"""Tests for market matching logic."""
from datetime import datetime, timezone
import pytest
from mcp_predictive_market.analysis.matching import MarketMatcher, MatchResult
from mcp_predictive_market.schema import Market
def make_market(platform: str, native_id: str, title: str) -> Market:
"""Create a test market."""
return Market(
platform=platform,
native_id=native_id,
url=f"https://{platform}.com/{native_id}",
title=title,
description="",
category="politics",
probability=0.5,
outcomes=[],
volume=1000,
liquidity=500,
created_at=datetime.now(timezone.utc),
closes_at=None,
resolved=False,
resolution=None,
last_fetched=datetime.now(timezone.utc),
price_history=[],
)
class TestMarketMatcher:
def test_find_matches_similar_titles(self):
"""Should match markets with similar titles."""
matcher = MarketMatcher()
target = make_market("manifold", "abc", "Will Trump win 2024 election?")
candidates = [
make_market("polymarket", "xyz", "Trump wins 2024 presidential election"),
make_market("kalshi", "pres", "Will Biden win 2024?"),
]
results = matcher.find_matches(target, candidates, min_confidence=0.3)
assert len(results) >= 1
assert results[0].market_b.platform == "polymarket"
assert results[0].confidence > 0.3
def test_find_matches_manual_mapping(self):
"""Should return manual mappings with confidence 1.0."""
matcher = MarketMatcher()
matcher.add_manual_mapping("manifold:abc", "polymarket:xyz")
target = make_market("manifold", "abc", "Some market")
candidates = [make_market("polymarket", "xyz", "Different title")]
results = matcher.find_matches(target, candidates)
assert len(results) == 1
assert results[0].confidence == 1.0
assert results[0].match_type == "manual"
def test_find_matches_excludes_same_market(self):
"""Should not match a market with itself."""
matcher = MarketMatcher()
market = make_market("manifold", "abc", "Test market")
results = matcher.find_matches(market, [market])
assert len(results) == 0
def test_text_similarity(self):
"""Should calculate Jaccard similarity."""
matcher = MarketMatcher()
# Identical (excluding stop words)
sim = matcher._text_similarity("Will Trump win?", "Trump win")
assert sim > 0.5
# Completely different
sim = matcher._text_similarity("Bitcoin price", "Football game")
assert sim == 0.0
def test_add_manual_mapping_bidirectional(self):
"""Should add mappings in both directions."""
matcher = MarketMatcher()
matcher.add_manual_mapping("a:1", "b:2")
assert matcher._is_manual_match("a:1", "b:2")
assert matcher._is_manual_match("b:2", "a:1")
def test_find_matches_respects_min_confidence(self):
"""Should filter out matches below min_confidence."""
matcher = MarketMatcher()
target = make_market("manifold", "abc", "Bitcoin price prediction")
candidates = [
make_market("polymarket", "xyz", "Bitcoin price forecast tomorrow"),
make_market("kalshi", "pres", "Ethereum gas fees"),
]
# With high threshold, should get fewer matches
high_threshold_results = matcher.find_matches(
target, candidates, min_confidence=0.8
)
low_threshold_results = matcher.find_matches(
target, candidates, min_confidence=0.1
)
assert len(high_threshold_results) <= len(low_threshold_results)
def test_find_matches_sorted_by_confidence(self):
"""Should return results sorted by confidence descending."""
matcher = MarketMatcher()
matcher.add_manual_mapping("manifold:abc", "polymarket:xyz")
target = make_market("manifold", "abc", "Will Trump win 2024?")
candidates = [
make_market("kalshi", "k1", "Trump wins 2024"),
make_market("polymarket", "xyz", "Some other market"), # manual match
]
results = matcher.find_matches(target, candidates, min_confidence=0.1)
# Manual match should come first with confidence 1.0
assert len(results) >= 1
assert results[0].match_type == "manual"
assert results[0].confidence == 1.0
def test_text_similarity_empty_after_stop_words(self):
"""Should return 0.0 if all words are stop words."""
matcher = MarketMatcher()
sim = matcher._text_similarity("the a an", "is to be")
assert sim == 0.0
def test_match_result_contains_markets(self):
"""Should include both markets in the result."""
matcher = MarketMatcher()
target = make_market("manifold", "abc", "Bitcoin prediction")
candidate = make_market("polymarket", "xyz", "Bitcoin prediction")
results = matcher.find_matches(target, [candidate], min_confidence=0.0)
assert len(results) == 1
assert results[0].market_a == target
assert results[0].market_b == candidate
