chore: drop the pymupdf extra — image extraction detects the library at runtime

This commit is contained in:
Ray 2026-07-19 16:31:37 +08:00
parent 36202a6029
commit 5fc1b81c62
3 changed files with 2 additions and 7 deletions

View file

@ -816,7 +816,7 @@ def get_page_tokens(pdf_path, model=None, pdf_parser="PyPDF2"):
page_list.append((page_text, token_length))
return page_list
elif pdf_parser == "PyMuPDF":
import pymupdf # optional dependency: pip install pageindex[pymupdf]
import pymupdf # optional dependency
if isinstance(pdf_path, BytesIO):
pdf_stream = pdf_path
doc = pymupdf.open(stream=pdf_stream, filetype="pdf")

View file

@ -56,7 +56,7 @@ class PdfParser:
if not _warned_no_pymupdf:
logging.getLogger(__name__).warning(
"PyMuPDF is not installed; skipping image extraction. "
'Install with: pip install "pageindex[pymupdf]"')
"Install with: pip install pymupdf")
_warned_no_pymupdf = True
return {}

View file

@ -22,8 +22,6 @@ packages = [{include = "pageindex"}]
[tool.poetry.dependencies]
python = ">=3.10"
litellm = ">=1.83.0"
# AGPL-licensed; optional so the default install stays permissively licensed.
pymupdf = {version = ">=1.26.0", optional = true}
PyPDF2 = ">=3.0.0"
python-dotenv = ">=1.0.0"
pyyaml = ">=6.0"
@ -34,9 +32,6 @@ httpx = {extras = ["socks"], version = ">=0.28.1"}
typing-extensions = ">=4.9.0"
pydantic = ">=2.5.0,<3.0.0"
[tool.poetry.extras]
pymupdf = ["pymupdf"]
[tool.poetry.group.dev.dependencies]
pytest = ">=7.0"