[xyu]

Proposed changes

serpextract no longer depends on pylru.

Why are these changes being made?

serpextract created a pylru cache at import time but never read or wrote it. That one line made pylru a dependency of every Parse.ly project that installs serpextract. Parse.ly's own code no longer uses pylru, so this is what still keeps it installed in dash, ML and the engineering monorepo.

Testing

pip install -r requirements.txt && pip install . && pytest -q in python:3.9-bullseye, with pylru confirmed absent from the environment:

pylru installed: False
Requires: iso3166, ruamel.yaml, tldextract
15 passed

Testing instructions

• Confirm nothing else references the removed cache:

  git grep -n -i lru -- '*.py' setup.py requirements.txt
  

  (no output)
  

🤖 Generated with Claude Code