Bottleneck Mass grave update< CAUSE IM A GOAT with No life please help me i need a life

src/pdf_import.rs

@@ -115,8 +115,16 @@ pub async fn import_pdf_directory(
             );
         }
 
-        let (count, completed) =
-            process_pdf_chunks(pdf_path, &chunks, &mut writer, ollama, model_name, pages_count, prompt_each_chunk).await?;
+        let (count, completed) = process_pdf_chunks(
+            pdf_path,
+            &chunks,
+            &mut writer,
+            ollama,
+            model_name,
+            pages_count,
+            prompt_each_chunk,
+        )
+        .await?;
         total_samples += count;
         if !completed {
             println!(
@@ -144,7 +152,14 @@ pub async fn import_pdf_path(
     prompt_each_chunk: bool,
 ) -> Result<bool> {
     if input_path.is_dir() {
-        import_pdf_directory(input_path, output_file, ollama, model_name, prompt_each_chunk).await
+        import_pdf_directory(
+            input_path,
+            output_file,
+            ollama,
+            model_name,
+            prompt_each_chunk,
+        )
+        .await
     } else if input_path.is_file() {
         if input_path
             .extension()
@@ -175,8 +190,16 @@ pub async fn import_pdf_path(
                 );
             }
 
-            let (total_samples, completed) =
-                process_pdf_chunks(input_path, &chunks, &mut writer, ollama, model_name, pages_count, prompt_each_chunk).await?;
+            let (total_samples, completed) = process_pdf_chunks(
+                input_path,
+                &chunks,
+                &mut writer,
+                ollama,
+                model_name,
+                pages_count,
+                prompt_each_chunk,
+            )
+            .await?;
             writer.flush()?;
             println!(
                 "✅ Wrote {} JSONL sample(s) to {}",
@@ -213,25 +236,35 @@ async fn process_pdf_chunks(
     prompt_each_chunk: bool,
 ) -> Result<(usize, bool)> {
     let mut total_samples = 0;
-    for (index, chunk) in chunks.iter().enumerate() {
-        let progress = format!("chunk {}/{}", index + 1, chunks.len());
-        let page_label = if pages_count > 0 {
-            let current_page = ((index * pages_count) / chunks.len()) + 1;
-            format!("page {}/{}", current_page, pages_count)
-        } else {
-            "page unknown".to_string()
-        };
-        println!("   ➜ Processing {} ({})", progress, page_label);
+    let batch_size: usize = 3;
+
+    for batch_start in (0..chunks.len()).step_by(batch_size) {
+        let batch_end = std::cmp::min(batch_start + batch_size, chunks.len());
+        let batch = &chunks[batch_start..batch_end];
+
+        for (offset, _chunk) in batch.iter().enumerate() {
+            let index = batch_start + offset;
+            let progress = format!("chunk {}/{}", index + 1, chunks.len());
+            let page_label = if pages_count > 0 {
+                let current_page = ((index * pages_count) / chunks.len()) + 1;
+                format!("page {}/{}", current_page, pages_count)
+            } else {
+                "page unknown".to_string()
+            };
+            println!("   ➜ Processing {} ({})", progress, page_label);
+        }
+
+        let owned_chunks: Vec<String> = batch.to_vec();
         let samples =
-            derive_samples_from_pdf_text(pdf_path, &[chunk.clone()], ollama, model_name).await?;
-        for sample in samples {
+            derive_samples_from_pdf_text(pdf_path, owned_chunks, ollama, model_name).await?;
+        for sample in &samples {
             let line = serde_json::to_string(&sample)?;
             writeln!(writer, "{}", line)?;
             total_samples += 1;
         }
 
-        if index + 1 < chunks.len() && prompt_each_chunk {
-            if !prompt_continue()? {
+        if batch_end < chunks.len() && prompt_each_chunk {
+            if !prompt_continue().await? {
                 println!("⚠️ Stopped early. Partial output written to disk.");
                 return Ok((total_samples, false));
             }
@@ -245,13 +278,16 @@ fn count_pdf_pages(pdf_path: &Path) -> Result<usize> {
     Ok(doc.get_pages().len())
 }
 
-fn prompt_continue() -> Result<bool> {
-    print!("Press Enter to continue to the next chunk, or type 'quit' to stop: ");
-    let stdout = io::stdout();
-    let mut handle = stdout.lock();
-    handle.flush()?;
+async fn prompt_continue() -> Result<bool> {
+    tokio::task::spawn_blocking(|| {
+        print!("Press Enter to continue to the next chunk, or type 'quit' to stop: ");
+        let stdout = io::stdout();
+        let mut handle = stdout.lock();
+        handle.flush()?;
 
-    let mut input = String::new();
-    io::stdin().read_line(&mut input)?;
-    Ok(!input.trim().eq_ignore_ascii_case("quit"))
+        let mut input = String::new();
+        io::stdin().read_line(&mut input)?;
+        Ok(!input.trim().eq_ignore_ascii_case("quit"))
+    })
+    .await?
 }