This commit addresses critical performance issues with vector visualization
search (reducing time from 40s to ~2s) and improves result visualization
through better visual encoding.
## Performance Fixes
### 1. Fix blocking sleep in retry decorator (base.py:51)
- Changed `time.sleep(5)` to `await anyio.sleep(5)` in @retry_on_429
- Prevents entire event loop from freezing during rate limit retries
- Impact: Reduced search time from 22s to 16s initially
### 2. Add concurrency limiting for verification (verification.py:77-93)
- Added `anyio.Semaphore(20)` to limit concurrent HTTP requests
- Prevents connection pool exhaustion (RequestError) from 90+ simultaneous requests
- Fixes false filtering (was filtering 77/90 results incorrectly)
- Note: Semaphore still in code but verification removed from viz endpoint
### 3. Remove unnecessary verification from viz endpoint (viz_routes.py:483-486)
- Visualization only needs Qdrant metadata (title, excerpt), not full content
- Verification only required for sampling (LLM needs full note content)
- Impact: Reduced search time from 43.7s to ~2s (final fix)
### 4. Restore streaming scanner pattern (scanner.py)
- Process notes one-at-a-time using async generator
- Avoids loading all notes into memory
## Visualization Improvements
### 5. Result-relative score normalization (viz_routes.py:489-504)
- Normalize scores within result set: best=1.0, worst=0.0
- Removes arbitrary RRF normalization (theoretical max didn't make sense)
- Makes visual encoding meaningful regardless of algorithm scores
### 6. Power scaling for marker sizes (userinfo_routes.py:743)
- Changed from linear `8 + (score * 12)` to power `6 + (score² * 14)`
- Creates dramatic visual contrast: 0.0→6px, 0.5→9.5px, 1.0→20px
- Combined with opacity (0.2-1.0) for clear visual hierarchy
### 7. Multi-channel visual encoding (userinfo_routes.py:740-745)
- Size: Exponentially scaled with score²
- Opacity: Linear 0.2-1.0 (keeps all points visible)
- Color: Viridis gradient (blue→yellow)
- Effect: Top results are large/bright/opaque, context results small/dim/transparent
## Result
- Search time: 40s → ~2s (20x faster)
- Visual contrast: Subtle → dramatic (clear result hierarchy)
- No arbitrary cutoffs: All results visible, best naturally highlighted
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Replace asyncio primitives with anyio equivalents throughout the codebase
to establish a single async pattern. This provides better structured
concurrency with automatic cancellation on errors and aligns with the
pytest anyio configuration.
Changes:
- hybrid.py: Replace asyncio.gather() with anyio task groups
- token_broker.py: Replace asyncio.Lock() with anyio.Lock()
- storage.py: Replace asyncio.run() with anyio.run()
- app.py: Replace tg.start_soon() with await tg.start() for task status
- processor.py: Add task_status parameter for structured startup
- scanner.py: Add task_status parameter for structured startup
- CLAUDE.md: Update async/await patterns guidance
The change from start_soon() to await tg.start() enables proper task
initialization signaling, ensuring background tasks are ready before
proceeding. This follows anyio best practices for structured concurrency.
All 118 unit tests pass with the new implementation.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Collect all notes to delete first, then delete concurrently
- Use anyio task group with semaphore (20 concurrent deletions)
- Add progress reporting and error tracking for deletions
- Show count of notes found before deletion starts
This significantly improves --force performance when refreshing large
corpuses (e.g., 3,633 notes now delete in ~1 minute instead of ~5 minutes).
