- _ensure_keyword_payload_indexes: distinguish 400 (schema conflict, warning)
from other status codes (5xx/network, error) so a transient outage doesn't
silently leave the collection unindexed.
- build_search_result_from_point: use .get("doc_id") + return None on missing
instead of KeyError-crashing the search; reverse metadata merge order so
payload-derived chunk_index/total_chunks win over caller-supplied extras.
- docs/configuration.md: restore the OpenAI/Mistral/Bedrock/Simple provider
sections + reference-table rows that were dropped in the rebase. Reword
the "Startup migrations" bullet to describe what the code actually does
(no sampling — full scroll, zero writes when clean). Add operator note
about the SemanticSearchResult.id TypeError path.
- tests: pytest.approx for float equality (Sonar python:S1244); coverage
for non-400 → ERROR, payload={doc_id: None}, and missing doc_id key.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Production was logging two cascading classes of Qdrant errors against the
welcomed-malamute deployment:
1. HTTP 400 — "Bad request: Index required but not found for \"doc_id\" of
one of the following types: [keyword]". The collection was created via
create_collection() with no payload indexes, so any FieldCondition
filter on doc_id failed at the Qdrant layer (placeholder writes/reads,
eviction, search context lookups).
2. Compounding the missing index, producers wrote a mix of int and str
doc_ids: webhook_parser stringified node_id, scanner stringified note
IDs, news IDs, and deck card IDs — but the file scanner passed the
numeric file_id through unchanged. A keyword index would not have
covered both kinds even if it had existed.
This change:
- Normalizes doc_id to str at every producer site (scanner.py:459,
DocumentTask.doc_id, indexed_*_ids reads from Qdrant).
- Tightens str|int annotations to str across placeholder.py,
eviction.py, search/verification.py, search/context.py,
SearchResult.id, and the auth/api visualization endpoints.
- Defensive str() coercion on doc_id reads in semantic.py /
bm25_hybrid.py / vector/visualization.py for the transition window
before the backfill runs.
- Adds an idempotent startup migration in get_qdrant_client():
- _ensure_keyword_payload_indexes creates KEYWORD indexes for
doc_id, user_id, and doc_type (tolerates "already exists" 400s).
- _backfill_doc_id_to_string scrolls the collection once and rewrites
int doc_ids to str. Skipped after a quick sample shows no legacy
int payloads.
- Public API preserved: SemanticSearchResult.id stays int via explicit
int(r.id) narrowing in server/semantic.py — surfaces a TypeError with
actionable context if a future doc_type ships non-numeric ids.
- Documents the startup migration in docs/configuration.md.
Tests: 11 new unit tests in tests/unit/vector/test_qdrant_client.py
covering happy path / already-exists / unrelated-400 for the index
helpers, and sample-skip / mixed-batch rewrite / payload=None edge cases
for the backfill. 889 unit tests pass.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Hide sensitive files/folders from the WebDAV MCP tool surface by
tagging them with a configured Nextcloud system tag. Defence-in-depth
control for users who connect LLMs to accounts holding contracts,
medical records, credentials, etc.
A new EXCLUDED_TAGS env var (comma-separated tag names, empty by
default) gates an exclusion layer that runs at the start of every
WebDAV tool call: tag names are resolved to tag IDs, those IDs are
expanded to the set of tagged paths, then listings/searches are
filtered and read/write/delete/move/copy operations on excluded paths
raise ToolError. Tagged folders exclude their descendants via prefix
match. Empty EXCLUDED_TAGS disables the feature entirely.
The threat model is preventing accidental data exfiltration via the
LLM tool surface — not hiding files from a determined operator. The
docs explicitly recommend creating exclusion tags with
user_assignable=false so the credentials the MCP server uses cannot
remove the tag.
Implementation:
- config.py: add `excluded_tags` to _DEFAULTS, Settings, and the
_field_map alongside other comma-separated env vars.
