {
}).to_string() }).to_string()
} }
// Helper function to strip emojis from a string
fn strip_emojis(s: &str) -> String {
// Remove all characters in the Emoji Unicode ranges
// This is a simple approach and may not cover all emojis, but works for most cases
// Unicode emoji ranges: https://unicode.org/Public/emoji/15.0/emoji-data.txt
s.chars()
.filter(|c| {
let c = *c as u32;
// Basic Emoji ranges
!( (c >= 0x1F600 && c <= 0x1F64F) // Emoticons
|| (c >= 0x1F300 && c <= 0x1F5FF) // Misc Symbols and Pictographs
|| (c >= 0x1F680 && c <= 0x1F6FF) // Transport and Map
|| (c >= 0x2600 && c <= 0x26FF) // Misc symbols
|| (c >= 0x2700 && c <= 0x27BF) // Dingbats
|| (c >= 0x1F900 && c <= 0x1F9FF) // Supplemental Symbols and Pictographs
|| (c >= 0x1FA70 && c <= 0x1FAFF) // Symbols and Pictographs Extended-A
|| (c >= 0x1F1E6 && c <= 0x1F1FF) // Regional Indicator Symbols
)
})
.collect()
}
static AMMONIA: Lazy<ammonia::Builder<'static>> = Lazy::new(|| {
let mut builder = ammonia::Builder::default();
builder.add_tag_attributes("h1", &["id"]);
builder.add_tag_attributes("h2", &["id"]);
builder.add_tag_attributes("h3", &["id"]);
builder.add_tag_attributes("h4", &["id"]);
builder.add_tag_attributes("h5", &["id"]);
builder.add_tag_attributes("h6", &["id"]);
builder
});
pub fn get_post_by_slug(slug: &str) -> Result<Post, Box<dyn std::error::Error>> { pub fn get_post_by_slug(slug: &str) -> Result<Post, Box<dyn std::error::Error>> {
let posts_dir = get_posts_directory(); let posts_dir = get_posts_directory();
let file_path = posts_dir.join(format!("{}.md", slug)); let file_path = posts_dir.join(format!("{}.md", slug));
@@ -122,7 +156,9 @@ pub fn get_post_by_slug(slug: &str) -> Result<Post, Box<dyn std::error::Error>>
}, },
Event::End(Tag::Heading(_, _, _)) => { Event::End(Tag::Heading(_, _, _)) => {
in_heading = false; in_heading = false;
let id = slugify(&heading_text); // Strip emojis before slugifying for the id
let heading_no_emoji = strip_emojis(&heading_text);
let id = slugify(&heading_no_emoji);
events.push(Event::Html(CowStr::Boxed(format!("<h{lvl} id=\"{id}\">", lvl=heading_level, id=id).into_boxed_str()))); events.push(Event::Html(CowStr::Boxed(format!("<h{lvl} id=\"{id}\">", lvl=heading_level, id=id).into_boxed_str())));
events.push(Event::Text(CowStr::Boxed(heading_text.clone().into_boxed_str()))); events.push(Event::Text(CowStr::Boxed(heading_text.clone().into_boxed_str())));
events.push(Event::Html(CowStr::Boxed(format!("</h{lvl}>", lvl=heading_level).into_boxed_str()))); events.push(Event::Html(CowStr::Boxed(format!("</h{lvl}>", lvl=heading_level).into_boxed_str())));
@@ -164,7 +200,7 @@ pub fn get_post_by_slug(slug: &str) -> Result<Post, Box<dyn std::error::Error>>
} }
html::push_html(&mut html_output, events.into_iter()); html::push_html(&mut html_output, events.into_iter());
let sanitized_html = clean(&html_output); let sanitized_html = AMMONIA.clean(&html_output).to_string();
Ok(Post { Ok(Post {
slug: slug.to_string(), slug: slug.to_string(),

View File

@@ -21,6 +21,16 @@ const parserStats = {
lastRustError: '', lastRustError: '',
}; };
// Add a slugify function that matches Rust's slug::slugify
function slugify(text: string): string {
return text
.toLowerCase()
.normalize('NFKD')
.replace(/[\u0300-\u036F]/g, '') // Remove diacritics
.replace(/[^a-z0-9]+/g, '-')
.replace(/^-+|-+$/g, '');
}
export default function PostPage({ params }: { params: { slug: string[] } }) { export default function PostPage({ params }: { params: { slug: string[] } }) {
const [post, setPost] = useState<Post | null>(null); const [post, setPost] = useState<Post | null>(null);
// Modal state for zoomed image // Modal state for zoomed image
@@ -315,35 +325,49 @@ export default function PostPage({ params }: { params: { slug: string[] } }) {
} }
} }
// Find the element, but only consider visible ones // Try to find the element by the raw ID first
const allElements = document.querySelectorAll(`#${id}`); let allElements = document.querySelectorAll(`#${id}`);
let element: HTMLElement | null = null; let element: HTMLElement | null = null;
// Check if we're on desktop or mobile
const isDesktop = window.innerWidth >= 640;
for (const el of Array.from(allElements)) { for (const el of Array.from(allElements)) {
const htmlEl = el as HTMLElement; const htmlEl = el as HTMLElement;
// Check if the element is visible (not hidden by CSS)
const rect = htmlEl.getBoundingClientRect(); const rect = htmlEl.getBoundingClientRect();
const isVisible = rect.width > 0 && rect.height > 0; const isVisible = rect.width > 0 && rect.height > 0;
if (isVisible) { if (isVisible) {
element = htmlEl; element = htmlEl;
break; break;
} }
} }
if (element) { if (element) {
console.log('Found target element:', element.textContent?.substring(0, 50)); console.log('Found target element (raw id):', element.textContent?.substring(0, 50));
scrollToElement(element); scrollToElement(element);
} else if (retryCount < 5) { return;
}
// If not found, try slugified version
const slugId = slugify(id);
if (slugId !== id) {
allElements = document.querySelectorAll(`#${slugId}`);
for (const el of Array.from(allElements)) {
const htmlEl = el as HTMLElement;
const rect = htmlEl.getBoundingClientRect();
const isVisible = rect.width > 0 && rect.height > 0;
if (isVisible) {
element = htmlEl;
break;
}
}
if (element) {
console.log('Found target element (slugified id):', element.textContent?.substring(0, 50));
scrollToElement(element);
return;
}
}
if (retryCount < 5) {
console.log(`Element not found for anchor: ${id}, retrying... (${retryCount + 1}/5)`); console.log(`Element not found for anchor: ${id}, retrying... (${retryCount + 1}/5)`);
setTimeout(() => { setTimeout(() => {
findAndScrollToElement(id, retryCount + 1); findAndScrollToElement(id, retryCount + 1);
}, 100); }, 100);
} else { } else {
console.warn(`Element with id "${id}" not found after retries`); console.warn(`Element with id "${id}" (or slugified "${slugId}") not found after retries`);
} }
}; };