Phase 5.1-5.4: PWA, Public Plugin Endpoints, Contacts Embedding, Search Coverage
Check Cross-Plugin Imports / check (push) Has been cancelled

5.1 Public Plugin Endpoints:
- PluginRouteDef.is_public field in manifest.py
- main.py: public routes mounted without auth dependency
- permissions/public_routes.py: token-based share link access (info, verify, download)
- permissions/plugin.py: public share route registered with is_public=True

5.2 PWA:
- vite.config.ts: VitePWA plugin configured (autoUpdate, workbox, runtime caching)
- frontend/public/manifest.json: PWA manifest with icons
- index.html: theme-color, manifest link, apple-touch-icon, apple-mobile-web-app meta
- Build generates sw.js + workbox (90 precache entries)

5.3 Contacts Embedding:
- contact.py: embedding column (Vector(768)) added to Contact model
- Migration 0002_embeddings.sql already exists (adds embedding + HNSW index)
- ContactSearchProvider already queries embedding column

5.4 Search Coverage:
- 5 new search providers: task, contactperson, tag, conversation, user
- All providers implement FTS search with tenant_id + deleted_at filters
- TagSearchProvider also supports vector search (384-dim embedding)
- provider_registry.py: all 5 new providers auto-registered
- Total: 10 search providers (was 5)
This commit is contained in:
Agent Zero
2026-08-04 14:49:35 +02:00
parent f704f7b032
commit cfb4c5ae8b
14 changed files with 994 additions and 8 deletions
@@ -115,6 +115,21 @@ async def auto_register_providers(db: AsyncSession) -> None:
from app.plugins.builtins.unified_search.providers.company_provider import (
CompanySearchProvider,
)
from app.plugins.builtins.unified_search.providers.task_provider import (
TaskSearchProvider,
)
from app.plugins.builtins.unified_search.providers.contactperson_provider import (
ContactPersonSearchProvider,
)
from app.plugins.builtins.unified_search.providers.tag_provider import (
TagSearchProvider,
)
from app.plugins.builtins.unified_search.providers.conversation_provider import (
ConversationSearchProvider,
)
from app.plugins.builtins.unified_search.providers.user_provider import (
UserSearchProvider,
)
registry = get_search_registry()
registry.clear()
@@ -126,6 +141,11 @@ async def auto_register_providers(db: AsyncSession) -> None:
FileSearchProvider,
EventSearchProvider,
CompanySearchProvider,
TaskSearchProvider,
ContactPersonSearchProvider,
TagSearchProvider,
ConversationSearchProvider,
UserSearchProvider,
]:
try:
registry.register(provider_cls())
@@ -0,0 +1,161 @@
"""ContactPerson search provider — queries contactpersons table."""
from __future__ import annotations
import logging
import uuid
from typing import Any
from sqlalchemy import text
from sqlalchemy.ext.asyncio import AsyncSession
from app.plugins.builtins.unified_search.base_provider import BaseSearchProvider
logger = logging.getLogger(__name__)
class ContactPersonSearchProvider(BaseSearchProvider):
"""Search provider for ContactPerson (Ansprechpartner) entities."""
entity_type = "contactperson"
async def _search_fts_filtered(
self,
db: AsyncSession,
tsquery: str,
tenant_id: uuid.UUID,
limit: int,
visible_ids: set[uuid.UUID] | None,
) -> list[dict[str, Any]]:
"""Full-text search on contactpersons fields."""
