fixed with pymupdf

app/tabs/submit.py

@@ -5,6 +5,7 @@ import re
 import time
 
 import certifi
+import fitz  # PyMuPDF
 import gradio as gr
 import pycurl
 import spaces
@@ -13,7 +14,7 @@ from gradio_modal import Modal
 from htrflow.pipeline.pipeline import Pipeline
 from htrflow.pipeline.steps import init_step
 from htrflow.volume.volume import Collection
-from pdf2image import convert_from_path
+from PIL import Image
 
 from app.pipelines import PIPELINES
 
@@ -70,6 +71,44 @@ class PipelineWithProgress(Pipeline):
         return collection
 
 
+def pdf_to_images(pdf_path):
+    """
+    Convert a PDF file to a list of PIL Image objects using PyMuPDF.
+    Extracts full-resolution images with no DPI adjustment.
+
+    Args:
+        pdf_path (str): Path to the PDF file
+
+    Returns:
+        list: List of PIL Image objects
+    """
+    # Open the PDF
+    pdf_document = fitz.open(pdf_path)
+
+    # List to store the images
+    images = []
+
+    # Iterate through each page
+    for page_num in range(len(pdf_document)):
+        # Get the page
+        page = pdf_document[page_num]
+
+        # Get the pixmap at default resolution
+        pixmap = page.get_pixmap(alpha=False)
+
+        # Convert pixmap to PIL Image
+        img_data = pixmap.tobytes("jpeg")
+        img = Image.open(io.BytesIO(img_data))
+
+        # Add the image to our list
+        images.append(img)
+
+    # Close the PDF
+    pdf_document.close()
+
+    return images
+
+
 @spaces.GPU
 def run_htrflow(custom_template_yaml, batch_image_gallery, progress=gr.Progress()):
     """
@@ -424,7 +463,7 @@ with gr.Blocks() as submit:
     image_url.submit(lambda url: [url], image_url, batch_image_gallery)
 
     pdf_file.upload(
-        lambda imgs: convert_from_path(imgs), inputs=pdf_file, outputs=pdf_gallery
+        lambda imgs: pdf_to_images(imgs), inputs=pdf_file, outputs=pdf_gallery
     )
 
     run_button.click(

pyproject.toml

@@ -24,7 +24,7 @@ dependencies = [
     "dill>=0.3.9",
     "spaces>=0.32.0",
     "pycurl",
-    "pdf2image>=1.17.0",
+    "pymupdf>=1.25.3",
 ]
 
 [project.urls]

requirements.txt

@@ -4,4 +4,4 @@ tqdm>=4.67.1
 gradio-modal>=0.0.4
 dill>=0.3.9
 pycurl>=7.45.6
-pdf2image>=1.17.0
+pymupdf>=1.25.3

uv.lock

@@ -543,8 +543,8 @@ dependencies = [
     { name = "gradio" },
     { name = "gradio-modal" },
     { name = "htrflow" },
-    { name = "pdf2image" },
     { name = "pycurl" },
+    { name = "pymupdf" },
     { name = "spaces" },
     { name = "tqdm" },
 ]
@@ -562,8 +562,8 @@ requires-dist = [
     { name = "gradio", specifier = ">=5.17.0" },
     { name = "gradio-modal", specifier = ">=0.0.4" },
     { name = "htrflow", specifier = "==0.2.5" },
-    { name = "pdf2image", specifier = ">=1.17.0" },
     { name = "pycurl" },
+    { name = "pymupdf", specifier = ">=1.25.3" },
     { name = "spaces", specifier = ">=0.32.0" },
     { name = "tqdm", specifier = ">=4.67.1" },
 ]
@@ -1185,18 +1185,6 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/29/d4/1244ab8edf173a10fd601f7e13b9566c1b525c4f365d6bee918e68381889/pandas-2.2.3-cp312-cp312-win_amd64.whl", hash = "sha256:59ef3764d0fe818125a5097d2ae867ca3fa64df032331b7e0917cf5d7bf66b13", size = 11504248 },
 ]
 
