Add geothermal electricity extraction support

[bpulluta]

Overview

This PR adds geothermal electricity as a supported extraction technology in COMPASS. It includes the extraction schema and plugin configuration needed to discover, retrieve, and extract structured ordinance data from jurisdictions governing utility-scale geothermal electricity generation.

Two bugs in the retrieval layer were discovered and fixed during development. Both affected all technologies, not just geothermal.

---

New: Geothermal Electricity Extraction

Files added:

• compass/extraction/geothermal_electricity/geothermal_schema.json — defines 29 extractable features including setbacks, permitting, noise limits, zoning classifications, decommissioning, and drilling requirements
• compass/extraction/geothermal_electricity/geothermal_plugin_config.yaml — configures search queries, website scoring keywords, heuristic filters, and document collection behavior tuned for geothermal electricity ordinances

The schema follows the standard COMPASS one-shot extraction format and is compatible with the existing compass process pipeline with no code changes required.

---

Bug Fix 1 — PDF URLs with spaces failed to download

crawl4ai can return document URLs with raw spaces in the path (e.g. a county storing files under a folder named Land Use Code/):

# broken — HTTP request fails silently
https://countygov.org/Land Use Code/53007.pdf

# fixed — percent-encoded, downloads correctly
https://countygov.org/Land%20Use%20Code/53007.pdf

_sanitize_doc_sources() now percent-encodes any source attribute with raw spaces before returning from download_jurisdiction_ordinances_from_website(). Uses Python stdlib urllib.parse, no new dependencies.

File: compass/scripts/download.py

---

Bug Fix 2 — Anchor text was never used in link scoring

When the COMPASS crawler parses a page, it reads each link's visible label and stored it in the title field. The upstream scorer (ELMLinkScorer) reads the text key. These never matched, so anchor text always scored zero regardless of keyword weights — only the URL filename was used.

# before — anchor text stored in title, scorer reads text (always empty)
_Link(title=title, href=..., base_domain=...)

# after — text field populated, scorer sees the anchor label
_Link(title=title, text=title, href=..., base_domain=...)

File: compass/web/website_crawl.py

---

Tests

Two regression tests added to tests/python/unit/web/test_web_crawl.py:

• test_extract_links_from_html_sets_text_from_anchor — verifies anchor text populates both title and text on the link object
• test_compass_link_scorer_scores_anchor_text — verifies the scorer uses text when assigning keyword scores

All 30 unit tests pass.