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

70
Cargo.lock generated
View file

@ -205,12 +205,23 @@ checksum = "bea68cab48b8459f17cf1c944c67ddc572d272d9f2b274140f223ecb1da4a3b7"
dependencies = [ dependencies = [
"log", "log",
"mac", "mac",
"markup5ever", "markup5ever 0.11.0",
"proc-macro2", "proc-macro2",
"quote", "quote",
"syn 1.0.109", "syn 1.0.109",
] ]
[[package]]
name = "html5ever"
version = "0.35.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "55d958c2f74b664487a2035fe1dadb032c48718a03b63f3ab0b8537db8549ed4"
dependencies = [
"log",
"markup5ever 0.35.0",
"match_token",
]
[[package]] [[package]]
name = "htmlua-apache-cgi" name = "htmlua-apache-cgi"
version = "0.1.0" version = "0.1.0"
@ -224,7 +235,9 @@ name = "htmlua-parser"
version = "0.1.0" version = "0.1.0"
dependencies = [ dependencies = [
"anyhow", "anyhow",
"html5ever 0.35.0",
"kuchikiki", "kuchikiki",
"markup5ever 0.11.0",
"mlua", "mlua",
"tendril", "tendril",
] ]
@ -256,7 +269,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f29e4755b7b995046f510a7520c42b2fed58b77bd94d5a87a8eb43d2fd126da8" checksum = "f29e4755b7b995046f510a7520c42b2fed58b77bd94d5a87a8eb43d2fd126da8"
dependencies = [ dependencies = [
"cssparser", "cssparser",
"html5ever", "html5ever 0.26.0",
"indexmap", "indexmap",
"matches", "matches",
"selectors", "selectors",
@ -329,6 +342,28 @@ dependencies = [
"tendril", "tendril",
] ]
[[package]]
name = "markup5ever"
version = "0.35.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "311fe69c934650f8f19652b3946075f0fc41ad8757dbb68f1ca14e7900ecc1c3"
dependencies = [
"log",
"tendril",
"web_atoms",
]
[[package]]
name = "match_token"
version = "0.35.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "ac84fd3f360fcc43dc5f5d186f02a94192761a080e8bc58621ad4d12296a58cf"
dependencies = [
"proc-macro2",
"quote",
"syn 2.0.104",
]
[[package]] [[package]]
name = "matches" name = "matches"
version = "0.1.10" version = "0.1.10"
@ -445,6 +480,15 @@ dependencies = [
"phf_shared 0.10.0", "phf_shared 0.10.0",
] ]
[[package]]
name = "phf"
version = "0.11.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "1fd6780a80ae0c52cc120a26a1a42c1ae51b247a253e4e06113d23d2c2edd078"
dependencies = [
"phf_shared 0.11.3",
]
[[package]] [[package]]
name = "phf_codegen" name = "phf_codegen"
version = "0.8.0" version = "0.8.0"
@ -465,6 +509,16 @@ dependencies = [
"phf_shared 0.10.0", "phf_shared 0.10.0",
] ]
[[package]]
name = "phf_codegen"
version = "0.11.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "aef8048c789fa5e851558d709946d6d79a8ff88c0440c587967f8e94bfb1216a"
dependencies = [
"phf_generator 0.11.3",
"phf_shared 0.11.3",
]
[[package]] [[package]]
name = "phf_generator" name = "phf_generator"
version = "0.8.0" version = "0.8.0"
@ -901,6 +955,18 @@ version = "0.11.1+wasi-snapshot-preview1"
source = "registry+https://github.com/rust-lang/crates.io-index" source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "ccf3ec651a847eb01de73ccad15eb7d99f80485de043efb2f370cd654f4ea44b" checksum = "ccf3ec651a847eb01de73ccad15eb7d99f80485de043efb2f370cd654f4ea44b"
[[package]]
name = "web_atoms"
version = "0.1.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "57ffde1dc01240bdf9992e3205668b235e59421fd085e8a317ed98da0178d414"
dependencies = [
"phf 0.11.3",
"phf_codegen 0.11.3",
"string_cache",
"string_cache_codegen",
]
[[package]] [[package]]
name = "which" name = "which"
version = "7.0.3" version = "7.0.3"

View file

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

View file

@ -1,14 +1,15 @@
use std::cell::RefCell; use std::cell::RefCell;
use std::fmt::Write; use std::fmt::Write;
use std::fs::read_to_string;
use std::path::PathBuf; use std::path::PathBuf;
use std::rc::Rc; use std::rc::Rc;
use anyhow::Result; use anyhow::Result;
use anyhow::anyhow; use anyhow::anyhow;
use kuchikiki::{NodeRef, traits::TendrilSink}; use kuchikiki::NodeRef;
use mlua::{Lua, Table}; use mlua::{Lua, Table};
use crate::helpers::read_doc_from_file;
fn create_htmlua_stdlib(l: &Lua, stdout: &Rc<RefCell<String>>) -> mlua::Result<Table> { fn create_htmlua_stdlib(l: &Lua, stdout: &Rc<RefCell<String>>) -> mlua::Result<Table> {
let t = l.create_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 let exported = from_node
.select_first(format!("exportelement.{name}").as_str()) .select_first(format!("exportelement.{name}").as_str())
.map_err(|()| anyhow!("Error finding exportelement"))?; .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") { while let Ok(i) = document.select_first("includeelement") {
i.as_node().detach(); i.as_node().detach();
} }
} }
for i in document for i in document
.select("include") .select("include")
.map_err(|()| anyhow!("Error finding 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") { if let Some(include_path) = attrs.get("path") {
let mut item_path = component_path.clone(); let mut item_path = component_path.clone();
item_path.push(include_path); item_path.push(include_path);
let component_text = read_to_string(item_path)?; let new_node = read_doc_from_file(item_path)?;
let new_doc = kuchikiki::parse_html().one(component_text); let replaced_node = expand_template(new_node, component_path, Some(i.as_node()))?;
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()))?;
i.as_node().insert_before(replaced_node); i.as_node().insert_before(replaced_node);
i.as_node().detach(); 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::execute_lua;
use super::render::expand_template; use super::render::expand_template;
use anyhow::Result; use anyhow::Result;
use kuchikiki::traits::TendrilSink;
use std::fs::read_to_string;
use std::path::{Path, PathBuf}; use std::path::{Path, PathBuf};
pub fn serve_content(request_uri: &str) -> Result<String> { 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 safe_path = Path::new(request_uri).strip_prefix("/")?;
let page_path = pages.join(safe_path); let page_path = pages.join(safe_path);
let page_string = read_to_string(&page_path)?; let doc = read_doc_from_file(page_path)?;
let doc = kuchikiki::parse_html().one(page_string);
let full_doc = expand_template(doc, &components, None)?; let full_doc = expand_template(doc, &components, None)?;
let executed_doc = execute_lua(full_doc)?; let executed_doc = execute_lua(full_doc)?;