tesseract single threaded

src/adapters/pdfpages.rs

@@ -41,38 +41,6 @@ impl GetMetadata for PdfPagesAdapter {
 	}
 }
 
-/*// todo: do this in an actually streaming fashion and less slow
-// IEND chunk + PDF magic
-// 4945 4e44 ae42 6082 8950 4e47 0d0a 1a0a
-let split_seq = hex_literal::hex!("4945 4e44 ae42 6082 8950 4e47 0d0a 1a0a");
-let split_seq_inx = 8;
-fn split_by_seq<'a>(
-	split_seq: &'a [u8],
-	split_inx: usize,
-	read: &mut Read,
-) -> Fallible<impl IntoIterator<Item = impl Read> + 'a> {
-	let regex = split_seq
-		.iter()
-		.map(|c| format!("\\x{:0>2x}", c))
-		.collect::<Vec<_>>()
-		.join("");
-	let restr = format!("(?-u){}", regex);
-	eprintln!("re: {}", restr);
-	let re = regex::bytes::Regex::new(&restr)?;
-
-	let mut all = Vec::new();
-	read.read_to_end(&mut all)?;
-	let mut out: Vec<Cursor<Vec<u8>>> = Vec::new();
-	let mut last = 0;
-	for (i, split) in re.find_iter(&all).enumerate() {
-		let pos = split.start() + split_inx;
-		out.push(Cursor::new(Vec::from(&all[last..pos])));
-		last = pos;
-	}
-	out.push(Cursor::new(Vec::from(&all[last..])));
-	Ok(out)
-}*/
-
 impl FileAdapter for PdfPagesAdapter {
 	fn adapt(&self, ai: AdaptInfo) -> Fallible<()> {
 		let AdaptInfo {
@@ -98,17 +66,13 @@ impl FileAdapter for PdfPagesAdapter {
 		let mut cmd = Command::new(exe_name);
 		cmd.arg("convert")
 			.arg("-density")
-			.arg("300")
+			.arg("200")
 			.arg(inp_fname)
 			.arg("+adjoin")
 			.arg(out_fname);
 
 		let mut cmd = cmd.spawn().map_err(|e| {
-			map_exe_error(
-				e,
-				exe_name,
-				"Could not find gm. Make sure you have graphicsmagick installed.",
-			)
+			map_exe_error(e, exe_name, "Make sure you have graphicsmagick installed.")
 		})?;
 		let args = config.args;
 		// TODO: how to handle this copying better?
@@ -133,7 +97,7 @@ impl FileAdapter for PdfPagesAdapter {
 				is_real_file: false,
 				inp: &mut ele,
 				oup,
-				line_prefix,
+				line_prefix: &format!("{}Page {}:", line_prefix, i + 1),
 				archive_recursion_depth: archive_recursion_depth + 1,
 				config: PreprocConfig { cache: None, args },
 			})?;
@@ -141,3 +105,35 @@ impl FileAdapter for PdfPagesAdapter {
 		Ok(())
 	}
 }
+
+/*// todo: do this in an actually streaming fashion and less slow
+// IEND chunk + PDF magic
+// 4945 4e44 ae42 6082 8950 4e47 0d0a 1a0a
+let split_seq = hex_literal::hex!("4945 4e44 ae42 6082 8950 4e47 0d0a 1a0a");
+let split_seq_inx = 8;
+fn split_by_seq<'a>(
+	split_seq: &'a [u8],
+	split_inx: usize,
+	read: &mut Read,
+) -> Fallible<impl IntoIterator<Item = impl Read> + 'a> {
+	let regex = split_seq
+		.iter()
+		.map(|c| format!("\\x{:0>2x}", c))
+		.collect::<Vec<_>>()
+		.join("");
+	let restr = format!("(?-u){}", regex);
+	eprintln!("re: {}", restr);
+	let re = regex::bytes::Regex::new(&restr)?;
+
+	let mut all = Vec::new();
+	read.read_to_end(&mut all)?;
+	let mut out: Vec<Cursor<Vec<u8>>> = Vec::new();
+	let mut last = 0;
+	for (i, split) in re.find_iter(&all).enumerate() {
+		let pos = split.start() + split_inx;
+		out.push(Cursor::new(Vec::from(&all[last..pos])));
+		last = pos;
+	}
+	out.push(Cursor::new(Vec::from(&all[last..])));
+	Ok(out)
+}*/

src/adapters/tesseract.rs

@@ -36,7 +36,8 @@ impl SpawningFileAdapter for TesseractAdapter {
 		"tesseract"
 	}
 	fn command(&self, _filepath_hint: &Path, mut cmd: Command) -> Command {
-		cmd.arg("-").arg("-");
+		// rg already does threading
+		cmd.env("OMP_THREAD_LIMIT", "1").arg("-").arg("-");
 		cmd
 	}
 }