- client/webdav.py: get_files_by_tag now requests <d:resourcetype/>
and surfaces is_directory so tagged directories can recursively
exclude descendants.
- server/tag_exclusion.py (new): get_excluded_tag_names,
get_excluded_file_paths, is_path_excluded.
- server/webdav.py: exclusion guards in all 11 WebDAV tools;
read/write/create/delete/move/copy raise ToolError, list/search
tools silently filter excluded entries. Existing f-string log
calls converted to lazy %-style.
- tests: 17 new unit tests covering path-matching edge cases
(shared-prefix non-match, descendants of excluded dirs), tag-name
parsing, and get_excluded_file_paths with mocked WebDAV; 1 new
client test asserting <d:resourcetype/> -> is_directory parsing.
- docs/configuration.md: new "Tag-Based File Exclusion" section with
per-tool effect table, security guidance, and per-call cost note.
- README.md: feature mention under Key Features.
Closes#710.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Rename `verified_count` → `verified_chunk_count` to make the count
granularity explicit at the field name (chunks vs unique docs).
- News verifier now fails open *per-item* on non-numeric stored doc_ids
(matches notes/files/deck shape); a single bad id no longer rescues
definitively-missing siblings from eviction.
- Update note-verifier integration test to use string doc_ids end-to-end
to match production storage (scanner.py:241 stringifies note ids).
- Add regression test for the closed-task-group race guard in
`verify_search_results` so the RuntimeError swallow is locked in.
- Convert remaining f-string logger calls in `server/semantic.py` to
lazy %-style formatting (per repo convention).
- Document `evict_on_missing` as a developer/test flag (no env var) and
flag the `get_file_info` 404→raise contract change in its docstring.
- Add a TODO(ADR-019) breadcrumb for the hardcoded 2× over-fetch so
future tuning has a clear hook.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Closes out the remaining nits flagged in the round-6 review.
Critical:
- _verify_files contract comment now enumerates all None-return cases
(404 + malformed PROPFIND XML) and documents the false-eviction
trade-off; self-healing via re-indexing recovers
- int(r.id) cast at the SemanticSearchResult boundary now raises a
TypeError with explicit doc_type/value context instead of bubbling
up as an opaque "Search failed: ..." McpError
Design observations:
- nc_semantic_search_answer docstring documents the per-note
round-trip cost from the post-verification race guard
- News verification latency hint added to configuration.md
- SemanticSearchResponse exposes verified_count + dropped_count so
short result pages on high-ghost-density indexes are
distinguishable from genuine scarcity. verify_search_results now
returns (kept, dropped_count); production caller and tests updated
Minor:
- Comment clarifies the .get() fallback in verify_search_results is
defensive only (run_verifier always populates the entry)
- Eviction task-group guard narrowed from except Exception to
except RuntimeError (the only documented failure mode of
TaskGroup.start_soon on a closed group)
- Indexer logs a warning when a deck_card task is missing
board_id/stack_id, surfacing data-quality issues at index time
rather than at verification time
- New unit test covers the news verifier's non-numeric-id fail-open
path (one bad doc_id keeps the entire batch)
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Tightens verifier consistency, closes test gaps, hardens the fire-and-forget
eviction snapshot, and routes the new concurrency knob through Settings.
- Pre-flight ``int()`` guard in ``_verify_notes`` mirrors ``_verify_deck_cards``,
so a non-numeric note id produces a type-specific log line instead of
falling through to the generic "unexpected error" branch.
- Adds explicit 403 tests for the file and news verifiers (symmetry with the
existing notes/deck 403 tests) plus a ``non_numeric_id_keeps`` test.
- ``AppContext`` and ``OAuthAppContext`` no longer snapshot
``_vector_sync_state.eviction_task_group`` at lifespan-yield time. Both
expose it as a ``@property`` that reads the singleton dynamically, removing
the order-sensitive race where a future startup-ordering change could
silently degrade fire-and-forget eviction to inline forever.
