"""Tests for journal quality ranking.
These tests verify:
1. Title normalization and abbreviation expansion
2. Journal lookup with exact, abbreviation, and fuzzy matching
3. Bundled CSV loading (Feature 4)
4. Override support (Feature 7)
5. Fuzzy threshold at 90% (Feature 7)
Tests are designed to FAIL LOUDLY if:
- Bundled CSV is missing or malformed
- Overrides file has syntax errors
- Fuzzy matching produces false positives
"""
from __future__ import annotations
import csv
import tempfile
from pathlib import Path
import pytest
from zotero_chunk_rag.journal_ranker import (
JournalRanker,
_expand_abbreviations,
_normalize_title,
)
# =============================================================================
# Test Title Normalization
# =============================================================================
class TestNormalizeTitle:
"""Test title normalization."""
def test_lowercase(self):
assert _normalize_title("Nature") == "nature"
assert _normalize_title("SCIENCE") == "science"
def test_punctuation_to_spaces(self):
assert _normalize_title("A & B") == "a b"
assert _normalize_title("A: B") == "a b"
assert _normalize_title("A-B") == "a b"
assert _normalize_title("A/B") == "a b"
def test_collapse_spaces(self):
assert _normalize_title("A B") == "a b"
assert _normalize_title("A & B") == "a b"
def test_strip(self):
assert _normalize_title(" Nature ") == "nature"
# =============================================================================
# Test Abbreviation Expansion
# =============================================================================
class TestExpandAbbreviations:
"""Test abbreviation expansion."""
def test_single_abbreviation(self):
expansions = _expand_abbreviations("J. Medicine")
assert "journal medicine" in expansions
def test_multiple_abbreviations(self):
expansions = _expand_abbreviations("IEEE Trans. Biomed. Eng.")
# Should contain full expansion
assert any(
"transactions" in e and "biomedical" in e and "engineering" in e
for e in expansions
)
def test_no_abbreviations(self):
expansions = _expand_abbreviations("Nature")
assert "nature" in expansions
# =============================================================================
# Test JournalRanker with Test CSV
# =============================================================================
class TestJournalRanker:
"""Test the journal ranker with a test CSV."""
@pytest.fixture
def test_csv(self, tmp_path: Path) -> Path:
"""Create a test CSV file."""
csv_path = tmp_path / "test_scimago.csv"
with open(csv_path, "w", newline="", encoding="utf-8") as f:
writer = csv.writer(f)
writer.writerow(["title_normalized", "quartile"])
writer.writerow(["nature", "Q1"])
writer.writerow(["science", "Q1"])
writer.writerow(["plos one", "Q1"])
# Note: abbreviation expander produces "ieee transactions biomedical engineering"
# (without "on"), so we match that form
writer.writerow(["ieee transactions biomedical engineering", "Q1"])
writer.writerow(["journal of physiology", "Q2"])
writer.writerow(["medical hypotheses", "Q4"])
return csv_path
def test_exact_match(self, test_csv: Path):
ranker = JournalRanker(test_csv)
assert ranker.lookup("Nature") == "Q1"
assert ranker.lookup("NATURE") == "Q1"
assert ranker.lookup("science") == "Q1"
def test_abbreviation_expansion(self, test_csv: Path):
"""Test that abbreviations are expanded and matched."""
ranker = JournalRanker(test_csv)
# "IEEE Trans. Biomed. Eng." expands to "ieee transactions biomedical engineering"
result = ranker.lookup("IEEE Trans. Biomed. Eng.")
assert result == "Q1"
def test_fuzzy_match_requires_90_percent(self, test_csv: Path):
"""Fuzzy matching should require 90% similarity (Feature 7)."""
ranker = JournalRanker(test_csv)
# "Journal of Physiology" is exact match
result = ranker.lookup("Journal of Physiology")
assert result == "Q2"
def test_fuzzy_match_rejects_low_similarity(self, test_csv: Path):
"""Fuzzy matching should reject matches below 90% threshold."""
