"""Wording must reach a frequency through the source's own metadata.

Baseline captured live on 2026-08-03 against the hosted Space: "typewriter" hit
oz.typewriter.posts, but "Robert", "Robert Messenger", "oztypewriter.blogspot.com"
and "typewriter history blog" all missed — every one of them a word that appears
in the signal's metadata but not in the frequency name.

Run: python3 -m pytest tests/test_alias_resolution.py
"""
import sys
from pathlib import Path

sys.path.insert(0, str(Path(__file__).resolve().parent.parent))

from core.signal_registry import SignalRegistry


def _mesh():
    SignalRegistry._instance = None
    reg = SignalRegistry()
    for title in ("The Ongoing Popularity of Portable Typewriters",
                  "THE UPPER AND LOWER CASE FOR (TYPEWRITTEN) LETTERS"):
        reg.broadcast(
            name="oz.typewriter.posts", source_type="feed_hydrator",
            data={"title": title, "link": "https://oztypewriter.blogspot.com/x.html",
                  "feed": "oz.Typewriter — Robert Messenger", "summary": ""},
            metadata={"feed_url": "https://oztypewriter.blogspot.com/feeds/posts/default"},
        )
    # Decoys: resolution must stay narrow, not answer everything with everything.
    reg.broadcast(name="aws_blog", source_type="feed_hydrator",
                  data={"title": "New in S3", "feed": "AWS Blog"},
                  metadata={"feed_url": "https://aws.amazon.com/blogs/aws/feed/"})
    reg.broadcast(name="krebs", source_type="feed_hydrator",
                  data={"title": "Phishing wave", "feed": "Krebs on Security"},
                  metadata={"feed_url": "https://krebsonsecurity.com/feed/"})
    return reg


def _hits(reg, kw):
    return {s["name"] for s in reg.tune_in([kw])}


def test_frequency_name_still_resolves():
    reg = _mesh()
    for kw in ("oz.typewriter.posts", "typewriter", "oz"):
        assert "oz.typewriter.posts" in _hits(reg, kw), kw


def test_author_resolves():
    """'What has Robert posted lately?' — the case that motivated all of this."""
    reg = _mesh()
    for kw in ("Robert Messenger", "What has Robert posted lately",
               "Robert Messenger's Typewriting Blog"):
        assert "oz.typewriter.posts" in _hits(reg, kw), kw


def test_hostname_resolves():
    reg = _mesh()
    assert "oz.typewriter.posts" in _hits(reg, "oztypewriter.blogspot.com")


def test_article_title_resolves():
    reg = _mesh()
    assert "oz.typewriter.posts" in _hits(reg, "upper and lower case for typewritten letters")


def test_download_phrasing_resolves():
    """Wording is irrelevant: 'tell me' / 'update me' / 'download' all route the same."""
    reg = _mesh()
    for kw in ("download Robert's latest typewriter posts",
               "update me on Robert Messenger",
               "tell me about oztypewriter.blogspot.com"):
        assert "oz.typewriter.posts" in _hits(reg, kw), kw


def test_resolution_stays_narrow():
    """An alias index that matches everything is worse than no alias index."""
    reg = _mesh()
    assert _hits(reg, "Krebs on Security") == {"krebs"}
    assert "oz.typewriter.posts" not in _hits(reg, "aws_blog")
    assert "krebs" not in _hits(reg, "Robert Messenger")


def test_honest_miss():
    reg = _mesh()
    assert _hits(reg, "quarterly revenue forecast") == set()


def test_tokens_split_dots_and_camel():
    t = SignalRegistry._tokens
    assert t("oz.Typewriter") == {"oz", "typewriter"}
    assert t("oz.typewriter.posts") == {"oz", "typewriter", "posts"}
    assert t("Robert Messenger's Typewriting Blog") == \
        {"robert", "messenger", "s", "typewriting", "blog"}


def test_trail_keyed_by_meaning_not_wording():
    """Two phrasings of one ask share a trail instead of each paying to learn it."""
    reg = _mesh()
    assert reg._trail_key("Robert's latest posts") == \
           reg._trail_key("latest posts by Robert")


if __name__ == "__main__":
    import pytest
    raise SystemExit(pytest.main([__file__, "-q"]))