feat: workspace-based multi-tenancy, replacing user as tenancy axis (#840)

Introduces `workspace` as the isolation boundary for config, flows,
library, and knowledge data. Removes `user` as a schema-level field
throughout the code, API specs, and tests; workspace provides the
same separation more cleanly at the trusted flow.workspace layer
rather than through client-supplied message fields.

Design
------
- IAM tech spec (docs/tech-specs/iam.md) documents current state,
  proposed auth/access model, and migration direction.
- Data ownership model (docs/tech-specs/data-ownership-model.md)
  captures the workspace/collection/flow hierarchy.

Schema + messaging
------------------
- Drop `user` field from AgentRequest/Step, GraphRagQuery,
  DocumentRagQuery, Triples/Graph/Document/Row EmbeddingsRequest,
  Sparql/Rows/Structured QueryRequest, ToolServiceRequest.
- Keep collection/workspace routing via flow.workspace at the
  service layer.
- Translators updated to not serialise/deserialise user.

API specs
---------
- OpenAPI schemas and path examples cleaned of user fields.
- Websocket async-api messages updated.
- Removed the unused parameters/User.yaml.

Services + base
---------------
- Librarian, collection manager, knowledge, config: all operations
  scoped by workspace. Config client API takes workspace as first
  positional arg.
- `flow.workspace` set at flow start time by the infrastructure;
  no longer pass-through from clients.
- Tool service drops user-personalisation passthrough.

CLI + SDK
---------
- tg-init-workspace and workspace-aware import/export.
- All tg-* commands drop user args; accept --workspace.
- Python API/SDK (flow, socket_client, async_*, explainability,
  library) drop user kwargs from every method signature.

MCP server
----------
- All tool endpoints drop user parameters; socket_manager no longer
  keyed per user.

Flow service
------------
- Closure-based topic cleanup on flow stop: only delete topics
  whose blueprint template was parameterised AND no remaining
  live flow (across all workspaces) still resolves to that topic.
  Three scopes fall out naturally from template analysis:
    * {id} -> per-flow, deleted on stop
    * {blueprint} -> per-blueprint, kept while any flow of the
      same blueprint exists
    * {workspace} -> per-workspace, kept while any flow in the
      workspace exists
    * literal -> global, never deleted (e.g. tg.request.librarian)
  Fixes a bug where stopping a flow silently destroyed the global
  librarian exchange, wedging all library operations until manual
  restart.

RabbitMQ backend
----------------
- heartbeat=60, blocked_connection_timeout=300. Catches silently
  dead connections (broker restart, orphaned channels, network
  partitions) within ~2 heartbeat windows, so the consumer
  reconnects and re-binds its queue rather than sitting forever
  on a zombie connection.

Tests
-----
- Full test refresh: unit, integration, contract, provenance.
- Dropped user-field assertions and constructor kwargs across
  ~100 test files.
- Renamed user-collection isolation tests to workspace-collection.
This commit is contained in:
cybermaggedon 2026-04-21 23:23:01 +01:00 committed by GitHub
parent 9332089b3d
commit d35473f7f7
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
377 changed files with 6868 additions and 5785 deletions

