2026-01-06 21:48:05 +00:00
|
|
|
"""
|
|
|
|
|
Unit tests for streaming behavior in message translators.
|
|
|
|
|
|
|
|
|
|
These tests verify that translators correctly handle empty strings and
|
|
|
|
|
end_of_stream flags in streaming responses, preventing bugs where empty
|
|
|
|
|
final chunks could be dropped due to falsy value checks.
|
|
|
|
|
"""
|
|
|
|
|
|
|
|
|
|
import pytest
|
|
|
|
|
from unittest.mock import MagicMock
|
|
|
|
|
from trustgraph.messaging.translators.retrieval import (
|
|
|
|
|
GraphRagResponseTranslator,
|
|
|
|
|
DocumentRagResponseTranslator,
|
|
|
|
|
)
|
|
|
|
|
from trustgraph.messaging.translators.prompt import PromptResponseTranslator
|
|
|
|
|
from trustgraph.messaging.translators.text_completion import TextCompletionResponseTranslator
|
|
|
|
|
from trustgraph.schema import (
|
|
|
|
|
GraphRagResponse,
|
|
|
|
|
DocumentRagResponse,
|
|
|
|
|
PromptResponse,
|
|
|
|
|
TextCompletionResponse,
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
class TestGraphRagResponseTranslator:
|
|
|
|
|
"""Test GraphRagResponseTranslator streaming behavior"""
|
|
|
|
|
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
def test_encode_with_empty_response(self):
|
2026-01-06 21:48:05 +00:00
|
|
|
"""Test that empty response strings are preserved"""
|
|
|
|
|
# Arrange
|
|
|
|
|
translator = GraphRagResponseTranslator()
|
|
|
|
|
response = GraphRagResponse(
|
|
|
|
|
response="",
|
|
|
|
|
end_of_stream=True,
|
|
|
|
|
error=None
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
# Act
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
result = translator.encode(response)
|
2026-01-06 21:48:05 +00:00
|
|
|
|
|
|
|
|
# Assert - Empty string should be included in result
|
|
|
|
|
assert "response" in result
|
|
|
|
|
assert result["response"] == ""
|
|
|
|
|
assert result["end_of_stream"] is True
|
|
|
|
|
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
def test_encode_with_non_empty_response(self):
|
2026-01-06 21:48:05 +00:00
|
|
|
"""Test that non-empty responses work correctly"""
|
|
|
|
|
# Arrange
|
|
|
|
|
translator = GraphRagResponseTranslator()
|
|
|
|
|
response = GraphRagResponse(
|
|
|
|
|
response="Some text",
|
|
|
|
|
end_of_stream=False,
|
|
|
|
|
error=None
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
# Act
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
result = translator.encode(response)
|
2026-01-06 21:48:05 +00:00
|
|
|
|
|
|
|
|
# Assert
|
|
|
|
|
assert result["response"] == "Some text"
|
|
|
|
|
assert result["end_of_stream"] is False
|
|
|
|
|
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
def test_encode_with_none_response(self):
|
2026-01-06 21:48:05 +00:00
|
|
|
"""Test that None response is handled correctly"""
|
|
|
|
|
# Arrange
|
|
|
|
|
translator = GraphRagResponseTranslator()
|
|
|
|
|
response = GraphRagResponse(
|
|
|
|
|
response=None,
|
|
|
|
|
end_of_stream=True,
|
|
|
|
|
error=None
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
# Act
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
result = translator.encode(response)
|
2026-01-06 21:48:05 +00:00
|
|
|
|
|
|
|
|
# Assert - None should not be included
|
|
|
|
|
assert "response" not in result
|
|
|
|
|
assert result["end_of_stream"] is True
|
|
|
|
|
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
def test_encode_with_completion_returns_correct_flag(self):
|
|
|
|
|
"""Test that encode_with_completion returns correct is_final flag"""
|
2026-01-06 21:48:05 +00:00
|
|
|
# Arrange
|
|
|
|
|
translator = GraphRagResponseTranslator()
|
|
|
|
|
|
|
|
|
|
# Test non-final chunk
|
|
|
|
|
response_chunk = GraphRagResponse(
|
|
|
|
|
response="chunk",
|
|
|
|
|
end_of_stream=False,
|
|
|
|
|
error=None
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
# Act
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
result, is_final = translator.encode_with_completion(response_chunk)
|
2026-01-06 21:48:05 +00:00
|
|
|
|
|
|
|
|
# Assert
|
|
|
|
|
assert is_final is False
|
|
|
|
|
assert result["end_of_stream"] is False
|
|
|
|
|
|
GraphRAG Query-Time Explainability (#677)
Implements full explainability pipeline for GraphRAG queries, enabling
traceability from answers back to source documents.
