Fix file output dir

chandra/scripts/cli.py

@@ -87,7 +87,7 @@ def save_merged_output(
 
         # Save extracted images if requested
         if save_images and result.images:
-            images_dir = file_output_dir / "images"
+            images_dir = file_output_dir
             images_dir.mkdir(exist_ok=True)
 
             for img_name, pil_image in result.images.items():
@@ -172,7 +172,7 @@ def save_merged_output(
 @click.option(
     "--batch-size",
     type=int,
-    default=1,
+    default=None,
     help="Number of pages to process in a batch.",
 )
 @click.option(
@@ -194,6 +194,16 @@ def main(
     batch_size: int,
     paginate_output: bool,
 ):
+    if method == "hf":
+        click.echo(
+            "When using '--method hf', ensure that the batch size is set correctly.  We will default to batch size of 1."
+        )
+        if batch_size is None:
+            batch_size = 1
+    elif method == "vllm":
+        if batch_size is None:
+            batch_size = 28
+
     click.echo("Chandra CLI - Starting OCR processing")
     click.echo(f"Input: {input_path}")
     click.echo(f"Output: {output_path}")

chandra/scripts/vllm.py

@@ -17,8 +17,6 @@ def main():
         "-v",
         f"{os.path.expanduser('~')}/.cache/huggingface:/root/.cache/huggingface",
         "--env",
-        f"HUGGING_FACE_HUB_TOKEN={os.getenv('HF_TOKEN')}",
-        "--env",
         "VLLM_ATTENTION_BACKEND=TORCH_SDPA",
         "-p",
         "8000:8000",

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "chandra-ocr"
-version = "0.1.3"
+version = "0.1.6"
 description = "OCR model that converts documents to markdown, HTML, or JSON."
 readme = "README.md"
 requires-python = ">=3.10"