ranker = JournalRanker(test_csv)
# This is too different to match anything at 90%
result = ranker.lookup("Journal of Something Completely Different")
assert result is None, "Low similarity should not match at 90% threshold"
def test_no_match(self, test_csv: Path):
ranker = JournalRanker(test_csv)
assert ranker.lookup("Unknown Journal of Stuff") is None
def test_empty_publication(self, test_csv: Path):
ranker = JournalRanker(test_csv)
assert ranker.lookup("") is None
assert ranker.lookup(None) is None
def test_caching(self, test_csv: Path):
ranker = JournalRanker(test_csv)
# First call
result1 = ranker.lookup("Nature")
# Second call should be cached
result2 = ranker.lookup("Nature")
assert result1 == result2 == "Q1"
assert "Nature" in ranker._cache
def test_loaded_property(self, test_csv: Path):
ranker = JournalRanker(test_csv)
assert ranker.loaded is True
def test_stats(self, test_csv: Path):
ranker = JournalRanker(test_csv)
stats = ranker.stats()
assert stats["total_journals"] == 6
assert stats["quartile_counts"]["Q1"] == 4
assert stats["quartile_counts"]["Q2"] == 1
assert stats["quartile_counts"]["Q4"] == 1
def test_missing_csv(self, tmp_path: Path):
"""Test graceful handling of missing CSV."""
ranker = JournalRanker(tmp_path / "nonexistent.csv")
assert ranker.loaded is False
assert ranker.lookup("Nature") is None
# =============================================================================
# Test Override Support (Feature 7)
# =============================================================================
class TestJournalOverrides:
"""Test manual override functionality."""
@pytest.fixture
def test_csv(self, tmp_path: Path) -> Path:
"""Create a test CSV file."""
csv_path = tmp_path / "test_scimago.csv"
with open(csv_path, "w", newline="", encoding="utf-8") as f:
writer = csv.writer(f)
writer.writerow(["title_normalized", "quartile"])
writer.writerow(["nature", "Q1"])
writer.writerow(["wrong journal match", "Q4"]) # Will be overridden
return csv_path
@pytest.fixture
def test_overrides(self, tmp_path: Path) -> Path:
"""Create a test overrides file."""
overrides_path = tmp_path / "overrides.csv"
with open(overrides_path, "w", encoding="utf-8") as f:
f.write("# This is a comment\n")
f.write("Correct Journal Name,Q1\n")
f.write("# Another comment\n")
f.write("wrong journal match,Q2\n") # Override the Q4 from main CSV
return overrides_path
def test_override_takes_precedence(self, test_csv: Path, test_overrides: Path):
"""Overrides should take precedence over SCImago data."""
ranker = JournalRanker(test_csv, overrides_path=test_overrides)
# "wrong journal match" is Q4 in CSV but Q2 in overrides
result = ranker.lookup("wrong journal match")
assert result == "Q2", "Override should take precedence over SCImago"
def test_override_exact_match(self, test_csv: Path, test_overrides: Path):
"""Overrides should work with exact matching."""
ranker = JournalRanker(test_csv, overrides_path=test_overrides)
result = ranker.lookup("Correct Journal Name")
assert result == "Q1"
def test_override_with_abbreviation_expansion(
self, test_csv: Path, tmp_path: Path
):
"""Overrides should work after abbreviation expansion."""
overrides_path = tmp_path / "overrides_abbrev.csv"
with open(overrides_path, "w", encoding="utf-8") as f:
# Note: "J. Biomed. Eng." expands to "journal biomedical engineering"
# (without "of"), so override must match the expanded form
f.write("journal biomedical engineering,Q1\n")
ranker = JournalRanker(test_csv, overrides_path=overrides_path)
# "J. Biomed. Eng." should expand and match override
result = ranker.lookup("J. Biomed. Eng.")
