"""Tests for VaultIndexer."""
from pathlib import Path
from unittest.mock import patch
import numpy as np
class TestVaultIndexerInit:
"""Tests for VaultIndexer initialization."""
def test_accepts_single_string_path(self, temp_vault: Path) -> None:
"""Test backward compatibility with single string path."""
with patch("semantic_search_mcp.indexer.SentenceTransformer") as mock_st:
mock_st.return_value.get_sentence_embedding_dimension.return_value = 384
mock_st.return_value.encode.return_value = np.array([[0.1] * 384])
from semantic_search_mcp.indexer import VaultIndexer
indexer = VaultIndexer(str(temp_vault))
assert len(indexer.vault_paths) == 1
assert indexer.vault_paths[0] == temp_vault
def test_accepts_list_of_paths(self, multi_vaults: list[Path]) -> None:
"""Test multiple paths as list."""
with patch("semantic_search_mcp.indexer.SentenceTransformer") as mock_st:
mock_st.return_value.get_sentence_embedding_dimension.return_value = 384
mock_st.return_value.encode.return_value = np.array([[0.1] * 384])
from semantic_search_mcp.indexer import VaultIndexer
paths = [str(v) for v in multi_vaults]
indexer = VaultIndexer(paths)
assert len(indexer.vault_paths) == 2
def test_creates_unique_index_dir_per_path_combination(self, multi_vaults: list[Path]) -> None:
"""Test different path combinations get different index directories."""
with patch("semantic_search_mcp.indexer.SentenceTransformer") as mock_st:
mock_st.return_value.get_sentence_embedding_dimension.return_value = 384
mock_st.return_value.encode.return_value = np.array([[0.1] * 384])
from semantic_search_mcp.indexer import VaultIndexer
indexer1 = VaultIndexer([str(multi_vaults[0])])
indexer2 = VaultIndexer([str(v) for v in multi_vaults])
# Different paths should have different content hashes
assert indexer1.index_dir.parent.name != indexer2.index_dir.parent.name
class TestVaultIndexerRebuild:
"""Tests for index rebuilding."""
def test_indexes_files_from_all_vaults(self, multi_vaults: list[Path]) -> None:
"""Test rebuild_index scans all configured directories."""
with patch("semantic_search_mcp.indexer.SentenceTransformer") as mock_st:
mock_st.return_value.get_sentence_embedding_dimension.return_value = 384
mock_st.return_value.encode.return_value = np.array([[0.1] * 384])
from semantic_search_mcp.indexer import VaultIndexer
paths = [str(v) for v in multi_vaults]
indexer = VaultIndexer(paths)
# Should have indexed files from both vaults
assert len(indexer.meta) == 2
class TestVaultIndexerFindDuplicates:
"""Tests for duplicate detection."""
def test_resolves_relative_path_against_all_vaults(self, multi_vaults: list[Path]) -> None:
"""Test relative paths are checked against all vault directories."""
with patch("semantic_search_mcp.indexer.SentenceTransformer") as mock_st:
mock_st.return_value.get_sentence_embedding_dimension.return_value = 384
mock_st.return_value.encode.return_value = np.array([[0.1] * 384])
from semantic_search_mcp.indexer import VaultIndexer
paths = [str(v) for v in multi_vaults]
indexer = VaultIndexer(paths)
# Relative path should be found in second vault
result = indexer.find_duplicates("note1.md")
# Should not return error
assert not isinstance(result, dict) or "error" not in result
class TestVaultIndexerInlineTags:
"""Tests for inline tag extraction."""
def test_extracts_inline_tags_from_body(self, tmp_path: Path) -> None:
"""Test inline #tags extracted and merged with frontmatter tags."""
with patch("semantic_search_mcp.indexer.SentenceTransformer") as mock_st:
mock_st.return_value.get_sentence_embedding_dimension.return_value = 384
mock_st.return_value.encode.return_value = np.array([[0.1] * 384])
from semantic_search_mcp.indexer import VaultIndexer
vault = tmp_path / "vault"
vault.mkdir()
# Create test file with frontmatter tags and inline #tags
test_file = vault / "test.md"
test_file.write_text("""---
title: Test Note
tags: [frontmatter-tag, duplicate]
---
# Test Note
This note has #inline-tag and #duplicate tags.
Also testing #EUR/USD format.
