vision-agent 0.2.214__py3-none-any.whl → 0.2.215__py3-none-any.whl

vision_agent/tools/__init__.py

@@ -32,6 +32,7 @@ from .tools import (
     countgd_sam2_video_tracking,
     depth_anything_v2,
     detr_segmentation,
+    document_analysis,
     extract_frames_and_timestamps,
     florence2_ocr,
     florence2_phrase_grounding,

vision_agent/tools/tools.py

@@ -1879,6 +1879,64 @@ def closest_box_distance(
     return cast(float, np.sqrt(horizontal_distance**2 + vertical_distance**2))
 
 
+def document_analysis(image: np.ndarray) -> Dict[str, Any]:
+    """'document_analysis' is an understanding tool that can handle various
+    types of document image layouts. It returns a structured output containing the text,
+    tables, pictures, charts and information caption, summary, labels, bounding boxes, etc
+    avoiding information loss.
+
+    Parameters:
+        image (np.ndarray): The document image to analyze
+
+    Returns:
+        Dict[str, Any]: A dictionary containing the extracted information.
+
+    Example
+    -------
+        >>> document_analysis(image)
+        {'pages': [{'bbox': [left_0, top_0, right_0, bottom_0],
+                    'chunks': [{'bbox': [left_1, top_1, right_1, bottom_1],
+                                'caption': 'TITLE',
+                                'label': 'page_header',
+                                'summary': 'The image contains a single word ...' },
+                               {'bbox': [left_2, top_2, right_2, bottom_2],
+                                'caption': {'data': [{'value': 200, 'year': '2024' ...},
+                                    'title': 'Total CapEx Spending',
+                                    'type': 'bar chart',
+                                    'unit': 'Billion USD',
+                                    'xAxis': 'Year',
+                                    'yAxis': 'Total CapEx Spending'},
+                                'label': 'picture',
+                                'summary': 'This bar chart illustrates the trend of ...'},
+                    ],
+    """
+
+    image_file = numpy_to_bytes(image)
+
+    files = [("image", image_file)]
+
+    payload = {
+        "model": "document-analysis",
+    }
+
+    response: dict[str, Any] = send_inference_request(
+        payload=payload,
+        endpoint_name="document-analysis",
+        files=files,
+        v2=True,
+        metadata_payload={"function_name": "document_analysis"},
+    )
+
+    _display_tool_trace(
+        document_analysis.__name__,
+        payload,
+        response,
+        files,
+    )
+
+    return response
+
+
 # Utility and visualization functions
 
 

{vision_agent-0.2.214.dist-info → vision_agent-0.2.215.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vision-agent
-Version: 0.2.214
+Version: 0.2.215
 Summary: Toolset for Vision Agent
 Author: Landing AI
 Author-email: dev@landing.ai

{vision_agent-0.2.214.dist-info → vision_agent-0.2.215.dist-info}/RECORD

@@ -26,12 +26,12 @@ vision_agent/fonts/default_font_ch_en.ttf,sha256=1YM0Z3XqLDjSNbF7ihQFSAIUdjF9m1r
 vision_agent/lmm/__init__.py,sha256=jyY1sJb_tYKg5-Wzs3p1lvwFkc-aUNZfMcLy3TOC4Zg,100
 vision_agent/lmm/lmm.py,sha256=x_nIyDNDZwq4-pfjnJTmcyyJZ2_B7TjkA5jZp88YVO8,17103
 vision_agent/lmm/types.py,sha256=ZEXR_ptBL0ZwDMTDYkgxUCmSZFmBYPQd2jreNzr_8UY,221
-vision_agent/tools/__init__.py,sha256=InL8zUTRN8i_9J6r2wAtYdtNrVkElqdO_p-e2OA8q5A,2770
+vision_agent/tools/__init__.py,sha256=Ny522Y4h1xDQTW6kBP_ceUM4jc0Y14dRhcHdtMDdr24,2793
 vision_agent/tools/meta_tools.py,sha256=TPeS7QWnc_PmmU_ndiDT03dXbQ5yDSP33E7U8cSj7Ls,28660
 vision_agent/tools/planner_tools.py,sha256=k7PPu-HhwDwusQgFSPTCWKRVVHBzPMeYB6h2xSEjdUo,13273
 vision_agent/tools/prompts.py,sha256=V1z4YJLXZuUl_iZ5rY0M5hHc_2tmMEUKr0WocXKGt4E,1430
 vision_agent/tools/tool_utils.py,sha256=LAnrb_nY6PNVamqJahRN-J0cuOy4gsKvCtSuXJf0RsI,10075
-vision_agent/tools/tools.py,sha256=ZcXEI0Pb54OGXnLWi690SFx22k7JlEmQ-N16LzRLHlk,90627
+vision_agent/tools/tools.py,sha256=xzN1uOkVQ9l1MaXsJxT_VlDp6nLQfdBX04kex_jE0fc,92692
 vision_agent/tools/tools_types.py,sha256=8hYf2OZhI58gvf65KGaeGkt4EQ56nwLFqIQDPHioOBc,2339
 vision_agent/utils/__init__.py,sha256=7fMgbZiEwbNS0fBOS_hJI5PuEYBblw36zLi_UjUzvj4,244
 vision_agent/utils/exceptions.py,sha256=booSPSuoULF7OXRr_YbC4dtKt6gM_HyiFQHBuaW86C4,2052
@@ -40,7 +40,7 @@ vision_agent/utils/image_utils.py,sha256=5uoYgXa6E0-lVrXR7K2XE7fe6r_n7pvK64HYQ50
 vision_agent/utils/sim.py,sha256=f1emBQM8SmyVKSrhj0NHItnfMHSeTw-Nk2pw-0eBZ5c,7462
 vision_agent/utils/type_defs.py,sha256=BE12s3JNQy36QvauXHjwyeffVh5enfcvd4vTzSwvEZI,1384
 vision_agent/utils/video.py,sha256=e1VwKhXzzlC5LcFMyrcQYrPnpnX4wxDpnQ-76sB4jgM,6001
-vision_agent-0.2.214.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-vision_agent-0.2.214.dist-info/METADATA,sha256=H-NMLaCs8bVHlxWxDQzqixEQjbqMOwYk4aYGkE13BqM,19071
-vision_agent-0.2.214.dist-info/WHEEL,sha256=7Z8_27uaHI_UZAc4Uox4PpBhQ9Y5_modZXWMxtUi4NU,88
-vision_agent-0.2.214.dist-info/RECORD,,
+vision_agent-0.2.215.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+vision_agent-0.2.215.dist-info/METADATA,sha256=nSGpnpDpzJmWmGYDSShBvfjD5dbB6ZWSgOXGQ2Ci_yM,19071
+vision_agent-0.2.215.dist-info/WHEEL,sha256=7Z8_27uaHI_UZAc4Uox4PpBhQ9Y5_modZXWMxtUi4NU,88
+vision_agent-0.2.215.dist-info/RECORD,,