View file

@ -91,11 +91,10 @@ class TestRowsGraphQLQueryLogic:
"""Test parsing of schema configuration"""
processor = MagicMock()
processor.schemas = {}
processor.schema_builders = {}
processor.graphql_schemas = {}
processor.config_key = "schema"
processor.schema_builder = MagicMock()
processor.schema_builder.clear = MagicMock()
processor.schema_builder.add_schema = MagicMock()
processor.schema_builder.build = MagicMock(return_value=MagicMock())
processor.query_cassandra = MagicMock()
processor.on_schema_config = Processor.on_schema_config.__get__(processor, Processor)
# Create test config
@ -129,11 +128,11 @@ class TestRowsGraphQLQueryLogic:
}
# Process config
await processor.on_schema_config(schema_config, version=1)
await processor.on_schema_config("default", schema_config, version=1)
# Verify schema was loaded
assert "customer" in processor.schemas
schema = processor.schemas["customer"]
assert "customer" in processor.schemas["default"]
schema = processor.schemas["default"]["customer"]
assert schema.name == "customer"
assert len(schema.fields) == 3
@ -147,39 +146,40 @@ class TestRowsGraphQLQueryLogic:
status_field = next(f for f in schema.fields if f.name == "status")
assert status_field.enum_values == ["active", "inactive"]
# Verify schema builder was called
processor.schema_builder.add_schema.assert_called_once()
processor.schema_builder.build.assert_called_once()
# Verify per-workspace schema builder was created and graphql schema built
assert "default" in processor.schema_builders
assert "default" in processor.graphql_schemas
@pytest.mark.asyncio
async def test_graphql_context_handling(self):
"""Test GraphQL execution context setup"""
processor = MagicMock()
processor.graphql_schema = AsyncMock()
graphql_schema = AsyncMock()
processor.graphql_schemas = {"default": graphql_schema}
processor.execute_graphql_query = Processor.execute_graphql_query.__get__(processor, Processor)
# Mock schema execution
mock_result = MagicMock()
mock_result.data = {"customers": [{"id": "1", "name": "Test"}]}
mock_result.errors = None
processor.graphql_schema.execute.return_value = mock_result
graphql_schema.execute.return_value = mock_result
result = await processor.execute_graphql_query(
workspace="default",
query='{ customers { id name } }',
variables={},
operation_name=None,
user="test_user",
collection="test_collection"
)
# Verify schema.execute was called with correct context
processor.graphql_schema.execute.assert_called_once()
call_args = processor.graphql_schema.execute.call_args
graphql_schema.execute.assert_called_once()
call_args = graphql_schema.execute.call_args
# Verify context was passed
context = call_args[1]['context_value']
assert context["processor"] == processor
assert context["user"] == "test_user"
assert context["workspace"] == "default"
assert context["collection"] == "test_collection"
# Verify result structure
@ -190,7 +190,8 @@ class TestRowsGraphQLQueryLogic:
async def test_error_handling_graphql_errors(self):
"""Test GraphQL error handling and conversion"""
processor = MagicMock()
processor.graphql_schema = AsyncMock()
graphql_schema = AsyncMock()
processor.graphql_schemas = {"default": graphql_schema}
processor.execute_graphql_query = Processor.execute_graphql_query.__get__(processor, Processor)
# Create a simple object to simulate GraphQL error
@ -212,13 +213,13 @@ class TestRowsGraphQLQueryLogic:
mock_result = MagicMock()
mock_result.data = None
mock_result.errors = [mock_error]
processor.graphql_schema.execute.return_value = mock_result
graphql_schema.execute.return_value = mock_result
result = await processor.execute_graphql_query(
workspace="default",
query='{ customers { invalid_field } }',
variables={},
operation_name=None,
user="test_user",
collection="test_collection"
)
@ -248,7 +249,6 @@ class TestRowsGraphQLQueryLogic:
# Create mock message
mock_msg = MagicMock()
mock_request = RowsQueryRequest(
user="test_user",
collection="test_collection",
query='{ customers { id name } }',
variables={},
@ -259,6 +259,7 @@ class TestRowsGraphQLQueryLogic:
# Mock flow
mock_flow = MagicMock()
mock_flow.workspace = "default"
mock_response_flow = AsyncMock()
mock_flow.return_value = mock_response_flow
@ -267,10 +268,10 @@ class TestRowsGraphQLQueryLogic:
# Verify query was executed
processor.execute_graphql_query.assert_called_once_with(
workspace="default",
query='{ customers { id name } }',
variables={},
operation_name=None,
user="test_user",
collection="test_collection"
)
@ -297,7 +298,6 @@ class TestRowsGraphQLQueryLogic:
# Create mock message
mock_msg = MagicMock()
mock_request = RowsQueryRequest(
user="test_user",
collection="test_collection",
query='{ invalid_query }',
variables={},
@ -357,7 +357,7 @@ class TestUnifiedTableQueries:
# Query with filter on indexed field
results = await processor.query_cassandra(
user="test_user",
workspace="test_workspace",
collection="test_collection",
schema_name="products",
row_schema=schema,
@ -374,7 +374,7 @@ class TestUnifiedTableQueries:
query = call_args[0][1]
params = call_args[0][2]
assert "SELECT data, source FROM test_user.rows" in query
assert "SELECT data, source FROM test_workspace.rows" in query
assert "collection = %s" in query
assert "schema_name = %s" in query
assert "index_name = %s" in query
@ -421,7 +421,7 @@ class TestUnifiedTableQueries:
# Query with filter on non-indexed field
results = await processor.query_cassandra(
user="test_user",
workspace="test_workspace",
collection="test_collection",
schema_name="products",
row_schema=schema,