Repo cleanup and code housekeeping

.github/workflows/ci.yml

@@ -20,13 +20,13 @@ jobs:
       - name: Install dependencies
         run: |
           pip install ruff black mypy
-          pip install -r requirements.txt
+          pip install $(python3 -c "import tomllib; print(' '.join(tomllib.load(open('pyproject.toml','rb'))['project']['dependencies']))")
 
       - name: Run ruff
-        run: ruff check src/ tests/
+        run: ruff check src/ tests/ scripts/
 
       - name: Run black (check only)
-        run: black --check src/ tests/
+        run: black --check src/ tests/ scripts/
 
       - name: Run mypy
         run: mypy src/ --ignore-missing-imports
@@ -57,7 +57,7 @@ jobs:
           python-version: "3.11"
 
       - name: Install dependencies
-        run: pip install -r requirements.txt
+        run: pip install $(python3 -c "import tomllib; print(' '.join(tomllib.load(open('pyproject.toml','rb'))['project']['dependencies']))") pytest pytest-asyncio
 
       - name: Initialize database schema
         env:

.gitignore

@@ -25,6 +25,10 @@ build/
 .coverage
 htmlcov/
 
+# Type checking / linting caches
+.mypy_cache/
+.ruff_cache/
+
 # IDE
 .vscode/
 .idea/

CHANGELOG.md

@@ -1,6 +1,6 @@
 # Changelog
 
-All notable changes to the Obsidian Graph MCP Server will be documented in this file.
+All notable changes to Obsidian Graph will be documented in this file.
 
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
@@ -53,18 +53,22 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ### Fixed
 - Stale database entries no longer persist after file deletions (Issue #2)
 - File moves now update paths correctly instead of creating duplicates
+- Embedding token limit errors on large/dense notes: dynamic batch sizing with retry-halving (#7)
+- Hub notes returning empty on first call: inline-await refresh instead of fire-and-forget (#8)
+- Event loop blocking during embedding API calls: async embedder methods (#9)
+- Missing database timeouts on 6 vector store methods (#9)
+- Schema trigger overwriting file modification times (#10)
+- Hub analyzer raising wrong exception type (#10)
+- Dead code cleanup and weak test assertions (#10)
 
 ### Changed
 - File watcher now defaults to polling mode in Docker (native filesystem events unreliable)
 - Startup scan now cleans up orphan paths before re-indexing stale files
-
-## [Unreleased - Planned]
+- Renamed project from "MCP Server" to "Obsidian Graph" (semantic knowledge graph engine)
+- Container names: obsidian-graph (app), obsidian-graph-pgvector (db)
 
 ### Planned
-- Additional embedding provider support (OpenAI, SentenceTransformers)
+- Separate src/ into engine/ and mcp/ packages
+- Additional embedding provider support
 - Cluster analysis tool (community detection)
-- Temporal statistics tool
-- Embedding validation tool
 - Performance optimizations for large vaults (>10k notes)
-- GitHub Actions CI/CD
-- Code coverage reporting

Dockerfile

@@ -5,16 +5,12 @@ WORKDIR /app
 # Create non-root user
 RUN useradd -m -u 1000 mcpuser
 
-# Install dependencies
-COPY requirements.txt .
-RUN pip install --no-cache-dir -r requirements.txt
+# Install dependencies from pyproject.toml (deps only, not the package itself)
+COPY pyproject.toml .
+RUN pip install --no-cache-dir $(python3 -c "import tomllib; print(' '.join(tomllib.load(open('pyproject.toml','rb'))['project']['dependencies']))")
 
-# Copy source code and tests and set ownership
+# Copy source code and set ownership
 COPY --chown=mcpuser:mcpuser src/ ./src/
-COPY --chown=mcpuser:mcpuser tests/ ./tests/
-
-# Copy configuration files for testing
-COPY --chown=mcpuser:mcpuser pyproject.toml mypy.ini ./
 
