feat: initial commit

2026-06-30 04:10:12 -04:00
commit e40ab61427
6 changed files with 6622 additions and 0 deletions
@@ -0,0 +1,84 @@
+[workspace]
+authors = [
+  "Alex Maldonado <alex@scient.ing>",
+]
+channels = ["conda-forge", "bioconda", "nvidia", "pytorch"]
+name = "pdf2markdown"
+platforms = ["linux-64"]
+version = "26.6.30"
+
+[environments]
+marker = ["marker"]
+chandra = ["chandra"]
+
+[feature.marker.pypi-dependencies]
+surya-ocr = "*"
+pdftext = "*"
+ebooklib = "*"
+transformers = "*"
+pillow = "*"
+pydantic = "*"
+pydantic-settings = "*"
+python-dotenv = "*"
+tqdm = "*"
+ftfy = "*"
+rapidfuzz = "*"
+regex = "*"
+markdownify = "*"
+click = "*"
+markdown2 = "*"
+filetype = "*"
+google-genai = "*"
+anthropic = "*"
+pre-commit = "*"
+scikit-learn = "*"
+mammoth = ">=1.9.0"
+openpyxl = ">=3.1.5"
+python-pptx = ">=1.0.2"
+weasyprint = ">=63.1"
+openai = ">=1.65.2"
+marker-pdf = ">=1.10.2, <2"
+
+[feature.marker.system-requirements]
+cuda = "12.0"
+
+[feature.marker.dependencies]
+python = ">=3.10"
+pytorch-gpu = ">=2.7.0"
+cuda-runtime = "==12.8"
+cuda-version = "==12.8"
+cuda-cudart-dev = ">=12.8"
+cuda-nvcc = ">=12.8"
+gcc = ">=14.3.0,<14.4"
+glib = ">=2.86.4,<3"
+
+[feature.marker.activation.env]
+PYTORCH_CUDA_ALLOC_CONF = "expandable_segments:True"
+RECOGNITION_BATCH_SIZE = "2"
+
+[feature.chandra.pypi-dependencies]
+beautifulsoup4 = "*"
+click = "*"
+filetype = "*"
+markdownify = "*"
+openai = "*"
+pillow = "*"
+pydantic = "*"
+pydantic-settings = "*"
+pypdfium2 = "*"
+python-dotenv = "*"
+six = "*"
+chandra-ocr = { version = ">=0.2.0, <0.3", extras = ["hf"] }
+
+[feature.chandra.system-requirements]
+cuda = "13.0"
+
+[feature.chandra.dependencies]
+python = ">=3.10"
+pytorch-gpu = ">=2.8.0"
+cuda-runtime = "==13.3.0"
+cuda-version = "==13.3.0"
+cuda-cudart-dev = ">=13.3.0"
+cuda-nvcc = ">=13.3.0"
+gcc = ">=14.3.0,<14.4"
+glib = ">=2.86.4,<3"