serp-zipfile

ZIP archives for Serpentine: read and write archives with stored (method 0) and DEFLATE (method 8) entries, UTF-8 names and CRC-verified extraction. Pure Serpentine ([P]) over serp-zlib — the same source dual-runs under CPython.

from serp_zipfile import ZipEntry, zip_write, zip_read, zip_extract, zip_names
from serpentine import move

entries: list[ZipEntry] = []
e: ZipEntry = ZipEntry("hello.txt", b"hello zip")
entries.append(move(e))
zip_write("out.zip", entries, True)

back: list[ZipEntry] = zip_read("out.zip")
print(zip_names(back))
print(zip_extract(back, "hello.txt"))

API

• ZipEntry(name: str, data: bytes) — one archive member
• zip_bytes(entries, compress) -> bytes / zip_parse(data) -> list[ZipEntry] — in-memory
• zip_write(path, entries, compress) / zip_read(path) -> list[ZipEntry] — on disk
• zip_names(entries) -> list[str], zip_has(entries, name) -> bool, zip_extract(entries, name) -> bytes

With compress=True each entry is deflated and kept only if smaller than the raw data (CPython zipfile behaves the same way). Written archives use a fixed DOS timestamp (1980-01-01) so output is deterministic.

Limits

No zip64 (4 GiB / 65535-entry caps), no encryption, no multi-disk archives, no on-the-fly streaming. Archives produced here open fine in CPython's zipfile and vice versa; exact archive bytes differ between worlds because the DEFLATE encoders differ.