""")
indexer = VaultIndexer(str(vault))
# Verify tags were extracted and merged
weighted_text = indexer._prepare_text_for_embedding(test_file, test_file.read_text())
# All tags should appear in the weighted text (lowercase, deduplicated)
assert "frontmatter-tag" in weighted_text.lower()
assert "inline-tag" in weighted_text.lower()
assert "eur/usd" in weighted_text.lower()
# Duplicate should only appear once in the merged set
# (checking exact count is harder due to weighting, but it should be present)
assert "duplicate" in weighted_text.lower()
def test_inline_tag_pattern_ignores_headers(self, tmp_path: Path) -> None:
"""Test that ## headers are not extracted as tags."""
with patch("semantic_search_mcp.indexer.SentenceTransformer") as mock_st:
mock_st.return_value.get_sentence_embedding_dimension.return_value = 384
mock_st.return_value.encode.return_value = np.array([[0.1] * 384])
from semantic_search_mcp.indexer import VaultIndexer
vault = tmp_path / "vault"
vault.mkdir()
test_file = vault / "test.md"
test_file.write_text("""# Header One
## Header Two
This has #real-tag but headers are not tags.
""")
indexer = VaultIndexer(str(vault))
inline_tags = indexer._extract_inline_tags(test_file.read_text())
# Should only extract #real-tag, not headers
assert "real-tag" in inline_tags
assert "Header" not in " ".join(inline_tags)
assert len(inline_tags) == 1
def test_frontmatter_single_string_tag(self, tmp_path: Path) -> None:
"""Test frontmatter tags as single string (not list)."""
with patch("semantic_search_mcp.indexer.SentenceTransformer") as mock_st:
mock_st.return_value.get_sentence_embedding_dimension.return_value = 384
mock_st.return_value.encode.return_value = np.array([[0.1] * 384])
from semantic_search_mcp.indexer import VaultIndexer
vault = tmp_path / "vault"
vault.mkdir()
test_file = vault / "test.md"
test_file.write_text("""---
tags: single-tag
---
Content with #inline-tag
""")
indexer = VaultIndexer(str(vault))
weighted_text = indexer._prepare_text_for_embedding(test_file, test_file.read_text())
assert "single-tag" in weighted_text.lower()
assert "inline-tag" in weighted_text.lower()
def test_case_insensitive_deduplication(self, tmp_path: Path) -> None:
"""Test tags deduplicated case-insensitively."""
with patch("semantic_search_mcp.indexer.SentenceTransformer") as mock_st:
mock_st.return_value.get_sentence_embedding_dimension.return_value = 384
mock_st.return_value.encode.return_value = np.array([[0.1] * 384])
from semantic_search_mcp.indexer import VaultIndexer
vault = tmp_path / "vault"
vault.mkdir()
test_file = vault / "test.md"
test_file.write_text("""---
tags: [Trading]
---
Content with #trading and #TRADING
""")
indexer = VaultIndexer(str(vault))
weighted_text = indexer._prepare_text_for_embedding(test_file, test_file.read_text())
# All variations should be present but deduplicated to lowercase
assert "trading" in weighted_text.lower()
# Count occurrences (difficult due to weighting, so just verify present)
def test_no_frontmatter_section(self, tmp_path: Path) -> None:
"""Test file without frontmatter section."""
with patch("semantic_search_mcp.indexer.SentenceTransformer") as mock_st:
mock_st.return_value.get_sentence_embedding_dimension.return_value = 384
mock_st.return_value.encode.return_value = np.array([[0.1] * 384])
from semantic_search_mcp.indexer import VaultIndexer
vault = tmp_path / "vault"
vault.mkdir()
test_file = vault / "test.md"
test_file.write_text("""# Simple Note
No frontmatter here, just #inline-tag and #another-tag.
""")
indexer = VaultIndexer(str(vault))
weighted_text = indexer._prepare_text_for_embedding(test_file, test_file.read_text())
assert "inline-tag" in weighted_text.lower()
assert "another-tag" in weighted_text.lower()
def test_tags_with_special_chars(self, tmp_path: Path) -> None:
"""Test tags with hyphens, underscores, and slashes."""
with patch("semantic_search_mcp.indexer.SentenceTransformer") as mock_st:
mock_st.return_value.get_sentence_embedding_dimension.return_value = 384
mock_st.return_value.encode.return_value = np.array([[0.1] * 384])
from semantic_search_mcp.indexer import VaultIndexer
vault = tmp_path / "vault"
vault.mkdir()
test_file = vault / "test.md"
test_file.write_text("""---
---
Testing #test-tag and #test_tag and #EUR/USD
""")
indexer = VaultIndexer(str(vault))
inline_tags = indexer._extract_inline_tags(test_file.read_text())
assert "test-tag" in inline_tags
assert "test_tag" in inline_tags
assert "EUR/USD" in inline_tags
assert len(inline_tags) == 3