if visible_ids is not None:
sql = text(
"""
SELECT cp.*, ts_rank(
to_tsvector('pg_catalog.german',
coalesce(cp.displayname, '') || ' ' ||
coalesce(cp.firstname, '') || ' ' ||
coalesce(cp.lastname, '') || ' ' ||
coalesce(cp.email, '') || ' ' ||
coalesce(cp.function, '') || ' ' ||
coalesce(cp.tags, '')
),
to_tsquery('pg_catalog.german', :q)
) AS rank
FROM contactpersons cp
WHERE cp.tenant_id = :tid
AND cp.deleted_at IS NULL
AND to_tsvector('pg_catalog.german',
coalesce(cp.displayname, '') || ' ' ||
coalesce(cp.firstname, '') || ' ' ||
coalesce(cp.lastname, '') || ' ' ||
coalesce(cp.email, '') || ' ' ||
coalesce(cp.function, '') || ' ' ||
coalesce(cp.tags, '')
) @@ to_tsquery('pg_catalog.german', :q)
AND cp.id = ANY(:visible_ids)
ORDER BY rank DESC
LIMIT :lim
"""
)
result = await db.execute(
sql,
{
"q": tsquery,
"tid": tenant_id,
"lim": limit,
"visible_ids": list(visible_ids),
},
)
else:
sql = text(
"""
SELECT cp.*, ts_rank(
to_tsvector('pg_catalog.german',
coalesce(cp.displayname, '') || ' ' ||
coalesce(cp.firstname, '') || ' ' ||
coalesce(cp.lastname, '') || ' ' ||
coalesce(cp.email, '') || ' ' ||
coalesce(cp.function, '') || ' ' ||
coalesce(cp.tags, '')
),
to_tsquery('pg_catalog.german', :q)
) AS rank
FROM contactpersons cp
WHERE cp.tenant_id = :tid
AND cp.deleted_at IS NULL
AND to_tsvector('pg_catalog.german',
coalesce(cp.displayname, '') || ' ' ||
coalesce(cp.firstname, '') || ' ' ||
coalesce(cp.lastname, '') || ' ' ||
coalesce(cp.email, '') || ' ' ||
coalesce(cp.function, '') || ' ' ||
coalesce(cp.tags, '')
) @@ to_tsquery('pg_catalog.german', :q)
ORDER BY rank DESC
LIMIT :lim
"""
)
result = await db.execute(
sql,
{"q": tsquery, "tid": tenant_id, "lim": limit},
)
rows = result.mappings().all()
return [dict(r) for r in rows]
async def _search_vector_filtered(
self,
db: AsyncSession,
embedding: list[float],
tenant_id: uuid.UUID,
limit: int,
visible_ids: set[uuid.UUID] | None,
) -> list[dict[str, Any]]:
"""Semantic search — contactpersons table has no embedding column yet."""
return []
async def get_embedding_text(
self, db: AsyncSession, entity_id: uuid.UUID, tenant_id: uuid.UUID
) -> str:
"""Get text for embedding generation."""
sql = text(
"""
SELECT displayname, firstname, lastname, email, function, tags
FROM contactpersons
WHERE id = :eid AND tenant_id = :tid
"""
)
result = await db.execute(sql, {"eid": entity_id, "tid": tenant_id})
row = result.mappings().first()
if not row:
return ""
parts = [
row.get("displayname", ""),
row.get("firstname", ""),
row.get("lastname", ""),
row.get("email", ""),
row.get("function", ""),
row.get("tags", ""),
]
return " ".join(str(p) for p in parts if p)
def to_search_result(self, entity: object) -> dict[str, Any]:
"""Convert contactperson to search result dict."""
if isinstance(entity, dict):
displayname = entity.get("displayname", "")
email = entity.get("email", "") or ""
entity_id = str(entity.get("id", ""))
contact_id = str(entity.get("contact_id", "")) if entity.get("contact_id") else ""
else:
displayname = getattr(entity, "displayname", "")
email = getattr(entity, "email", "") or ""
entity_id = str(getattr(entity, "id", ""))
contact_id = str(getattr(entity, "contact_id", "")) if getattr(entity, "contact_id", None) else ""
return {
"entity_type": self.entity_type,
"entity_id": entity_id,
"title": displayname,
"snippet": email[:200],
"score": 0.0,
"data": {"contact_id": contact_id},
}
@@ -0,0 +1,137 @@
"""Conversation search provider — queries comm_conversations table."""
from __future__ import annotations
import logging
import uuid
from typing import Any
from sqlalchemy import text
from sqlalchemy.ext.asyncio import AsyncSession
from app.plugins.builtins.unified_search.base_provider import BaseSearchProvider
logger = logging.getLogger(__name__)
class ConversationSearchProvider(BaseSearchProvider):
"""Search provider for CommConversation entities."""