- Adds ``verification_concurrency`` (env var ``VERIFICATION_CONCURRENCY``,
default 20) to ``Settings`` with a dynaconf validator; ``verify_search_results``
resolves the cap lazily from settings when the caller doesn't override it.
- Enriches the news verifier TODO to call out that ``batch_size=-1`` is
intentional — a numeric ceiling would silently break correctness because
any item beyond the cap would be missing from ``present_ids`` and dropped.
- Updates ``Optional[TaskGroup]`` to ``TaskGroup | None`` per project style.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Implements fire-and-forget eviction (ADR-019 §"Lazy eviction"): the
search response no longer waits on Qdrant deletes, instead spawning
evict() on a long-lived lifespan-owned task group. Falls back to inline
eviction in modes without vector sync and in unit tests.
Also: harden _verify_news_items against non-numeric ids (fail open
instead of crashing the verifier); document the get_file_info None-on-404
contract; add INDEXED_DOC_TYPES single source of truth in vector/scanner.py
referenced by the CI-guard test; write a Verify-on-Read Latency Budget
section in docs/configuration.md covering the unbounded news.get_items
fetch. Closes the two remaining ADR-019 implementation checklist items.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The docs claimed scopes are mcp:-prefixed (mcp:notes.read,
mcp:notes.write) and that the notes.* pair "covers all Nextcloud
apps". Both are false. Per @require_scopes decorators across
nextcloud_mcp_server/server/, scopes are unprefixed and per-app:
notes.read/write, talk.read/write, files.read/write,
calendar.read/write, contacts.read/write, deck.read/write,
news.read, tables.read/write, cookbook.read/write,
todo.read/write, collectives.read/write, sharing.write,
semantic.read, plus standard OIDC scopes.
Changes:
- login-flow-v2.md: replace the false 2-row "covers all apps"
scope table with the real per-app reference (links to
scope_authorization.discover_all_scopes() as authoritative
source); strip mcp: prefix from intro paragraph, sequence
diagrams, @require_scopes example, WWW-Authenticate header
example. Also fix sticky-session keying advice per reviewer:
route on user identity (sub claim) rather than the raw bearer
token, since tokens rotate on refresh.
- auth-flows.md: clarify "Astrolabe (hosted UI) → MCP" matrix
column header; strip mcp: from sequence diagram and key
characteristics bullet; correct "issued by MCP server" to
"issued by configured IdP" on the Login Flow v2 token.
- authentication.md: strip mcp: from the high-level diagram and
scope-enforcement prose; cross-link to the scope reference.
- configuration.md: add NEXTCLOUD_OIDC_CLIENT_ID,
NEXTCLOUD_OIDC_CLIENT_SECRET, and OIDC_DISCOVERY_URL to the
Login Flow v2 vars table — these were undocumented in the
table after the round-2 multi-IdP fix.
- running.md: drop deprecated `version: '3.8'` from compose
snippets (Compose v2 ignores it and emits warnings).
- testing-oidc-consent.md: fix sample authorize URL and consent
description to use real scope names instead of mcp:-prefixed
ones (the manual test as written would have failed with
invalid_scope).
- CLAUDE.md: replace dead links to deleted oauth-architecture.md,
oauth-setup.md, and audience-validation-setup.md with
login-flow-v2.md.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Previous round narrowed the framing too far in the other direction —
made it sound like Nextcloud OIDC is *the* IdP. The MCP server
actually supports any OIDC-compliant provider (Nextcloud's built-in
OIDC, Keycloak, AWS Cognito, Auth0, etc.) selected via
`OIDC_DISCOVERY_URL`. `NEXTCLOUD_OIDC_CLIENT_ID/SECRET` are generic
OIDC client credentials despite the Nextcloud-flavored naming.