assert result == "Q1"
def test_empty_overrides_file(self, test_csv: Path, tmp_path: Path):
"""Empty overrides file should not cause errors."""
overrides_path = tmp_path / "empty_overrides.csv"
overrides_path.write_text("# Just comments\n# No actual overrides\n")
ranker = JournalRanker(test_csv, overrides_path=overrides_path)
assert ranker.loaded is True
assert ranker.lookup("Nature") == "Q1"
def test_missing_overrides_file(self, test_csv: Path, tmp_path: Path):
"""Missing overrides file should not cause errors."""
ranker = JournalRanker(
test_csv, overrides_path=tmp_path / "nonexistent.csv"
)
assert ranker.loaded is True
assert ranker.lookup("Nature") == "Q1"
# =============================================================================
# Test Bundled CSV (Feature 4)
# =============================================================================
class TestBundledCSV:
"""Test that bundled CSV is accessible and valid.
THESE TESTS FAIL LOUDLY if the bundled CSV is missing or malformed.
"""
def test_bundled_csv_exists(self):
"""Bundled CSV must exist in package data directory."""
import zotero_chunk_rag
csv_path = (
Path(zotero_chunk_rag.__file__).parent / "data" / "scimago_quartiles.csv"
)
assert csv_path.exists(), (
f"CRITICAL: Bundled SCImago CSV not found at {csv_path}. "
"Feature 4 requires this file to be present. "
"Run: python scripts/prepare_scimago.py --input <scimago_download.csv> "
f"--output {csv_path}"
)
def test_bundled_csv_is_valid(self):
"""Bundled CSV must have correct format and data."""
import zotero_chunk_rag
csv_path = (
Path(zotero_chunk_rag.__file__).parent / "data" / "scimago_quartiles.csv"
)
if not csv_path.exists():
pytest.fail(
f"Bundled CSV not found at {csv_path}. "
"Cannot verify format."
)
with open(csv_path, "r", encoding="utf-8") as f:
reader = csv.DictReader(f)
# Verify header columns
assert reader.fieldnames is not None, "CSV has no header"
assert "title_normalized" in reader.fieldnames, (
"CSV missing 'title_normalized' column"
)
assert "quartile" in reader.fieldnames, (
"CSV missing 'quartile' column"
)
# Verify data
row_count = 0
quartile_counts = {"Q1": 0, "Q2": 0, "Q3": 0, "Q4": 0}
for row in reader:
row_count += 1
title = row.get("title_normalized", "").strip()
quartile = row.get("quartile", "").strip()
assert title, f"Row {row_count} has empty title"
assert quartile in ("Q1", "Q2", "Q3", "Q4"), (
f"Row {row_count} has invalid quartile: {quartile}"
)
quartile_counts[quartile] += 1
assert row_count >= 10, (
f"Bundled CSV only has {row_count} journals. "
"Expected at least 10 for minimal functionality."
)
def test_bundled_csv_loads_automatically(self):
"""JournalRanker should load bundled data automatically."""
ranker = JournalRanker() # No path specified
assert ranker.loaded, (
"JournalRanker should load bundled CSV automatically. "
"Ensure data/scimago_quartiles.csv exists."
)
def test_bundled_csv_has_common_journals(self):
"""Bundled CSV should include common high-impact journals."""
ranker = JournalRanker()
assert ranker.loaded, (
"CRITICAL: JournalRanker failed to load bundled CSV. "
"Ensure data/scimago_quartiles.csv exists in package."
)
# These journals should definitely be in the bundled data
common_journals = [
("Nature", "Q1"),
("Science", "Q1"),
("PLOS ONE", "Q1"),
]
for journal, expected_quartile in common_journals:
result = ranker.lookup(journal)
assert result == expected_quartile, (
f"Expected {journal} to be {expected_quartile}, got {result}. "
"Bundled CSV may be incomplete."