Renamed throughout for clarity:
- provenance_callback → explain_callback
- provenance_id → explain_id
- provenance_collection → explain_collection
- message_type "provenance" → "explain"
- Queue name "provenance" → "explainability"
GraphRAG queries now emit explainability events as they execute:
1. Session - query text and timestamp
2. Retrieval - edges retrieved from subgraph
3. Selection - selected edges with LLM reasoning (JSONL with id +
reasoning)
4. Answer - reference to synthesized response
Events stream via explain_callback during query(), enabling
real-time UX.
- Answers stored in librarian service (not inline in graph - too large)
- Document ID as URN: urn:trustgraph:answer:{session_id}
- Graph stores tg:document reference (IRI) to librarian document
- Added librarian producer/consumer to graph-rag service
- get_labelgraph() now returns (labeled_edges, uri_map)
- uri_map maps edge_id(label_s, label_p, label_o) →
(uri_s, uri_p, uri_o)
- Explainability data stores original URIs, not labels
- Enables tracing edges back to reifying statements via tg:reifies
- Added serialize_triple() to query service (matches storage format)
- get_term_value() now handles TRIPLE type terms
- Enables querying by quoted triple in object position:
?stmt tg:reifies <<s p o>>
- Displays real-time explainability events during query
- Resolves rdfs:label for edge components (s, p, o)
- Traces source chain via prov:wasDerivedFrom to root document
- Output: "Source: Chunk 1 → Page 2 → Document Title"
- Label caching to avoid repeated queries
GraphRagResponse:
- explain_id: str | None
- explain_collection: str | None
- message_type: str ("chunk" or "explain")
- end_of_session: bool
trustgraph-base/trustgraph/provenance/:
- namespaces.py - Added TG_DOCUMENT predicate
- triples.py - answer_triples() supports document_id reference
- uris.py - Added edge_selection_uri()
trustgraph-base/trustgraph/schema/services/retrieval.py:
- GraphRagResponse with explain_id, explain_collection, end_of_session
trustgraph-flow/trustgraph/retrieval/graph_rag/:
- graph_rag.py - URI preservation, streaming answer accumulation
- rag.py - Librarian integration, real-time explain emission
trustgraph-flow/trustgraph/query/triples/cassandra/service.py:
- Quoted triple serialization for query matching
trustgraph-cli/trustgraph/cli/invoke_graph_rag.py:
- Full explainability display with label resolution and source tracing
2026-03-10 10:00:01 +00:00
|
|
|
# Test final message with end_of_session=True
|
2026-01-06 21:48:05 +00:00
|
|
|
final_response = GraphRagResponse(
|
|
|
|
|
response="",
|
|
|
|
|
end_of_stream=True,
|
GraphRAG Query-Time Explainability (#677)
Implements full explainability pipeline for GraphRAG queries, enabling
traceability from answers back to source documents.
Renamed throughout for clarity:
- provenance_callback → explain_callback
- provenance_id → explain_id
- provenance_collection → explain_collection
- message_type "provenance" → "explain"
- Queue name "provenance" → "explainability"
GraphRAG queries now emit explainability events as they execute:
1. Session - query text and timestamp
2. Retrieval - edges retrieved from subgraph
3. Selection - selected edges with LLM reasoning (JSONL with id +
reasoning)
4. Answer - reference to synthesized response
Events stream via explain_callback during query(), enabling
real-time UX.