Code references:
- IdP discovery: app.py:607-668 (auto-detects integrated vs external
by comparing discovered issuer to NEXTCLOUD_HOST)
- JWKS: unified_verifier.py:71-73 (dynamically discovered, not
hard-coded to Nextcloud)
- IdP selection knob: OIDC_DISCOVERY_URL (config.py)
Changes:
- login-flow-v2.md: redraw "How It Works" diagram to show the IdP as
a separate component; replace "Nextcloud OIDC" with "configurable
IdP" framing throughout; add OIDC_DISCOVERY_URL to the env-var
reference; clarify NEXTCLOUD_OIDC_CLIENT_ID/SECRET are generic OIDC
creds; rename "OAuth Endpoints" subtitle to point at "the configured
IdP".
- running.md: rewrite the OAuth Mode intro and Quick Start note to
mention IdP configurability and OIDC_DISCOVERY_URL.
- configuration.md: update Best Practices "For Production" multi-user
bullet to reference the IdP selector and generic-creds caveat.
- auth-flows.md: generalize Astrolabe-flow and Login Flow v2
characteristics bullets — IdP and JWKS source are configurable.
- keycloak-multi-client-validation.md: REMOVE the "deprecated"
banner I added in 35c115e. The doc covers active behavior in
external-IdP mode (realm-level token validation by user_oidc),
not retired direct-OAuth-to-Nextcloud architecture. Replaced with
a scope note pointing at when this applies.
oauth-impersonation-findings.md keeps its deprecation banner — that
doc *is* about the rejected service-account / impersonation path
(ADR-002 Tier 2, "Will Not Implement"), so the deprecation framing
remains correct there.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The previous round of review feedback rested on a misunderstanding —
that the MCP server is "the OAuth issuer" under Login Flow v2 and that
NEXTCLOUD_OIDC_CLIENT_ID/SECRET are external-IdP-only. Code says
otherwise (app.py:619/625/703-717, unified_verifier.py:72):
- The MCP server is an OIDC relying party of Nextcloud OIDC. Tokens are
signed by Nextcloud and validated against Nextcloud's JWKS in all
modes — the server has no private signing keys.
- Static NEXTCLOUD_OIDC_CLIENT_ID/SECRET are the preferred way to
register the MCP server as that relying party; RFC 7591 DCR is a
fallback when both are unset.
- Login Flow v2 layers per-user app-password acquisition on top — it
governs the MCP→Nextcloud data leg, not the relying-party setup.
This commit reverts the inaccuracies introduced by 35c115e and reframes
the original `login-flow-v2.md` to match what the code does:
- login-flow-v2.md: revise "How It Works" to describe the MCP server
as an OIDC RP + OAuth facade (not a standalone issuer); rename
"OAuth Issuer Endpoints" → "OAuth Endpoints" with a note that those
endpoints front Nextcloud OIDC; add NEXTCLOUD_OIDC_CLIENT_ID/SECRET
to the required env vars with DCR documented as fallback.
- running.md: restore the static-creds Docker example (deleted in
35c115e on the wrong reasoning that it was tied to the retired
direct-OAuth-to-Nextcloud flow); rewrite the OAuth Mode section
intro to describe the actual relying-party + facade architecture.
- configuration.md: fix Best Practices "For Production" to mention
static creds as preferred / DCR as fallback; restore the .oauth
Docker volume alongside data so DCR-registered MCP-client state and
the encrypted app-password DB both persist.
- auth-flows.md: drop the note added in 35c115e that wrongly claimed
the MCP server validates Bearer tokens against its own JWKS under
Login Flow v2 — it validates against Nextcloud's JWKS in all modes;
reword the Login Flow v2 "Key characteristics" bullet that called
the MCP server "the OAuth authorization server".
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Round-2 cleanup of PR #743 review comments not covered by d153e96:
- configuration.md: fix broken `#multi-user-oauth-modes` anchor; replace
with the two real anchors (Multi-User BasicAuth, Login Flow v2). Rewrite
the stale "always use OAuth2/OIDC with pre-configured clients" Best
Practices section to reflect the post-pivot mode matrix, and update the
Docker volume example to mount the encrypted app-password store
(`TOKEN_STORAGE_DB`) rather than obsolete `.oauth` client storage.