)
def test_bundled_overrides_exists(self):
"""Bundled overrides CSV must exist (can be empty)."""
import zotero_chunk_rag
overrides_path = (
Path(zotero_chunk_rag.__file__).parent / "data" / "journal_overrides.csv"
)
assert overrides_path.exists(), (
f"CRITICAL: Bundled overrides CSV not found at {overrides_path}. "
"Feature 7 requires this file to be present (can be empty with comments)."
)
# =============================================================================
# Test Fuzzy Match Threshold (Feature 7)
# =============================================================================
class TestFuzzyThreshold:
"""Test that fuzzy matching uses 90% threshold.
THESE TESTS FAIL LOUDLY if false positives occur.
"""
@pytest.fixture
def ranker_with_known_journals(self, tmp_path: Path) -> JournalRanker:
"""Create a ranker with specific journals for threshold testing."""
csv_path = tmp_path / "threshold_test.csv"
with open(csv_path, "w", newline="", encoding="utf-8") as f:
writer = csv.writer(f)
writer.writerow(["title_normalized", "quartile"])
writer.writerow(["journal of cardiology", "Q1"])
writer.writerow(["european journal of cardiology", "Q2"])
return JournalRanker(csv_path)
def test_exact_match_works(self, ranker_with_known_journals: JournalRanker):
"""Exact matches should work."""
assert ranker_with_known_journals.lookup("Journal of Cardiology") == "Q1"
def test_near_exact_match_works(self, ranker_with_known_journals: JournalRanker):
"""Very similar names should match."""
# "journal of cardiology" vs "Journal of Cardiology" - should match
assert ranker_with_known_journals.lookup("Journal Of Cardiology") == "Q1"
def test_false_positive_prevention(
self, ranker_with_known_journals: JournalRanker
):
"""Different journals should NOT match due to 90% threshold.
At 85% threshold, "Journal of Cardiology" might match
"Journal of Neurology" (they share "Journal of ____ology").
At 90%, this should be rejected.
"""
# This is deliberately different and should NOT match
result = ranker_with_known_journals.lookup("Journal of Neurology")
assert result is None, (
"FALSE POSITIVE: 'Journal of Neurology' incorrectly matched. "
"Fuzzy threshold may be too low."
)
def test_very_different_no_match(
self, ranker_with_known_journals: JournalRanker
):
"""Very different names should definitely not match."""
result = ranker_with_known_journals.lookup("Proceedings of the ACM")
assert result is None
# =============================================================================
# Test Hot Reload (existing feature)
# =============================================================================
class TestHotReload:
"""Test hot reload functionality."""
def test_is_stale_after_modification(self, tmp_path: Path):
"""is_stale should detect when CSV has been modified."""
csv_path = tmp_path / "test.csv"
with open(csv_path, "w", newline="", encoding="utf-8") as f:
writer = csv.writer(f)
writer.writerow(["title_normalized", "quartile"])
writer.writerow(["nature", "Q1"])
ranker = JournalRanker(csv_path)
assert not ranker.is_stale()
# Modify the file
import time
time.sleep(0.1) # Ensure mtime changes
with open(csv_path, "a", encoding="utf-8") as f:
f.write("science,Q1\n")
assert ranker.is_stale()
def test_reload_if_stale(self, tmp_path: Path):
"""reload_if_stale should reload modified CSV."""
csv_path = tmp_path / "test.csv"
with open(csv_path, "w", newline="", encoding="utf-8") as f:
writer = csv.writer(f)
writer.writerow(["title_normalized", "quartile"])
writer.writerow(["nature", "Q1"])
ranker = JournalRanker(csv_path)
assert ranker.lookup("Science") is None
# Modify the file to add Science
import time
time.sleep(0.1)
with open(csv_path, "a", encoding="utf-8") as f:
f.write("science,Q1\n")
# Reload
reloaded = ranker.reload_if_stale()
assert reloaded is True
# Now Science should be found
assert ranker.lookup("Science") == "Q1"
