EPUB Parsing & Rendering Engine

Every e-reader, library app, and conversion pipeline has to answer the same question: given a.epub file from anywhere, how do you turn it into pages a person can actually read? This handbook answers it end to end, from the first byte of the ZIP container to a working reference engine you can run yourself. You'll open the container correctly — and see exactly how a miscompressed mimetype file breaks readers silently — parse the OPF package document's metadata, manifest, and spine, and build a unified table of contents from both the EPUB 3 navigation document and its EPUB 2 NCX predecessor. From there, the book treats every content document as what it actually is in the wild: untrusted HTML that has to be sanitized against script injection, tracking beacons, and path-traversal attacks before it's safe to render. The second half covers what most tutorials skip. The EPUB CSS profile, and how the cascade really resolves across reading-system, author, and user stylesheets. The exact IDPF font-obfuscation algorithm, byte-for-byte, so embedded fonts stop failing with no explanation. Pagination for reflowable text and the EPUB CFI addressing scheme that keeps a reader's place through a font-size change. Fixed-layout rendition, synthetic spreads, and the accessibility gap fixed layout introduces. Validation against EPUBCheck, and the performance discipline a library of a few thousand books actually demands. Every technique is demonstrated in real, runnable Python, including a compact reference engine in the closing chapter that ties the whole pipeline together — tested against this book's own EPUB file. Written for engineers building a reading app, a documentation viewer, a conversion tool, or QA tooling around the EPUB format, not for casual readers wondering how their e-reader works.