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Add --force flag to delete all existing notes in target category before upload
- Implement concurrent uploads using anyio task groups (20 concurrent max)
- Add semaphore to limit concurrent requests and avoid overwhelming server
- Improve progress reporting with upload count and error tracking
- Update README with --force flag documentation
Performance improvement: Concurrent uploads significantly reduce upload time
from ~10-15 minutes to ~2-3 minutes for 3,633 documents.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- HuggingFace BeIR/nfcorpus only has 'corpus' and 'queries' configs
- Download qrels from original BEIR ZIP file (nfcorpus.zip)
- Use synchronous httpx.Client for download (simpler than async)
- Remove deprecated trust_remote_code parameter
Tested with successful corpus download and qrels extraction.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Use NextcloudClient with BasicAuth instead of raw httpx
- Replace direct HTTP POST with notes.create_note() method
- Add close() method to LLMProvider Protocol for proper cleanup
- Fix type annotations for dataset iteration
This improves code reuse and consistency with the rest of the codebase.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Add ADR-013 documenting RAG evaluation architecture
- Implement two-part evaluation: Context Recall (retrieval) + Answer Correctness (generation)
- Create Click CLI for ground truth generation and corpus upload
- Add pytest fixtures and tests for retrieval/generation quality
- Use BeIR/nfcorpus dataset with 5 selected test queries
- Support Ollama and Anthropic LLM providers
- Generate synthetic ground truth answers offline
- Add comprehensive documentation in tests/rag_evaluation/README.md
The framework separates one-time setup (generate/upload) from test execution,
making tests much faster (~6-12 min vs ~15-25 min per run).
Tests are manual only (not in CI) and require external LLM access.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Improve user comprehension by scaling RRF scores to match the intuitive
0-1 range used by other search algorithms.
## Problem
RRF (Reciprocal Rank Fusion) scores had a drastically different scale
than semantic/keyword/fuzzy scores:
- Semantic similarity: 0.0 to 1.0 (typical: 0.5-0.9)
- RRF scores: 0.0 to ~0.016 (typical: 0.005-0.015)
This caused user confusion - a score of 0.0078 looked terrible but was
actually excellent (near theoretical maximum).
## Solution
Normalize RRF scores using the formula:
`normalized_score = rrf_score * (rrf_k + 1) / total_weight`
Where:
- rrf_k = 60 (RRF constant)
- total_weight = sum of algorithm weights (default: 1.0)
**Example transformation:**
- Before: 0.0078 (confusing)
- After: 0.477 (intuitive)
## Changes
**nextcloud_mcp_server/search/hybrid.py:**
- Store total_weight as instance variable (line 63)
- Calculate normalization factor in _reciprocal_rank_fusion() (line 209)
- Apply normalization to all RRF scores (line 217)
- Preserve raw RRF score in metadata for debugging (line 222)
## Impact
**User Experience:**
- Hybrid search scores now comparable with semantic/keyword/fuzzy
- Score of 0.5 indicates good match across all algorithms
- Consistent scale improves score threshold usability
**Backward Compatibility:**
- Raw RRF scores preserved in metadata["rrf_score_raw"]
- Result ordering unchanged (normalization is linear transformation)
- Breaking change: Existing score thresholds need adjustment
**Performance:**
- Negligible overhead (single multiplication per result)
## Testing
Verified with nc_semantic_search and nc_semantic_search_answer:
- Hybrid scores now 0.47-0.7 range (was 0.003-0.011)
- Semantic scores unchanged (0.75)
- Result ordering preserved
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Move access verification from individual search algorithms to final output
stage, eliminating redundant API calls and improving performance.
## Changes
**New:**
- `search/verification.py`: Centralized verification using anyio task groups
- Deduplicates results by (doc_id, doc_type) before verification
- Verifies all unique documents in parallel using structured concurrency
- Filters out inaccessible documents in single pass
**Modified Search Algorithms:**
- `search/semantic.py`: Removed _deduplicate_and_verify() and _verify_document_access()
- `search/keyword.py`: Removed _verify_access() and parallel verification
- `search/fuzzy.py`: Removed _verify_access() and parallel verification
- `search/hybrid.py`: Removed nextcloud_client parameter passing
All algorithms now return unverified results from Qdrant payload.
**Modified Output Stages:**
- `server/semantic.py`: Added verify_search_results() call after search
- `auth/viz_routes.py`: Added verify_search_results() call after search
Both endpoints now verify access once at final stage with deduplication.
