Crate pdfmuse_core 

pdfmuse-core — deterministic PDF/DOCX parser core.

The naive parse() lands in PER-33 and the self-written content-stream interpreter (the real value) in PER-36. The unified IR — the data foundation that every binding serializes byte-identically — lives in ir.

Re-exports

pub use error::PdfmuseError;

pub use error::Result;

Modules

backend

Pluggable vision backend — the ML boundary.

error

Structured error type.

ir

Unified intermediate representation (IR).

Structs

Chunk

A retrieval unit: a block’s text plus the context needed to cite it.

Enums

Format

Source-format hint for parse.

Functions

chunk

Split doc into chunks (one per non-empty block), tracking heading context.

parse

Parse data into the unified ir::Document.

parse_with_password

Like parse, but supplies a password for encrypted PDFs.

to_json

Serialize the entire Document to pretty-printed JSON.

to_markdown

Render doc to GitHub-flavored Markdown, pages and blocks in order.

to_text

Render doc to plain reading-order text — no Markdown syntax, just the block text joined by newlines. The cheapest useful output for search / ATS / feeding an LLM, and (via the bindings) avoids materializing the full IR on the host.