vision-agent 1.1.1__py3-none-any.whl → 1.1.2__py3-none-any.whl

vision_agent/tools/tools.py

@@ -19,6 +19,7 @@ from IPython.display import display
 from PIL import Image, ImageDraw, ImageFont
 from pillow_heif import register_heif_opener  # type: ignore
 from pytube import YouTube  # type: ignore
+import pymupdf  # type: ignore
 
 from vision_agent.lmm.lmm import LMM, AnthropicLMM, OpenAILMM
 from vision_agent.utils.execute import FileSerializer, MimeType
@@ -3147,6 +3148,56 @@ def save_image(image: np.ndarray, file_path: str) -> None:
     pil_image.save(file_path)
 
 
+def load_pdf(pdf_path: str) -> List[np.ndarray]:
+    """'load_pdf' is a utility function that loads a PDF from the given file path string and converts each page to an image.
+
+    Parameters:
+        pdf_path (str): The path to the PDF file.
+
+    Returns:
+        List[np.ndarray]: A list of images as NumPy arrays, one for each page of the PDF.
+
+    Example
+    -------
+        >>> load_pdf("path/to/document.pdf")
+    """
+
+    # Handle URL case
+    if pdf_path.startswith(("http", "https")):
+        _, pdf_suffix = os.path.splitext(pdf_path)
+        with tempfile.NamedTemporaryFile(delete=False, suffix=pdf_suffix) as tmp_file:
+            # Download the PDF and save it to the temporary file
+            with urllib.request.urlopen(pdf_path) as response:
+                tmp_file.write(response.read())
+            pdf_path = tmp_file.name
+
+    # Open the PDF
+    doc = pymupdf.open(pdf_path)
+    images = []
+
+    # Convert each page to an image
+    for page_num in range(len(doc)):
+        page = doc.load_page(page_num)
+
+        # Render page to an image
+        pix = page.get_pixmap(matrix=pymupdf.Matrix(2, 2))
+
+        # Convert to PIL Image
+        img = Image.frombytes("RGB", (pix.width, pix.height), pix.samples)
+
+        # Convert to numpy array
+        images.append(np.array(img))
+
+    # Close the document
+    doc.close()
+
+    # Clean up temporary file if it was a URL
+    if pdf_path.startswith(("http", "https")):
+        os.unlink(pdf_path)
+
+    return images
+
+
 def save_video(
     frames: List[np.ndarray], output_video_path: Optional[str] = None, fps: float = 5
 ) -> str:

{vision_agent-1.1.1.dist-info → vision_agent-1.1.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: vision-agent
-Version: 1.1.1
+Version: 1.1.2
 Summary: Toolset for Vision Agent
 Author: Landing AI
 Author-email: dev@landing.ai
@@ -28,6 +28,7 @@ Requires-Dist: pandas (==2.*)
 Requires-Dist: pillow (==10.*)
 Requires-Dist: pillow-heif (>=0.16.0,<0.17.0)
 Requires-Dist: pydantic (>=2.0.0,<3.0.0)
+Requires-Dist: pymupdf (>=1.23.0,<2.0.0)
 Requires-Dist: pytube (==15.0.0)
 Requires-Dist: requests (==2.*)
 Requires-Dist: rich (>=13.7.1,<14.0.0)

{vision_agent-1.1.1.dist-info → vision_agent-1.1.2.dist-info}/RECORD

@@ -30,7 +30,7 @@ vision_agent/tools/__init__.py,sha256=o9lfWBVopT_qSoSi26WcgQJTKQYNgbXv7r4z_o5j2E
 vision_agent/tools/meta_tools.py,sha256=9iJilpGYEiXW0nYPTYAWHa7l23wGN8IM5KbE7mWDOT0,6798
 vision_agent/tools/planner_tools.py,sha256=iQWtTgXdomn0IWrbmvXXM-y8Q_RSEOxyP04HIRLrgWI,19576
 vision_agent/tools/prompts.py,sha256=V1z4YJLXZuUl_iZ5rY0M5hHc_2tmMEUKr0WocXKGt4E,1430
-vision_agent/tools/tools.py,sha256=yr45nu0Hr9_KT5sQn6ggaI0FP65XP0dlAmBtHnKihPU,124180
+vision_agent/tools/tools.py,sha256=dKKrfKxqQYVDFRsLjMMpp1z4_5k68pkaoZUMf1BMc_Q,125694
 vision_agent/utils/__init__.py,sha256=mANUs_84VL-3gpZbXryvV2mWU623eWnRlJCSUHtMjuw,122
 vision_agent/utils/agent.py,sha256=2ifTP5QElItnr4YHOJR6L5P1PUzV0GhChTTqVxuVyQg,15153
 vision_agent/utils/exceptions.py,sha256=zis8smCbdEylBVZBTVfEUfAh7Rb7cWV3MSPambu6FsQ,1837
@@ -40,7 +40,7 @@ vision_agent/utils/tools.py,sha256=XbH5wuD1nlaKe6zBeVPqUbQDsK6D-eCskKND3rRHOzo,8
 vision_agent/utils/tools_doc.py,sha256=yFue6KSXoa_Z1ngCdBEc4SdPZOWF1rVLeaHu02I8Wis,2523
 vision_agent/utils/video.py,sha256=rjsQ1sKKisaQ6AVjJz0zd_G4g-ovRweS_rs4JEhenoI,5340
 vision_agent/utils/video_tracking.py,sha256=DZLFpNCuzuPJQzbQoVNcp-m4dKxgiKdCNM5QTh_zURE,12245
-vision_agent-1.1.1.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-vision_agent-1.1.1.dist-info/METADATA,sha256=T9CSlGabaZwR1u2ZQlV2wTkXNpQZi5Nn1KwJMSo7s2o,12532
-vision_agent-1.1.1.dist-info/WHEEL,sha256=XbeZDeTWKc1w7CSIyre5aMDU_-PohRwTQceYnisIYYY,88
-vision_agent-1.1.1.dist-info/RECORD,,
+vision_agent-1.1.2.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+vision_agent-1.1.2.dist-info/METADATA,sha256=JxWPwfrAwtWx0Fpqq9b9Se7LZi22Ddqiw-YxX6nHe0A,12573
+vision_agent-1.1.2.dist-info/WHEEL,sha256=XbeZDeTWKc1w7CSIyre5aMDU_-PohRwTQceYnisIYYY,88
+vision_agent-1.1.2.dist-info/RECORD,,