 # Create directories for data and cache
 RUN mkdir -p /home/mcpuser/.obsidian-graph/cache && \

README.md

@@ -217,10 +217,6 @@ POSTGRES_MIN_CONNECTIONS=5
 POSTGRES_MAX_CONNECTIONS=20
 EMBEDDING_BATCH_SIZE=128
 EMBEDDING_REQUESTS_PER_MINUTE=300
-
-# HNSW index (advanced)
-HNSW_M=16
-HNSW_EF_CONSTRUCTION=64
 ```
 
 ### Cloud Sync Support (iCloud, Google Drive, Dropbox, OneDrive)
@@ -268,7 +264,7 @@ By default, the indexer excludes common system and tool folders:
 drafts/
 ```
 
-See [`.obsidian-graph.conf.example`](.obsidian-graph.conf.example) for more patterns and examples.
+See [`docs/obsidian-graph.conf.example`](docs/obsidian-graph.conf.example) for more patterns and examples.
 
 **Pattern Syntax:**
 

docker-compose.yml

@@ -24,8 +24,6 @@ services:
       - OBSIDIAN_WATCH_POLLING_INTERVAL=${OBSIDIAN_WATCH_POLLING_INTERVAL:-30}
       - POSTGRES_MIN_CONNECTIONS=${POSTGRES_MIN_CONNECTIONS:-5}
       - POSTGRES_MAX_CONNECTIONS=${POSTGRES_MAX_CONNECTIONS:-20}
-      - HNSW_M=${HNSW_M:-16}
-      - HNSW_EF_CONSTRUCTION=${HNSW_EF_CONSTRUCTION:-64}
       - EMBEDDING_BATCH_SIZE=${EMBEDDING_BATCH_SIZE:-128}
       - EMBEDDING_REQUESTS_PER_MINUTE=${EMBEDDING_REQUESTS_PER_MINUTE:-300}
       - MCP_SERVER_NAME=${MCP_SERVER_NAME:-obsidian-graph}

pyproject.toml

@@ -3,6 +3,24 @@ name = "obsidian-graph"
 version = "1.0.0"
 description = "Semantic knowledge graph engine for markdown vaults"
 requires-python = ">=3.11"
+dependencies = [
+    "mcp>=1.0.0",
+    "asyncpg>=0.29.0",
+    "pgvector>=0.3.0",
+    "voyageai>=0.3.0",
+    "watchdog>=4.0.0",
+    "loguru>=0.7.2",
+    "python-dotenv>=1.0.0",
+]
+
+[project.optional-dependencies]
+dev = [
+    "pytest>=8.0.0",
+    "pytest-asyncio>=0.23.0",
+    "black",
+    "ruff",
+    "mypy",
+]
 
 [tool.pytest.ini_options]
 testpaths = ["tests"]
@@ -19,7 +37,7 @@ markers = [
 source = ["src"]
 omit = [
     "tests/*",
-    "src/diagnose_vault.py",  # Diagnostic utility, not core functionality
+    "scripts/diagnose_vault.py",  # Diagnostic utility, not core functionality
 ]
 
 [tool.coverage.report]
@@ -60,7 +78,7 @@ ignore = [
 
 [tool.ruff.lint.per-file-ignores]
 "tests/*" = ["S101", "S106", "T201", "S108", "B007", "F841", "E402"]  # Allow assert, passwords, print, /tmp, unused vars, non-top imports in tests
-"src/diagnose_vault.py" = ["T201"]  # CLI diagnostic script uses print() for output
+"scripts/diagnose_vault.py" = ["T201"]  # CLI diagnostic script uses print() for output
 
 [tool.black]
 line-length = 100
@@ -76,3 +94,36 @@ extend-exclude = '''
   | __pycache__
 )/
 '''
+
+[tool.mypy]
+python_version = "3.11"
+# Gradual adoption - relaxed settings for existing code
+disallow_untyped_defs = false
+check_untyped_defs = false
+warn_return_any = false
+warn_unused_configs = true
+warn_redundant_casts = true
+warn_unused_ignores = false
+warn_no_return = true
+warn_unreachable = false
+strict_optional = false
+show_error_codes = true
+pretty = true
+# Disable error codes for gradual adoption (enable as codebase is typed)
+disable_error_code = ["var-annotated", "arg-type", "assignment", "call-overload", "return-value", "attr-defined"]
+
+[[tool.mypy.overrides]]
+module = "src.validation"
+disallow_untyped_defs = true
+disallow_any_unimported = true
+warn_return_any = true
+
+[[tool.mypy.overrides]]
+module = "src.security_utils"
+disallow_untyped_defs = true
+disallow_any_unimported = true
+warn_return_any = true
+
+[[tool.mypy.overrides]]
+module = ["watchdog.*", "voyageai.*", "pgvector.*", "mcp.*", "loguru.*"]
+ignore_missing_imports = true

