[Question]: when parsing a PDF, is it possible to include page number, into individual node's meta data ?

[jzhao62]

Question Validation

• I have searched both the documentation and discord for an answer.

Question

• as the question goes, it would be great if each node's meta data could contain page Number that points to its origianl occurance in the PDF

[dosubot]

Yes, when parsing a PDF using the PyMuPDFReader class, you can include the page number in each individual node's metadata. This is done by adding the page number to the extra_info dictionary with the key "source". Here's a code snippet that demonstrates how to do this:

return [
    Document(
        text=page.get_text().encode("utf-8"),
        extra_info=dict(
            extra_info,
            **{
                "source": f"{page.number+1}",
            },
        ),
    )
    for page in doc
]

In this snippet, page.number+1 is used to get the page number (starting from 1) and include it in the metadata for each document node ^[1].

To continue talking to Dosu, mention @dosu.

---

^{Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Other}