entity_type = "conversation"
async def _search_fts_filtered(
self,
db: AsyncSession,
tsquery: str,
tenant_id: uuid.UUID,
limit: int,
visible_ids: set[uuid.UUID] | None,
) -> list[dict[str, Any]]:
"""Full-text search on comm_conversations.title and last_msg_preview."""
if visible_ids is not None:
sql = text(
"""
SELECT c.*, ts_rank(
to_tsvector('pg_catalog.german',
coalesce(c.title, '') || ' ' || coalesce(c.last_msg_preview, '')
),
to_tsquery('pg_catalog.german', :q)
) AS rank
FROM comm_conversations c
WHERE c.tenant_id = :tid
AND c.deleted_at IS NULL
AND to_tsvector('pg_catalog.german',
coalesce(c.title, '') || ' ' || coalesce(c.last_msg_preview, '')
) @@ to_tsquery('pg_catalog.german', :q)
AND c.id = ANY(:visible_ids)
ORDER BY rank DESC
LIMIT :lim
"""
)
result = await db.execute(
sql,
{
"q": tsquery,
"tid": tenant_id,
"lim": limit,
"visible_ids": list(visible_ids),
},
)
else:
sql = text(
"""
SELECT c.*, ts_rank(
to_tsvector('pg_catalog.german',
coalesce(c.title, '') || ' ' || coalesce(c.last_msg_preview, '')
),
to_tsquery('pg_catalog.german', :q)
) AS rank
FROM comm_conversations c
WHERE c.tenant_id = :tid
AND c.deleted_at IS NULL
AND to_tsvector('pg_catalog.german',
coalesce(c.title, '') || ' ' || coalesce(c.last_msg_preview, '')
) @@ to_tsquery('pg_catalog.german', :q)
ORDER BY rank DESC
LIMIT :lim
"""
)
result = await db.execute(
sql,
{"q": tsquery, "tid": tenant_id, "lim": limit},
)
rows = result.mappings().all()
return [dict(r) for r in rows]
async def _search_vector_filtered(
self,
db: AsyncSession,
embedding: list[float],
tenant_id: uuid.UUID,
limit: int,
visible_ids: set[uuid.UUID] | None,
) -> list[dict[str, Any]]:
"""Semantic search — comm_conversations table has no embedding column yet."""
return []
async def get_embedding_text(
self, db: AsyncSession, entity_id: uuid.UUID, tenant_id: uuid.UUID
) -> str:
"""Get text for embedding generation."""
sql = text(
"""
SELECT title, last_msg_preview
FROM comm_conversations
WHERE id = :eid AND tenant_id = :tid
"""
)
result = await db.execute(sql, {"eid": entity_id, "tid": tenant_id})
row = result.mappings().first()
if not row:
return ""
parts = [
row.get("title", ""),
row.get("last_msg_preview", ""),
]
return " ".join(str(p) for p in parts if p)
def to_search_result(self, entity: object) -> dict[str, Any]:
"""Convert conversation to search result dict."""
if isinstance(entity, dict):
title = entity.get("title", "") or ""
preview = entity.get("last_msg_preview", "") or ""
entity_id = str(entity.get("id", ""))
is_direct = entity.get("is_direct", False)
else:
title = getattr(entity, "title", "") or ""
preview = getattr(entity, "last_msg_preview", "") or ""
entity_id = str(getattr(entity, "id", ""))
is_direct = getattr(entity, "is_direct", False)
return {
"entity_type": self.entity_type,
"entity_id": entity_id,
"title": title,
"snippet": preview[:200],
"score": 0.0,
"data": {"is_direct": is_direct},
}
@@ -0,0 +1,156 @@
"""Tag search provider — queries tags table."""
from __future__ import annotations
import logging
import uuid
from typing import Any
from sqlalchemy import text
from sqlalchemy.ext.asyncio import AsyncSession
from app.plugins.builtins.unified_search.base_provider import BaseSearchProvider
logger = logging.getLogger(__name__)
class TagSearchProvider(BaseSearchProvider):
"""Search provider for Tag entities."""