- semantic-search-architecture.md: rename remaining body references from
the deprecated `VECTOR_SYNC_ENABLED` to `ENABLE_SEMANTIC_SEARCH` so the
doc matches configuration.md / troubleshooting.md.
- running.md: relabel "OAuth Mode (Recommended)" as
"Login Flow v2 / OAuth issuer mode (--oauth)", drop the misleading
"(Legacy)" suffix from BasicAuth, drop the
`NEXTCLOUD_OIDC_CLIENT_ID/SECRET` example (tied to the retired
direct-OAuth-to-Nextcloud flow), and add a note explaining what
`--oauth` actually enables post-pivot.
- keycloak-multi-client-validation.md, oauth-impersonation-findings.md:
add a deprecation banner pointing at ADR-022 / Login Flow v2. Files
retained because ADR-002 and CLAUDE.md still cite them.
- auth-flows.md: clarify under the Astrolabe → MCP diagram that the
Nextcloud-OIDC JWKS path applies to Multi-User BasicAuth; under
Login Flow v2 the MCP server validates tokens against its own JWKS.
- login-flow-v2.md: clarify the sticky-session note — affinity must key
on the OAuth bearer token (or user-bound cookie), not source IP, since
MCP clients may not maintain stable IPs across the provisioning flow.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Fix issues raised by reviewer on PR #743:
- troubleshooting.md: renumber "Getting Help" steps (4→3, 5→4) after
earlier consolidation left a gap
- installation.md: drop stale "OIDC app" prerequisite; admin access is
now optional under Login Flow v2 (works on stock Nextcloud 16+)
- semantic-search-architecture.md: rename VECTOR_SYNC_ENABLED to
ENABLE_SEMANTIC_SEARCH in the Status callout (renamed in v0.58.0)
- configuration.md: remove Quick Start references to deprecated
oauth-multi-user / oauth-advanced templates and point to
login-flow-v2.md; update "OAuth, Multi-User BasicAuth" label to
"Login Flow v2, Multi-User BasicAuth"
- auth-flows.md: fix background-sync diagram so Encrypt+persist step
no longer crosses into the Nextcloud column
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replace the seven OAuth-to-Nextcloud docs (oauth-setup, quickstart-oauth,
oauth-architecture, oauth-upstream-status, oauth-troubleshooting,
jwt-oauth-reference, audience-validation-setup) with a single new
docs/login-flow-v2.md. The deprecated flow required upstream user_oidc
patches that were never merged; Login Flow v2 is the forward-looking
multi-user mode (see ADR-022), and works with stock Nextcloud 16+.
Rewrite docs/authentication.md and docs/auth-flows.md around three modes:
Single-User BasicAuth, Multi-User BasicAuth pass-through, and Login Flow v2.
Update README to add an Astrolabe Cloud (https://astrolabecloud.com)
callout for users who prefer not to self-host, drop the OAuth deployment
mode from the auth table, simplify the Docker block, and trim the
Examples and Security sections.
Sweep configuration.md, installation.md, troubleshooting.md, running.md,
and semantic-search-architecture.md to replace links to the deleted docs
and update deprecated mode names.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Use dynaconf (get_settings()) instead of os.getenv for OIDC_RESOURCE_SERVER_ID
- Re-add Settings field, _field_map entry, and settings.toml default
- Add trailing-slash guard (.rstrip("/")) to prevent double-slash in scopes
- Add double-prefixing guard: skip scopes already carrying the prefix
- Add @pytest.mark.unit to test module
- Add test for already-prefixed scopes
- Document OIDC_RESOURCE_SERVER_ID in docs/configuration.md
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Smithery is no longer a supported deployment mode. Remove all Smithery-specific
code paths, middleware, configuration, and tests. This simplifies the codebase
by eliminating DeploymentMode enum, SmitheryConfigMiddleware, session config
context variables, and the smithery_main entrypoint.