src/diagnose_vault.py → scripts/diagnose_vault.py

@@ -5,7 +5,7 @@
 import sys
 from pathlib import Path
 
-sys.path.insert(0, "/app")
+sys.path.insert(0, str(Path(__file__).parent.parent))
 
 from src.embedder import VoyageEmbedder
 from src.exceptions import EmbeddingError

src/exceptions.py

@@ -20,7 +20,7 @@ class EmbeddingError(ObsidianGraphError):
     - cause: Original exception that triggered the failure
     """
 
-    def __init__(self, message: str, text_preview: str = "", cause: Exception = None):
+    def __init__(self, message: str, text_preview: str = "", cause: Exception | None = None):
         self.text_preview = text_preview[:100]  # First 100 chars for debugging
         self.cause = cause
         super().__init__(message)

src/indexer.py

@@ -157,7 +157,7 @@ async def index_vault(vault_path: str, batch_size: int = 100):
                 try:
                     embeddings_list, total_chunks = await embedder.embed_with_chunks(
                         note_data["content"],
-                        chunk_size=2000,  # oachatbot standard
+                        chunk_size=2000,
                         input_type="document",
                     )
                 except EmbeddingError as e:

src/vector_store.py

@@ -1,11 +1,9 @@
 """
-PostgreSQL+pgvector Vector Store for Obsidian Graph MCP Server.
+PostgreSQL+pgvector Vector Store for Obsidian Graph.
 
-Adapted from oachatbot's PostgreSQL store, simplified for Obsidian notes:
-- Stores whole notes and chunked documents (automatic chunking for large notes)
-- Uses 'path' as identifier (not document_id)
-- No site_id or publish_date (Obsidian-specific)
-- Adds connection_count materialization for graph queries
+Stores notes (whole or chunked) with vector embeddings for semantic search,
+graph analysis, and hub/orphan detection. Uses connection_count materialization
+for fast graph queries.
 """
 
 import asyncio

tests/conftest.py

@@ -1,33 +1,18 @@
 """
-Shared test fixtures for Obsidian Graph MCP Server.
+Shared test fixtures for Obsidian Graph.
 
 Provides reusable fixtures for:
-- Event loops (module-scoped for async tests)
 - Mock stores and embedders
 - Temporary vaults with test data
 - Server contexts for integration testing
 """
 
-import asyncio
-from datetime import datetime
+from datetime import UTC, datetime
 from unittest.mock import AsyncMock, MagicMock
 
 import pytest
 
 
-@pytest.fixture(scope="module")
-def event_loop():
-    """
-    Module-scoped event loop for async tests.
-
-    This ensures all async tests in a module share the same event loop,
-    which is more efficient and avoids loop recreation overhead.
-    """
-    loop = asyncio.get_event_loop_policy().new_event_loop()
-    yield loop
-    loop.close()
-
-
 @pytest.fixture
 def mock_store():
     """
@@ -203,7 +188,7 @@ def sample_notes():
             title="Test 1",
             content="This is test note 1",
             embedding=[0.1] * 1024,
-            modified_at=datetime.now(),
+            modified_at=datetime.now(UTC),
             file_size_bytes=100,
             chunk_index=0,
             total_chunks=1,
@@ -213,7 +198,7 @@ def sample_notes():
             title="Test 2",
             content="This is test note 2",
             embedding=[0.2] * 1024,
-            modified_at=datetime.now(),
+            modified_at=datetime.now(UTC),
             file_size_bytes=150,
             chunk_index=0,
             total_chunks=1,