Extractor rewrite

crates/pile-value/src/extract/item/epub/epub_text.rs

@@ -0,0 +1,105 @@
+use epub::doc::EpubDoc;
+use pile_config::Label;
+use std::{
+	collections::HashMap,
+	sync::{Arc, OnceLock},
+};
+use tracing::debug;
+
+use crate::{
+	extract::traits::ObjectExtractor,
+	value::{Item, PileValue, SyncReadBridge},
+};
+
+pub struct EpubTextExtractor {
+	item: Item,
+	output: OnceLock<HashMap<Label, PileValue>>,
+}
+
+impl EpubTextExtractor {
+	pub fn new(item: &Item) -> Self {
+		Self {
+			item: item.clone(),
+			output: OnceLock::new(),
+		}
+	}
+
+	async fn get_inner(&self) -> Result<&HashMap<Label, PileValue>, std::io::Error> {
+		if let Some(x) = self.output.get() {
+			return Ok(x);
+		}
+
+		let key = self.item.key();
+		let ext = key.as_str().rsplit('.').next();
+		if !matches!(ext, Some("epub")) {
+			return Ok(self.output.get_or_init(HashMap::new));
+		}
+
+		let reader = SyncReadBridge::new_current(self.item.read().await?);
+		let raw_text = tokio::task::spawn_blocking(move || {
+			let mut doc = EpubDoc::from_reader(reader)
+				.map_err(|e| std::io::Error::new(std::io::ErrorKind::InvalidData, e.to_string()))?;
+
+			let mut text_parts: Vec<String> = Vec::new();
+
+			loop {
+				if let Ok(content) = doc.get_current_str() {
+					text_parts.push(strip_html(&content));
+				}
+				if doc.go_next().is_err() {
+					break;
+				}
+			}
+
+			Ok::<_, std::io::Error>(text_parts.join(" "))
+		})
+		.await
+		.map_err(std::io::Error::other)?;
+
+		let raw_text = match raw_text {
+			Ok(x) => x,
+			Err(error) => {
+				debug!(message = "Could not process epub", ?error, key = ?self.item.key());
+				return Ok(self.output.get_or_init(HashMap::new));
+			}
+		};
+
+		#[expect(clippy::unwrap_used)]
+		let output = HashMap::from([(
+			Label::new("text").unwrap(),
+			PileValue::String(Arc::new(raw_text.into())),
+		)]);
+
+		let _ = self.output.set(output);
+		#[expect(clippy::unwrap_used)]
+		return Ok(self.output.get().unwrap());
+	}
+}
+
+/// Strip HTML/XHTML tags from a string, leaving only text nodes.
+fn strip_html(html: &str) -> String {
+	let mut result = String::with_capacity(html.len());
+	let mut in_tag = false;
+
+	for c in html.chars() {
+		match c {
+			'<' => in_tag = true,
+			'>' => in_tag = false,
+			_ if !in_tag => result.push(c),
+			_ => {}
+		}
+	}
+
+	result
+}
+
+#[async_trait::async_trait]
+impl ObjectExtractor for EpubTextExtractor {
+	async fn field(&self, name: &Label) -> Result<Option<PileValue>, std::io::Error> {
+		Ok(self.get_inner().await?.get(name).cloned())
+	}
+
+	async fn fields(&self) -> Result<Vec<Label>, std::io::Error> {
+		Ok(self.get_inner().await?.keys().cloned().collect())
+	}
+}