## Performance Impact
**Before:**
- Hybrid mode (limit=10): 30 API calls (10 per algorithm × 3 algorithms)
- Single algorithm: 10-20 API calls (with verification buffer)
**After:**
- Hybrid mode (limit=10): 10 API calls (deduplicated verification)
- Single algorithm: 10 API calls (deduplicated verification)
**Performance Gain:** 3x reduction in API calls for hybrid search
## Architecture Benefits
- **Separation of concerns**: Algorithms handle scoring, output stage handles security
- **Deduplication**: Each document verified exactly once
- **Parallel execution**: All verifications run concurrently via anyio task groups
- **Consistency**: Same verification logic across MCP tools and viz endpoints
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Vector Visualization Improvements:
- Add interactive vector viz tab with Alpine.js and Plotly.js to user info page
- Refactor viz route CSS for better scoping and maintainability
- Remove unused nextcloud_host variable
Performance Optimizations:
- Parallelize access verification in fuzzy and keyword search algorithms
- Use asyncio.gather() to verify multiple documents concurrently
- Add exception handling with return_exceptions=True for resilience
Dependencies:
- Update third_party/oidc submodule to include RFC 9728 resource_url support
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Skip tracing for /app/vector-sync/status to reduce noise from HTMX polling.
Metrics collection continues for this endpoint.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Move Webhooks tab to the right (User Info | Vector Sync | Vector Viz | Webhooks)
- Use request.user.display_name instead of session for viz routes
- Fixes session middleware error when accessing via iframe
- Add /app/vector-viz endpoint for interactive search testing
- Implement server-side PCA dimensionality reduction (768-dim → 2D)
- Support multi-select document type filter for cross-app search
- Support all search algorithms: semantic, keyword, fuzzy, hybrid
- Display 2D scatter plot of vector embeddings using Plotly
- Show search results with scores and document types
- Register viz routes in app.py
- Add custom PCA implementation using numpy eigendecomposition
- Replace sklearn.decomposition.PCA with custom implementation
- Maintains same API (fit, transform, fit_transform)
- Supports explained_variance_ratio_ for variance analysis
- Removes scikit-learn dependency from project
- Add type hints and assertion for type safety
BREAKING CHANGE: Search algorithms now require Qdrant to be populated.
Vector sync must be enabled and documents indexed for search to work.
- Keyword and fuzzy search now query Qdrant scroll API for title/excerpt
- Remove inefficient Nextcloud API fetching pattern
- Add optional Nextcloud verification for security
- Deduplicate by (doc_id, doc_type) tuple, keeping chunk_index=0
- Align with document processor pattern that already stores text in Qdrant
Implements NextcloudClientProtocol for multi-document type search following
user requirement that document types are not 1:1 with apps (e.g., Notes app
specializes in markdown, while Files/WebDAV handles multiple file types).
Key Changes:
- NextcloudClientProtocol: Generic protocol with app-specific client properties
- get_indexed_doc_types(): Query Qdrant for actually-indexed document types
- Document dispatch: All algorithms check Qdrant before attempting access
- Cross-type deduplication: Use (doc_id, doc_type) tuples in hybrid RRF
Search Algorithm Updates:
- Semantic: Added _verify_document_access() with dispatch to appropriate client
- Deduplication by (doc_id, doc_type) tuple
- Only "note" verification implemented, others return None with info log
- Keyword: Added _fetch_documents() dispatch method
- Queries Qdrant for available types before fetching
- Supports cross-app search when doc_type=None
- Fuzzy: Same pattern as keyword search
- Hybrid: Already uses (doc_id, doc_type) for deduplication (no changes needed)
Future-Proof Design:
- File/calendar verification stubs in place
- Clear logging when unsupported types found
- Easy to extend when processor indexes new document types
Currently Supported:
- "note" documents fully implemented and tested
- Other types gracefully handled (logged but skipped)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Implements ADR-012 by adding multi-algorithm support to the MCP tool.
Key changes:
- Added algorithm parameter: "semantic"|"keyword"|"fuzzy"|"hybrid" (default: "hybrid")
- Added weight parameters for hybrid mode configuration
- Replaced direct Qdrant/embedding calls with search module abstractions
- Updated docstring to describe all four algorithms
- Simplified implementation: ~50 lines vs ~150 lines (67% reduction)
- Better error handling for missing vector sync
Algorithm selection:
- semantic: Pure vector similarity (requires VECTOR_SYNC_ENABLED=true)
- keyword: Token-based matching with weighted title/content scoring
- fuzzy: Character overlap for typo tolerance
- hybrid: RRF fusion with configurable weights (default: 0.5/0.3/0.2)
Backward compatibility:
- Tool name unchanged (nc_semantic_search)
- New parameters have sensible defaults
- Existing clients get hybrid search automatically (better than pure semantic)
- search_method field in response reflects actual algorithm used
Weight validation:
- Performed in HybridSearchAlgorithm constructor
- Must sum to ≤1.0 and all non-negative
- At least one weight must be > 0
- Clear error messages on validation failure
Next: Update viz pane to use same algorithms
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Updates ADR-012 to clarify that all search and filtering operations
must happen server-side, not in the browser.
