This commit is contained in:
ldy
2026-06-17 08:38:15 -04:00
commit f13b8fc1ca
28 changed files with 894 additions and 0 deletions

View File

@@ -0,0 +1,13 @@
"""Cheap-LLM link classification for careers page and job links (Stage 2, tier 5 / Stage 3, tier 4).
Scaffold stub -- not implemented yet.
"""
# TODO (Stage 2 tier 5 / Stage 3 tier 4): implement per CLAUDE.md "Cheap-LLM classification".
# Uses Pydantic AI (model-agnostic) with the `classifier_model` from config.
# Two typed tasks:
# 1. classify_careers_link(anchors: list[Anchor]) -> CareerLinkResult
# Given extracted <a> tags from a page, pick the careers/jobs page URL.
# 2. classify_job_link(anchors: list[Anchor]) -> JobLinkResult
# Given extracted <a> tags from a careers page, pick one open-position URL.
# Both return a typed Pydantic result including the chosen URL and confidence.
# Graceful degradation: if llm_api_key is placeholder or call fails, return None.