Files deleted: Dockerfile.smithery, smithery.yaml, smithery_main.py
ADR-016 retained with deprecated status for historical reference.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add NEXTCLOUD_VERIFY_SSL and NEXTCLOUD_CA_BUNDLE env vars to configure
TLS certificate verification for all outbound Nextcloud connections.
Centralizes SSL config via a new HTTP client factory (http.py) used by
all 27 Nextcloud-bound call sites, including API clients, OIDC endpoints,
OAuth flows, and health checks.
Closes#560
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Remove astrolabe-specific docs and sections that belong in the
astrolabe repo. Update remaining references to point to the
astrolabe repo where appropriate.
- Fix .gitmodules SSH → HTTPS URL for astrolabe submodule
- Remove bump-version.yml stale "astrolabe" scope comment
- Delete blog-introducing-astrolabe.md (moved to astrolabe repo)
- Remove "Astrolabe Background Token Refresh" section from auth-flows.md
- Replace "Astrolabe User Setup" section in authentication.md with link
- Remove "Astrolabe Internal URL" section from configuration.md
- Remove "Webhook Presets (via Astrolabe UI)" from webhook guide
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This PR enables safe switching between embedding models and multi-server
deployments by implementing auto-generated Qdrant collection names based on
deployment ID and model name.
## Problem
Previously, all deployments used a single hardcoded collection name
"nextcloud_content", which caused two critical issues:
1. **Dimension mismatches when switching models**: Changing
OLLAMA_EMBEDDING_MODEL (e.g., nomic-embed-text at 768D → all-minilm at
384D) would cause runtime errors as vectors couldn't be inserted into a
collection with incompatible dimensions.
2. **Collection collisions in multi-server setups**: Multiple MCP servers
sharing a single Qdrant instance would overwrite each other's data,
making horizontal scaling impossible.
## Solution
### Auto-Generated Collection Naming
Collections are now automatically named using the pattern:
\`{deployment-id}-{model-name}\`
**Deployment ID**: Uses \`OTEL_SERVICE_NAME\` if configured (and not default
value), otherwise falls back to \`hostname\` for simple Docker deployments.
**Model Name**: From \`OLLAMA_EMBEDDING_MODEL\` with path separators sanitized.
**Examples**:
- \`my-mcp-server-nomic-embed-text\` (with OTEL_SERVICE_NAME=my-mcp-server)
- \`mcp-container-all-minilm\` (simple Docker, hostname=mcp-container)
**Override**: Users can still set \`QDRANT_COLLECTION\` explicitly to bypass
auto-generation for backward compatibility.
### Dimension Validation
Added startup validation that checks collection dimensions match the
embedding service. If a mismatch is detected, the server fails fast with a
clear error message explaining:
- Expected vs actual dimensions
- Likely cause (model change)
- Solutions (delete collection, use different name, or revert model)
### Improved Sampling Error Handling
Enhanced MCP sampling rejection handling to treat user rejections as normal
behavior rather than errors:
- **User rejections** ("rejected", "denied") → INFO log, no traceback
- **Unsupported clients** → INFO log, no traceback
- **Other MCP errors** → WARNING log, no traceback
- **Unexpected errors** → ERROR log WITH traceback
This aligns with the MCP specification where clients SHOULD prompt users for
approval/denial of sampling requests.