entity_type = "tag"
async def _search_fts_filtered(
self,
db: AsyncSession,
tsquery: str,
tenant_id: uuid.UUID,
limit: int,
visible_ids: set[uuid.UUID] | None,
) -> list[dict[str, Any]]:
"""Full-text search on tags.search_tsv."""
if visible_ids is not None:
sql = text(
"""
SELECT t.*, ts_rank(t.search_tsv, to_tsquery('pg_catalog.german', :q)) AS rank
FROM tags t
WHERE t.tenant_id = :tid
AND t.deleted_at IS NULL
AND t.search_tsv @@ to_tsquery('pg_catalog.german', :q)
AND t.id = ANY(:visible_ids)
ORDER BY rank DESC
LIMIT :lim
"""
)
result = await db.execute(
sql,
{
"q": tsquery,
"tid": tenant_id,
"lim": limit,
"visible_ids": list(visible_ids),
},
)
else:
sql = text(
"""
SELECT t.*, ts_rank(t.search_tsv, to_tsquery('pg_catalog.german', :q)) AS rank
FROM tags t
WHERE t.tenant_id = :tid
AND t.deleted_at IS NULL
AND t.search_tsv @@ to_tsquery('pg_catalog.german', :q)
ORDER BY rank DESC
LIMIT :lim
"""
)
result = await db.execute(
sql,
{"q": tsquery, "tid": tenant_id, "lim": limit},
)
rows = result.mappings().all()
return [dict(r) for r in rows]
async def _search_vector_filtered(
self,
db: AsyncSession,
embedding: list[float],
tenant_id: uuid.UUID,
limit: int,
visible_ids: set[uuid.UUID] | None,
) -> list[dict[str, Any]]:
"""Semantic search on tags.embedding (384-dim)."""
if visible_ids is not None:
sql = text(
"""
SELECT t.*, 1 - (t.embedding <=> cast(:emb AS vector)) AS score
FROM tags t
WHERE t.tenant_id = :tid
AND t.deleted_at IS NULL
AND t.embedding IS NOT NULL
AND t.id = ANY(:visible_ids)
ORDER BY t.embedding <=> cast(:emb AS vector)
LIMIT :lim
"""
)
result = await db.execute(
sql,
{
"emb": str(embedding),
"tid": tenant_id,
"lim": limit,
"visible_ids": list(visible_ids),
},
)
else:
sql = text(
"""
SELECT t.*, 1 - (t.embedding <=> cast(:emb AS vector)) AS score
FROM tags t
WHERE t.tenant_id = :tid
AND t.deleted_at IS NULL
AND t.embedding IS NOT NULL
ORDER BY t.embedding <=> cast(:emb AS vector)
LIMIT :lim
"""
)
result = await db.execute(
sql,
{"emb": str(embedding), "tid": tenant_id, "lim": limit},
)
rows = result.mappings().all()
return [dict(r) for r in rows]
async def get_embedding_text(
self, db: AsyncSession, entity_id: uuid.UUID, tenant_id: uuid.UUID
) -> str:
"""Get text for embedding generation."""
sql = text(
"""
SELECT name
FROM tags
WHERE id = :eid AND tenant_id = :tid
"""
)
result = await db.execute(sql, {"eid": entity_id, "tid": tenant_id})
row = result.mappings().first()
if not row:
return ""
return row.get("name", "") or ""
def to_search_result(self, entity: object) -> dict[str, Any]:
"""Convert tag to search result dict."""
if isinstance(entity, dict):
name = entity.get("name", "")
color = entity.get("color", "")
entity_id = str(entity.get("id", ""))
else:
name = getattr(entity, "name", "")
color = getattr(entity, "color", "")
entity_id = str(getattr(entity, "id", ""))
return {
"entity_type": self.entity_type,
"entity_id": entity_id,
"title": name,
"snippet": "",
"score": 0.0,
"data": {"color": color},
}
@@ -0,0 +1,129 @@
"""Task search provider — queries tasks table."""
from __future__ import annotations
import logging
import uuid
from typing import Any
from sqlalchemy import text
from sqlalchemy.ext.asyncio import AsyncSession
from app.plugins.builtins.unified_search.base_provider import BaseSearchProvider
logger = logging.getLogger(__name__)
class TaskSearchProvider(BaseSearchProvider):
"""Search provider for Task entities."""
