Merge pull request #39 from datalab-to/dev

Enable passing bbox scale

chandra/model/__init__.py

@@ -4,6 +4,7 @@ from chandra.model.hf import load_model, generate_hf
 from chandra.model.schema import BatchInputItem, BatchOutputItem
 from chandra.model.vllm import generate_vllm
 from chandra.output import parse_markdown, parse_html, parse_chunks, extract_images
+from chandra.settings import settings
 
 
 class InferenceManager:
@@ -26,19 +27,27 @@ class InferenceManager:
             output_kwargs["include_headers_footers"] = kwargs.pop(
                 "include_headers_footers"
             )
+        bbox_scale = kwargs.get("bbox_scale", settings.BBOX_SCALE)
 
         if self.method == "vllm":
             results = generate_vllm(
-                batch, max_output_tokens=max_output_tokens, **kwargs
+                batch,
+                max_output_tokens=max_output_tokens,
+                bbox_scale=bbox_scale,
+                **kwargs,
             )
         else:
             results = generate_hf(
-                batch, self.model, max_output_tokens=max_output_tokens, **kwargs
+                batch,
+                self.model,
+                max_output_tokens=max_output_tokens,
+                bbox_scale=bbox_scale,
+                **kwargs,
             )
 
         output = []
         for result, input_item in zip(results, batch):
-            chunks = parse_chunks(result.raw, input_item.image)
+            chunks = parse_chunks(result.raw, input_item.image, bbox_scale=bbox_scale)
             output.append(
                 BatchOutputItem(
                     markdown=parse_markdown(result.raw, **output_kwargs),

chandra/model/hf.py

@@ -10,12 +10,18 @@ from chandra.settings import settings
 
 
 def generate_hf(
-    batch: List[BatchInputItem], model, max_output_tokens=None, **kwargs
+    batch: List[BatchInputItem],
+    model,
+    max_output_tokens=None,
+    bbox_scale: int = settings.BBOX_SCALE,
+    **kwargs,
 ) -> List[GenerationResult]:
     if max_output_tokens is None:
         max_output_tokens = settings.MAX_OUTPUT_TOKENS
 
-    messages = [process_batch_element(item, model.processor) for item in batch]
+    messages = [
+        process_batch_element(item, model.processor, bbox_scale) for item in batch
+    ]
     text = model.processor.apply_chat_template(
         messages, tokenize=False, add_generation_prompt=True
     )
@@ -48,12 +54,12 @@ def generate_hf(
     return results
 
 
-def process_batch_element(item: BatchInputItem, processor):
+def process_batch_element(item: BatchInputItem, processor, bbox_scale: int):
     prompt = item.prompt
     prompt_type = item.prompt_type
 
     if not prompt:
-        prompt = PROMPT_MAPPING[prompt_type]
+        prompt = PROMPT_MAPPING[prompt_type].replace("{bbox_scale}", str(bbox_scale))
 
     content = []
     image = scale_to_fit(item.image)  # Guarantee max size

chandra/model/vllm.py

@@ -28,6 +28,7 @@ def generate_vllm(
     max_workers: int | None = None,
     custom_headers: dict | None = None,
     max_failure_retries: int | None = None,
+    bbox_scale: int = settings.BBOX_SCALE,
 ) -> List[GenerationResult]:
     client = OpenAI(
         api_key=settings.VLLM_API_KEY,
@@ -54,7 +55,9 @@ def generate_vllm(
     ) -> GenerationResult:
         prompt = item.prompt
         if not prompt:
-            prompt = PROMPT_MAPPING[item.prompt_type]
+            prompt = PROMPT_MAPPING[item.prompt_type].replace(
+                "{bbox_scale}", str(bbox_scale)
+            )
 
         content = []
         image = scale_to_fit(item.image)

chandra/output.py

@@ -221,12 +221,12 @@ class LayoutBlock:
     content: str
 
 
-def parse_layout(html: str, image: Image.Image):
+def parse_layout(html: str, image: Image.Image, bbox_scale=settings.BBOX_SCALE):
     soup = BeautifulSoup(html, "html.parser")
     top_level_divs = soup.find_all("div", recursive=False)
     width, height = image.size
-    width_scaler = width / settings.BBOX_SCALE
-    height_scaler = height / settings.BBOX_SCALE
+    width_scaler = width / bbox_scale
+    height_scaler = height / bbox_scale
     layout_blocks = []
     for div in top_level_divs:
         bbox = div.get("data-bbox")
@@ -255,7 +255,7 @@ def parse_layout(html: str, image: Image.Image):
     return layout_blocks
 
 
-def parse_chunks(html: str, image: Image.Image):
-    layout = parse_layout(html, image)
+def parse_chunks(html: str, image: Image.Image, bbox_scale=settings.BBOX_SCALE):
+    layout = parse_layout(html, image, bbox_scale=bbox_scale)
     chunks = [asdict(block) for block in layout]
     return chunks

chandra/prompts.py

@@ -1,5 +1,3 @@
-from chandra.settings import settings
-
 ALLOWED_TAGS = [
     "math",
     "br",
@@ -67,7 +65,7 @@ Guidelines:
 """.strip()
 
 OCR_LAYOUT_PROMPT = f"""
-OCR this image to HTML, arranged as layout blocks.  Each layout block should be a div with the data-bbox attribute representing the bounding box of the block in [x0, y0, x1, y1] format.  Bboxes are normalized 0-{settings.BBOX_SCALE}. The data-label attribute is the label for the block.
+OCR this image to HTML, arranged as layout blocks.  Each layout block should be a div with the data-bbox attribute representing the bounding box of the block in [x0, y0, x1, y1] format.  Bboxes are normalized 0-{{bbox_scale}}. The data-label attribute is the label for the block.
 
 Use the following labels:
 - Caption