## Changes
### Core Implementation
- **nextcloud_mcp_server/config.py**: Added \`get_collection_name()\` method
with deployment ID detection and model name sanitization
- **nextcloud_mcp_server/vector/qdrant_client.py**: Dimension validation on
collection open with helpful error messages
- **nextcloud_mcp_server/vector/{scanner,processor}.py**: Updated to use
\`get_collection_name()\`
- **nextcloud_mcp_server/auth/userinfo_routes.py**: Vector sync status uses
\`get_collection_name()\`
- **nextcloud_mcp_server/server/semantic.py**:
- Updated semantic search tools to use \`get_collection_name()\`
- Improved sampling rejection error handling (McpError vs Exception)
### Documentation
- **docs/semantic-search-architecture.md**: New comprehensive architecture
document (557 lines) covering background sync, semantic search flow, RAG
implementation, and deployment modes
- **docs/configuration.md**: Added detailed "Qdrant Collection Naming"
section with examples and multi-server deployment guidance
- **docker-compose.yml**: Added comments explaining collection naming behavior
- **README.md**: Updated semantic search descriptions to clarify
experimental status, Notes-only support, and infrastructure requirements
## Migration Guide
**For existing single-server deployments:**
Option 1 (Recommended): Use explicit collection name for continuity
\`\`\`bash
QDRANT_COLLECTION=nextcloud_content # Keep existing collection
\`\`\`
Option 2: Allow auto-generation and re-embed
\`\`\`bash
# Remove QDRANT_COLLECTION override
# New collection will be created based on deployment ID + model
# Requires re-embedding all documents (may take time)
\`\`\`
**For new multi-server deployments:**
Set unique OTEL service names per server:
\`\`\`bash
# Server 1
OTEL_SERVICE_NAME=mcp-prod
OLLAMA_EMBEDDING_MODEL=nomic-embed-text
# → Collection: "mcp-prod-nomic-embed-text"
# Server 2
OTEL_SERVICE_NAME=mcp-staging
OLLAMA_EMBEDDING_MODEL=nomic-embed-text
# → Collection: "mcp-staging-nomic-embed-text"
\`\`\`
## Benefits
✅ **Safe model switching**: Each model gets its own collection, preventing
dimension mismatch errors
✅ **Multi-server support**: Multiple MCP servers can share one Qdrant
instance without conflicts
✅ **Clear ownership**: Collection names show which deployment and model owns
the data
✅ **Better error messages**: Dimension validation provides actionable
guidance
✅ **Backward compatible**: Existing deployments can continue using
\`QDRANT_COLLECTION\` override
## Testing
Validated with:
- Single-server deployments (default hostname-based naming)
- Multi-server deployments (OTEL service name-based naming)
- Model switching scenarios (dimension validation)
- Collection override scenarios (backward compatibility)
Next steps: Testing various Ollama embedding models to investigate optimal
chunk sizes and performance characteristics.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Adds flexible Qdrant deployment modes to reduce infrastructure requirements
for local development and smaller deployments:
**Configuration Changes:**
- Add QDRANT_LOCATION environment variable (mutually exclusive with QDRANT_URL)
- Three modes: network (URL), in-memory (:memory:, default), persistent (file path)
- Settings dataclass validation via __post_init__ ensures mutual exclusivity
- API key warning when set in local mode (ignored, only for network mode)
**Client Initialization:**
- Auto-detect mode: network (url + api_key) vs local (:memory: or path=)
- In-memory: AsyncQdrantClient(":memory:") - zero config default
- Persistent: AsyncQdrantClient(path="/app/data/qdrant") - file storage
- Network: AsyncQdrantClient(url, api_key) - production mode
**Docker Compose Updates:**
- Qdrant service moved to optional profile (--profile qdrant)
- MCP service uses QDRANT_LOCATION=:memory: by default
- Added mcp-data volume for persistent storage (/app/data)
- No hard dependency on qdrant service
**Documentation:**
- Comprehensive configuration guide in docs/configuration.md
- All three modes documented with pros/cons
- Docker Compose examples for each mode
- Environment variable reference table
**Tests:**
- 13 new config validation tests (mutual exclusivity, defaults, warnings)
- Persistent mode integration test (create, close, reopen, verify persistence)
- All 82 unit tests + 5 smoke tests pass
**Breaking Change:**
- Default changed from QDRANT_URL=http://qdrant:6333 to QDRANT_LOCATION=:memory:
- Simplifies local development (no external service needed)
- Production deployments: explicitly set QDRANT_URL or QDRANT_LOCATION
Related: ADR-007 background vector sync implementation
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