entity_type = "task"
async def _search_fts_filtered(
self,
db: AsyncSession,
tsquery: str,
tenant_id: uuid.UUID,
limit: int,
visible_ids: set[uuid.UUID] | None,
) -> list[dict[str, Any]]:
"""Full-text search on tasks.title and tasks.description."""
if visible_ids is not None:
sql = text(
"""
SELECT t.*, ts_rank(
to_tsvector('pg_catalog.german', coalesce(t.title, '') || ' ' || coalesce(t.description, '')),
to_tsquery('pg_catalog.german', :q)
) AS rank
FROM tasks t
WHERE t.tenant_id = :tid
AND t.deleted_at IS NULL
AND to_tsvector('pg_catalog.german', coalesce(t.title, '') || ' ' || coalesce(t.description, '')) @@ to_tsquery('pg_catalog.german', :q)
AND t.id = ANY(:visible_ids)
ORDER BY rank DESC
LIMIT :lim
"""
)
result = await db.execute(
sql,
{
"q": tsquery,
"tid": tenant_id,
"lim": limit,
"visible_ids": list(visible_ids),
},
)
else:
sql = text(
"""
SELECT t.*, ts_rank(
to_tsvector('pg_catalog.german', coalesce(t.title, '') || ' ' || coalesce(t.description, '')),
to_tsquery('pg_catalog.german', :q)
) AS rank
FROM tasks t
WHERE t.tenant_id = :tid
AND t.deleted_at IS NULL
AND to_tsvector('pg_catalog.german', coalesce(t.title, '') || ' ' || coalesce(t.description, '')) @@ to_tsquery('pg_catalog.german', :q)
ORDER BY rank DESC
LIMIT :lim
"""
)
result = await db.execute(
sql,
{"q": tsquery, "tid": tenant_id, "lim": limit},
)
rows = result.mappings().all()
return [dict(r) for r in rows]
async def _search_vector_filtered(
self,
db: AsyncSession,
embedding: list[float],
tenant_id: uuid.UUID,
limit: int,
visible_ids: set[uuid.UUID] | None,
) -> list[dict[str, Any]]:
"""Semantic search — tasks table has no embedding column yet."""
return []
async def get_embedding_text(
self, db: AsyncSession, entity_id: uuid.UUID, tenant_id: uuid.UUID
) -> str:
"""Get text for embedding generation."""
sql = text(
"""
SELECT title, description
FROM tasks
WHERE id = :eid AND tenant_id = :tid
"""
)
result = await db.execute(sql, {"eid": entity_id, "tid": tenant_id})
row = result.mappings().first()
if not row:
return ""
parts = [
row.get("title", ""),
row.get("description", ""),
]
return " ".join(str(p) for p in parts if p)
def to_search_result(self, entity: object) -> dict[str, Any]:
"""Convert task to search result dict."""
if isinstance(entity, dict):
title = entity.get("title", "")
description = entity.get("description", "") or ""
entity_id = str(entity.get("id", ""))
status = entity.get("status", "")
else:
title = getattr(entity, "title", "")
description = getattr(entity, "description", "") or ""
entity_id = str(getattr(entity, "id", ""))
status = getattr(entity, "status", "")
return {
"entity_type": self.entity_type,
"entity_id": entity_id,
"title": title,
"snippet": description[:200],
"score": 0.0,
"data": {"status": status},
}
@@ -0,0 +1,162 @@
"""User search provider — queries users table via user_tenants for tenant scope."""
from __future__ import annotations
import logging
import uuid
from typing import Any
from sqlalchemy import text
from sqlalchemy.ext.asyncio import AsyncSession
from app.plugins.builtins.unified_search.base_provider import BaseSearchProvider
logger = logging.getLogger(__name__)
class UserSearchProvider(BaseSearchProvider):
"""Search provider for User entities.
Users are global entities (no tenant_id on users table).
Tenant membership is resolved via user_tenants join.
