fix: stabilize test_text_query_word_weights by using component assertions

[thakoreh]

Summary

• Fixes Fix flaky test_text_query_word_weights and re-enable test #497
• Replaced exact string match with component-based assertions
• Removes @pytest.mark.skip decorator

Problem

The test was flaky because it relied on exact string matching with specific token ordering. Dict iteration order could cause tokens to appear in different positions across Python versions.

Solution

Check for expected query components individually rather than matching the entire string exactly:

# Before (flaky)
assert str(query) == "@description:(query | string | alpha=>{$weight:2} | ...) LIMIT 0 10"

# After (stable)
query_str = str(query)
assert "@description:(" in query_str
assert "alpha=>{$weight:2}" in query_str
assert "SCORER BM25STD" in query_str
# ... etc

Testing

• Test now passes consistently regardless of dict ordering
• All existing test assertions preserved (value checks, error cases)

---

Note

Low Risk
Low risk: only modifies unit test assertions to be less order-dependent, with no production code changes.

Overview
Re-enables test_text_query_word_weights by removing the skip marker and replacing an exact full query-string equality check with component-based in assertions.

This makes the test resilient to non-deterministic token ordering (e.g., dict iteration) while still verifying weighted tokens and key query options (scorer, dialect, limit) are present.

^{Written by Cursor Bugbot for commit c9f9a0d. This will update automatically on new commits. Configure here.}

[Copilot]

Pull request overview

Stabilizes the previously flaky test_text_query_word_weights unit test by removing an exact full-string assertion and re-enabling the test in CI (fix for #497).

Changes:

• Removed the @pytest.mark.skip decorator from test_text_query_word_weights.
• Replaced an exact str(query) == ... assertion with component-based substring assertions.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

These substring assertions are much less strict than the previous exact-match and could pass even if key semantics regress (e.g., only one of the two alpha occurrences is weighted, or the @description:(...) clause is malformed). Consider adding stable-but-strong checks like counting weighted occurrences (e.g., alpha appears twice), and asserting structure (e.g., startswith @description:( and that the text clause closes before SCORER).

Suggested change

	# Check query components without relying on exact token ordering
	query_str = str(query)
	assert "@description:(" in query_str
	assert "alpha=>{$weight:2}" in query_str
	assert "delta=>{$weight:0.555}" in query_str
	assert "query" in query_str
	assert "string" in query_str
	assert "bravo" in query_str
	assert "tango" in query_str
	# Check query components without relying on exact token ordering,
	# but with stronger structural guarantees.
	query_str = str(query)
	# The description clause should be present and properly delimited
	assert "@description:(" in query_str
	description_start = query_str.index("@description:(")
	description_close = query_str.index(")", description_start)
	scorer_pos = query_str.index("SCORER")
	# Ensure the text clause closes before the SCORER section
	assert description_close < scorer_pos
	# The weighted tokens should appear the expected number of times
	assert query_str.count("alpha=>{$weight:2}") == 2
	assert query_str.count("delta=>{$weight:0.555}") == 1
	# Ensure the main text tokens appear inside the description clause
	description_clause = query_str[description_start:description_close]
	assert "query" in description_clause
	assert "string" in description_clause
	assert "bravo" in description_clause
	assert "tango" in description_clause
	# Other expected query modifiers