-[[package]]
-name = "pdf2image"
-version = "1.17.0"
-source = { registry = "https://pypi.org/simple" }
-dependencies = [
-    { name = "pillow" },
-]
-sdist = { url = "https://files.pythonhosted.org/packages/00/d8/b280f01045555dc257b8153c00dee3bc75830f91a744cd5f84ef3a0a64b1/pdf2image-1.17.0.tar.gz", hash = "sha256:eaa959bc116b420dd7ec415fcae49b98100dda3dd18cd2fdfa86d09f112f6d57", size = 12811 }
-wheels = [
-    { url = "https://files.pythonhosted.org/packages/62/33/61766ae033518957f877ab246f87ca30a85b778ebaad65b7f74fa7e52988/pdf2image-1.17.0-py3-none-any.whl", hash = "sha256:ecdd58d7afb810dffe21ef2b1bbc057ef434dabbac6c33778a38a3f7744a27e2", size = 11618 },
-]
-
 [[package]]
 name = "pfzy"
 version = "0.3.4"
@@ -1492,6 +1480,21 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/8a/0b/9fcc47d19c48b59121088dd6da2488a49d5f72dacf8262e2790a1d2c7d15/pygments-2.19.1-py3-none-any.whl", hash = "sha256:9ea1544ad55cecf4b8242fab6dd35a93bbce657034b0611ee383099054ab6d8c", size = 1225293 },
 ]
 
+[[package]]
+name = "pymupdf"
+version = "1.25.3"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/06/47/b61c1c44b87cbdaeecdec3f43ce524ed6b3c72172bc6184eb82c94fbc43d/pymupdf-1.25.3.tar.gz", hash = "sha256:b640187c64c5ac5d97505a92e836da299da79c2f689f3f94a67a37a493492193", size = 67259841 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/61/9b/98ef4b98309e9db3baa9fe572f0e61b6130bb9852d13189970f35b703499/pymupdf-1.25.3-cp39-abi3-macosx_10_9_x86_64.whl", hash = "sha256:96878e1b748f9c2011aecb2028c5f96b5a347a9a91169130ad0133053d97915e", size = 19343576 },
+    { url = "https://files.pythonhosted.org/packages/14/62/4e12126db174c8cfbf692281cda971cc4046c5f5226032c2cfaa6f83e08d/pymupdf-1.25.3-cp39-abi3-macosx_11_0_arm64.whl", hash = "sha256:6ef753005b72ebfd23470f72f7e30f61e21b0b5e748045ec5b8f89e6e3068d62", size = 18580114 },
+    { url = "https://files.pythonhosted.org/packages/ec/c5/cf7ecf005e4f8ba3664d6aaa0613adeba4c2ab524832c452c69857e7184f/pymupdf-1.25.3-cp39-abi3-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:cbff443d899f37b17f1e67563cc03673d50b4bf33ccc237e73d34f18f3a07ccf", size = 19442580 },
+    { url = "https://files.pythonhosted.org/packages/52/de/bd1418e31f73d37b8381cd5deacfd681e6be702b8890e123e83724569ee1/pymupdf-1.25.3-cp39-abi3-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:46d90c4f9e62d1856e8db4b9f04a202ff4a7f086a816af73abdc86adb7f5e25a", size = 19999825 },
+    { url = "https://files.pythonhosted.org/packages/42/ee/3c449b0de061440ba1ac984aa845315e9e2dca0ff2003c5adfc6febff203/pymupdf-1.25.3-cp39-abi3-musllinux_1_2_x86_64.whl", hash = "sha256:a5de51efdbe4d486b6c1111c84e8a231cbfb426f3d6ff31ab530ad70e6f39756", size = 21123157 },
+    { url = "https://files.pythonhosted.org/packages/83/53/71faaaf91c56f2883b13f3dd849bf2697f012eb35eb7b952d62734cff41f/pymupdf-1.25.3-cp39-abi3-win32.whl", hash = "sha256:bca72e6089f985d800596e22973f79cc08af6cbff1d93e5bda9248326a03857c", size = 15094211 },
+    { url = "https://files.pythonhosted.org/packages/09/e0/d72e88a1d5e23aa381fd463057dc3d0fb29090e1e7308a870c334716579c/pymupdf-1.25.3-cp39-abi3-win_amd64.whl", hash = "sha256:4fb357438c9129fbf939b5af85323434df64e36759c399c376b62ad6da95498c", size = 16542949 },
+]
+
 [[package]]
 name = "pyparsing"
 version = "3.2.1"