Key changes:
- Enhanced viz pane data flow showing server-side processing
- Added performance benefits section (384x bandwidth reduction)
- Detailed server-side filtering approach:
* Query execution via search/algorithms.py
* User ID filtering (multi-tenant security)
* Document type filtering
* PCA reduction (768-dim → 2D) on server
* Only 2D coordinates + metadata sent to client
- Updated Phase 3 implementation plan:
* Remove ALL client-side search logic
* Implement /app/vector-viz server endpoint
* htmx form submission for queries
* Performance optimizations (caching, streaming)
This ensures:
- Minimal bandwidth usage (only 2 floats per doc vs 768)
- Client handles only visualization, not computation
- Can visualize 10,000+ documents without client lag
- Raw vectors never leave server (security)
- Same search logic as MCP tool (consistency)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Enhances ADR-012 with detailed architecture visualization and UI mockup
for the vector visualization pane.
Added sections:
- Architecture diagram showing MCP tool and viz pane integration
- Data flow diagrams for both MCP requests and viz pane interactions
- Detailed UI mockup with ASCII art showing:
* Search configuration controls
* Algorithm selector with weight sliders
* Interactive 2D scatter plot (Plotly.js)
* Results panel with scores
* Performance comparison table
- Technology stack details (htmx, Alpine.js, Plotly.js, Tailwind CSS)
The diagrams illustrate how the viz pane and MCP tool share the same
search algorithm implementations from search/algorithms.py, ensuring
consistency between user testing interface and programmatic API.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Changes:
- Remove streamable-http transport override from mcp service in docker-compose.yml
- Service now uses CLI default SSE transport on /sse endpoint
- Add create_mcp_client_session_sse() helper for SSE connections
- Update nc_mcp_client fixture to use SSE transport
- Fix unpacking for SSE client (yields 2 values vs 3 for streamable-http)
Testing:
- All 4 smoke tests pass with SSE transport
- 32/34 affected tests pass (2 skipped for vector sync)
- OAuth services remain on streamable-http (unchanged)
Note: SSE transport is being deprecated in favor of streamable-http.
This enables minimal validation testing before deprecation.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
After comprehensive research, the hybrid OAuth + AppAPI architecture is NOT
being implemented due to fundamental architectural incompatibilities.
Key updates:
- Status: Proposed → Not Planned
- Added validation from Nextcloud Context Agent project
- Context Agent (official NC ExApp with MCP) faces IDENTICAL limitations
- Proves constraints are architectural, not implementation-specific
Context Agent findings:
- ExApp with MCP server endpoint (~28 tools exposed)
- Uses Task Processing API for confirmations (NOT MCP elicitation)
- Works around AppAPI proxy limitations by changing protocol
- MCP endpoint is secondary feature with documented constraints
- Primary use: In-app Assistant integration, not external MCP clients
Critical features impossible through AppAPI proxy:
- ❌ MCP sampling (eliminates RAG/LLM features)
- ❌ MCP elicitation (user prompts)
- ❌ Real-time progress updates
- ❌ Bidirectional streaming
- Validated by Context Agent facing same limitations
Decision rationale:
- MCP requires multi-turn nested interactions
- AppAPI provides stateless request/response proxy only
- No implementation effort can bridge this fundamental gap
- Would require complete AppAPI redesign (WebSocket, message routing)
- Even official Nextcloud projects work around these limitations
Alternative considered for future:
- Register as Task Processing provider (different product)
- Use Nextcloud Assistant UI (not external MCP clients)
- Accept different capabilities (no sampling, custom flows)
OAuth mode remains sole solution for external MCP client integration.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Chris Coutinho