page_range not working with .pptx documents

[JOSHMT0744]

Bug

when setting page_range parameter of convert function, the document converter still parses the whole document regardless.

Steps to reproduce

python```
doc = "example.pptx"
output_dir = Path("out")

doc_converter = create_doc_converter()
conv_result = doc_converter.convert(doc, page_range=(1,2))

Export Markdown format without image referencing

with (output_dir / f"{conv_result.input.file.stem}.md").open("w", encoding="utf-8") as fp:
fp.write(conv_result.document.export_to_markdown())

Given example.pptx is more than 2 pages, output will still be whole document

### Docling version
Docling version: 2.41.0
Docling Core version: 2.42.0
Docling IBM Models version: 3.8.1
Docling Parse version: 4.1.0

### Python version
Python: cpython-39 (3.9.21)

[dosubot]

Hi @JOSHMT0744! I'm Dosu and I’m helping the docling team.

The page_range parameter isn't currently implemented for .pptx files—when you set it, the backend still processes all slides regardless. This is because the MsPowerpointDocumentBackend, which handles PowerPoint files, doesn't use the page_range setting in its conversion logic, even though the pipeline passes it through and claims to support pagination. So, this is a missing feature rather than a bug or a documented limitation, and there aren't any workarounds in Docling right now. Implementing this would require updating the backend to only process the slides within the specified range. If you want pointers on where to start with a code change, let me know!
Reference: MsPowerpointDocumentBackend ignores page_range

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}   

[dolfim-ibm]

@dosu that could be a nice addition. Can you please provide the required code changes, ready to be submitted as a PR?

[dosubot]

Here are the code changes needed to add page_range support for .pptx files in MsPowerpointDocumentBackend:

1. In MsPowerpointDocumentBackend.convert, extract the page_range from self.in_doc.limits.page_range, convert it from 1-based to 0-based indices, and slice the slides accordingly.
2. Update walk_linear to accept a list of slides (or add a parameter for the filtered slides), and process only those slides.
3. Handle edge cases: if the range is out of bounds, reversed, or empty, process nothing or raise a clear error. The page_range is already validated upstream, so you mainly need to handle out-of-bounds gracefully.
4. Update/add tests in tests/test_backend_pptx.py to cover: normal usage (e.g., (1,2)), single-slide, reversed/empty, and out-of-bounds ranges.
5. Add docstrings to convert and walk_linear to document the new behavior.
6. Update user-facing docs and usage guides to mention slide-range support for PowerPoint.

Example patch for the backend (core logic):

# In MsPowerpointDocumentBackend.convert
def convert(self) -> DoclingDocument:
    origin = DocumentOrigin(
        filename=self.file.name or "file",
        mimetype="application/vnd.ms-powerpoint",
        binary_hash=self.document_hash,
    )
    doc = DoclingDocument(name=self.file.stem or "file", origin=origin)
    # Extract and convert page_range to 0-based indices
    start, end = self.in_doc.limits.page_range
    slides = list(self.pptx_obj.slides)
    # Clamp indices to valid range
    start_idx = max(0, start - 1)
    end_idx = min(len(slides), end)
    selected_slides = slides[start_idx:end_idx]
    doc = self.walk_linear(selected_slides, doc)
    return doc

# Update walk_linear to accept a list of slides
def walk_linear(self, slides, doc) -> DoclingDocument:
    # ... use 'slides' instead of 'pptx_obj.slides' in the loop ...

Tests and documentation should be updated as described above. This change is self-contained and does not affect other backends or pipelines. Let me know if you want a more detailed patch or test examples!

References:
MsPowerpointDocumentBackend ignores page_range
Validation and usage of page_range in DocumentLimits

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}