Fixed document vs fragment parsing

This commit is contained in:
gyoder 2025-07-07 11:33:26 -06:00
parent 9a0038dacd
commit f9b2e74b45
6 changed files with 113 additions and 14 deletions

View file

@ -5,7 +5,9 @@ edition = "2024"
[dependencies]
anyhow = "1.0.98"
html5ever = "0.35.0"
kuchikiki = "0.8.2"
markup5ever = "0.11.0"
mlua = { version = "0.10.5", features = ["lua54", "vendored", "serialize"] }
tendril = "0.4.3"

View file

@ -0,0 +1,32 @@
use std::{fs::File, io::{BufRead, BufReader, Read, Seek, SeekFrom}, path::PathBuf};
use anyhow::Result;
use kuchikiki::NodeRef;
use markup5ever::{LocalName, Namespace, QualName};
use tendril::TendrilSink;
pub fn read_doc_from_file(path: PathBuf) -> Result<NodeRef> {
let mut file = File::open(path)?;
let mut reader = BufReader::new(&mut file);
let mut first_line = String::new();
reader.read_line(&mut first_line)?;
let first_line_trimmed = first_line.trim();
let is_whole_doc = first_line_trimmed.eq_ignore_ascii_case("<!DOCTYPE html>");
file.seek(SeekFrom::Start(0))?;
let mut page_string = String::new();
file.read_to_string(&mut page_string)?;
if is_whole_doc {
Ok(kuchikiki::parse_html().one(page_string))
} else {
let ctx_name = QualName::new(None, Namespace::from("http://www.w3.org/1999/xhtml"), LocalName::from("div"));
Ok(kuchikiki::parse_fragment(ctx_name, Vec::new()).one(page_string))
}
}

View file

@ -1,2 +1,3 @@
pub mod render;
pub mod serve;
pub mod helpers;

View file

@ -1,14 +1,15 @@
use std::cell::RefCell;
use std::fmt::Write;
use std::fs::read_to_string;
use std::path::PathBuf;
use std::rc::Rc;
use anyhow::Result;
use anyhow::anyhow;
use kuchikiki::{NodeRef, traits::TendrilSink};
use kuchikiki::NodeRef;
use mlua::{Lua, Table};
use crate::helpers::read_doc_from_file;
fn create_htmlua_stdlib(l: &Lua, stdout: &Rc<RefCell<String>>) -> mlua::Result<Table> {
let t = l.create_table()?;
@ -79,13 +80,14 @@ pub fn expand_template(document: NodeRef, component_path: &PathBuf, include_from
let exported = from_node
.select_first(format!("exportelement.{name}").as_str())
.map_err(|()| anyhow!("Error finding exportelement"))?;
exported.as_node().children().for_each(|c| i.as_node().insert_after(c));
exported.as_node().children().rev().for_each(|c| i.as_node().insert_after(c));
}
}
while let Ok(i) = document.select_first("includeelement") {
i.as_node().detach();
}
}
for i in document
.select("include")
.map_err(|()| anyhow!("Error finding include"))?
@ -97,10 +99,8 @@ pub fn expand_template(document: NodeRef, component_path: &PathBuf, include_from
if let Some(include_path) = attrs.get("path") {
let mut item_path = component_path.clone();
item_path.push(include_path);
let component_text = read_to_string(item_path)?;
let new_doc = kuchikiki::parse_html().one(component_text);
let new_node = new_doc.select_first("body").map_err(|()| anyhow!("Failed to get body"))?;
let replaced_node = expand_template(new_node.as_node().clone(), component_path, Some(i.as_node()))?;
let new_node = read_doc_from_file(item_path)?;
let replaced_node = expand_template(new_node, component_path, Some(i.as_node()))?;
i.as_node().insert_before(replaced_node);
i.as_node().detach();
}

View file

@ -1,8 +1,8 @@
use crate::helpers::read_doc_from_file;
use super::render::execute_lua;
use super::render::expand_template;
use anyhow::Result;
use kuchikiki::traits::TendrilSink;
use std::fs::read_to_string;
use std::path::{Path, PathBuf};
pub fn serve_content(request_uri: &str) -> Result<String> {
@ -13,9 +13,7 @@ pub fn serve_content(request_uri: &str) -> Result<String> {
let safe_path = Path::new(request_uri).strip_prefix("/")?;
let page_path = pages.join(safe_path);
let page_string = read_to_string(&page_path)?;
let doc = kuchikiki::parse_html().one(page_string);
let doc = read_doc_from_file(page_path)?;
let full_doc = expand_template(doc, &components, None)?;
let executed_doc = execute_lua(full_doc)?;