"""
entity_type = "user"
async def _search_fts_filtered(
self,
db: AsyncSession,
tsquery: str,
tenant_id: uuid.UUID,
limit: int,
visible_ids: set[uuid.UUID] | None,
) -> list[dict[str, Any]]:
"""Full-text search on users fields, filtered by tenant membership."""
if visible_ids is not None:
sql = text(
"""
SELECT u.*, ts_rank(
to_tsvector('pg_catalog.german',
coalesce(u.name, '') || ' ' ||
coalesce(u.first_name, '') || ' ' ||
coalesce(u.last_name, '') || ' ' ||
coalesce(u.email, '')
),
to_tsquery('pg_catalog.german', :q)
) AS rank
FROM users u
JOIN user_tenants ut ON ut.user_id = u.id
WHERE ut.tenant_id = :tid
AND u.deleted_at IS NULL
AND u.is_active = true
AND to_tsvector('pg_catalog.german',
coalesce(u.name, '') || ' ' ||
coalesce(u.first_name, '') || ' ' ||
coalesce(u.last_name, '') || ' ' ||
coalesce(u.email, '')
) @@ to_tsquery('pg_catalog.german', :q)
AND u.id = ANY(:visible_ids)
ORDER BY rank DESC
LIMIT :lim
"""
)
result = await db.execute(
sql,
{
"q": tsquery,
"tid": tenant_id,
"lim": limit,
"visible_ids": list(visible_ids),
},
)
else:
sql = text(
"""
SELECT u.*, ts_rank(
to_tsvector('pg_catalog.german',
coalesce(u.name, '') || ' ' ||
coalesce(u.first_name, '') || ' ' ||
coalesce(u.last_name, '') || ' ' ||
coalesce(u.email, '')
),
to_tsquery('pg_catalog.german', :q)
) AS rank
FROM users u
JOIN user_tenants ut ON ut.user_id = u.id
WHERE ut.tenant_id = :tid
AND u.deleted_at IS NULL
AND u.is_active = true
AND to_tsvector('pg_catalog.german',
coalesce(u.name, '') || ' ' ||
coalesce(u.first_name, '') || ' ' ||
coalesce(u.last_name, '') || ' ' ||
coalesce(u.email, '')
) @@ to_tsquery('pg_catalog.german', :q)
ORDER BY rank DESC
LIMIT :lim
"""
)
result = await db.execute(
sql,
{"q": tsquery, "tid": tenant_id, "lim": limit},
)
rows = result.mappings().all()
return [dict(r) for r in rows]
async def _search_vector_filtered(
self,
db: AsyncSession,
embedding: list[float],
tenant_id: uuid.UUID,
limit: int,
visible_ids: set[uuid.UUID] | None,
) -> list[dict[str, Any]]:
"""Semantic search — users table has no embedding column yet."""
return []
async def get_embedding_text(
self, db: AsyncSession, entity_id: uuid.UUID, tenant_id: uuid.UUID
) -> str:
"""Get text for embedding generation."""
sql = text(
"""
SELECT name, first_name, last_name, email
FROM users
WHERE id = :eid
"""
)
result = await db.execute(sql, {"eid": entity_id})
row = result.mappings().first()
if not row:
return ""
parts = [
row.get("name", ""),
row.get("first_name", ""),
row.get("last_name", ""),
row.get("email", ""),
]
return " ".join(str(p) for p in parts if p)
def to_search_result(self, entity: object) -> dict[str, Any]:
"""Convert user to search result dict."""
if isinstance(entity, dict):
name = entity.get("name", "")
email = entity.get("email", "") or ""
entity_id = str(entity.get("id", ""))
first_name = entity.get("first_name", "") or ""
last_name = entity.get("last_name", "") or ""
else:
name = getattr(entity, "name", "")
email = getattr(entity, "email", "") or ""
entity_id = str(getattr(entity, "id", ""))
first_name = getattr(entity, "first_name", "") or ""
last_name = getattr(entity, "last_name", "") or ""
display_name = f"{first_name} {last_name}".strip() or name
return {
"entity_type": self.entity_type,
"entity_id": entity_id,
"title": display_name,
"snippet": email[:200],
"score": 0.0,
"data": {"email": email},
}