- Answers stored in librarian service (not inline in graph - too large)
- Document ID as URN: urn:trustgraph:answer:{session_id}
- Graph stores tg:document reference (IRI) to librarian document
- Added librarian producer/consumer to graph-rag service
- get_labelgraph() now returns (labeled_edges, uri_map)
- uri_map maps edge_id(label_s, label_p, label_o) →
(uri_s, uri_p, uri_o)
- Explainability data stores original URIs, not labels
- Enables tracing edges back to reifying statements via tg:reifies
- Added serialize_triple() to query service (matches storage format)
- get_term_value() now handles TRIPLE type terms
- Enables querying by quoted triple in object position:
?stmt tg:reifies <<s p o>>
- Displays real-time explainability events during query
- Resolves rdfs:label for edge components (s, p, o)
- Traces source chain via prov:wasDerivedFrom to root document
- Output: "Source: Chunk 1 → Page 2 → Document Title"
- Label caching to avoid repeated queries
GraphRagResponse:
- explain_id: str | None
- explain_collection: str | None
- message_type: str ("chunk" or "explain")
- end_of_session: bool
trustgraph-base/trustgraph/provenance/:
- namespaces.py - Added TG_DOCUMENT predicate
- triples.py - answer_triples() supports document_id reference
- uris.py - Added edge_selection_uri()
trustgraph-base/trustgraph/schema/services/retrieval.py:
- GraphRagResponse with explain_id, explain_collection, end_of_session
trustgraph-flow/trustgraph/retrieval/graph_rag/:
- graph_rag.py - URI preservation, streaming answer accumulation
- rag.py - Librarian integration, real-time explain emission
trustgraph-flow/trustgraph/query/triples/cassandra/service.py:
- Quoted triple serialization for query matching
trustgraph-cli/trustgraph/cli/invoke_graph_rag.py:
- Full explainability display with label resolution and source tracing
2026-03-10 10:00:01 +00:00
|
|
|
end_of_session=True,
|
2026-01-06 21:48:05 +00:00
|
|
|
error=None
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
# Act
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
result, is_final = translator.encode_with_completion(final_response)
|
2026-01-06 21:48:05 +00:00
|
|
|
|
GraphRAG Query-Time Explainability (#677)
Implements full explainability pipeline for GraphRAG queries, enabling
traceability from answers back to source documents.
Renamed throughout for clarity:
- provenance_callback → explain_callback
- provenance_id → explain_id
- provenance_collection → explain_collection
- message_type "provenance" → "explain"
- Queue name "provenance" → "explainability"
GraphRAG queries now emit explainability events as they execute:
1. Session - query text and timestamp
2. Retrieval - edges retrieved from subgraph
3. Selection - selected edges with LLM reasoning (JSONL with id +
reasoning)
4. Answer - reference to synthesized response
Events stream via explain_callback during query(), enabling
real-time UX.
- Answers stored in librarian service (not inline in graph - too large)
- Document ID as URN: urn:trustgraph:answer:{session_id}
- Graph stores tg:document reference (IRI) to librarian document
- Added librarian producer/consumer to graph-rag service
- get_labelgraph() now returns (labeled_edges, uri_map)
- uri_map maps edge_id(label_s, label_p, label_o) →
(uri_s, uri_p, uri_o)
- Explainability data stores original URIs, not labels
- Enables tracing edges back to reifying statements via tg:reifies
- Added serialize_triple() to query service (matches storage format)
- get_term_value() now handles TRIPLE type terms
- Enables querying by quoted triple in object position:
?stmt tg:reifies <<s p o>>
- Displays real-time explainability events during query
- Resolves rdfs:label for edge components (s, p, o)
- Traces source chain via prov:wasDerivedFrom to root document
- Output: "Source: Chunk 1 → Page 2 → Document Title"
- Label caching to avoid repeated queries
GraphRagResponse:
- explain_id: str | None
- explain_collection: str | None
- message_type: str ("chunk" or "explain")
- end_of_session: bool
trustgraph-base/trustgraph/provenance/:
- namespaces.py - Added TG_DOCUMENT predicate
- triples.py - answer_triples() supports document_id reference
- uris.py - Added edge_selection_uri()
trustgraph-base/trustgraph/schema/services/retrieval.py:
- GraphRagResponse with explain_id, explain_collection, end_of_session
trustgraph-flow/trustgraph/retrieval/graph_rag/:
- graph_rag.py - URI preservation, streaming answer accumulation
- rag.py - Librarian integration, real-time explain emission
trustgraph-flow/trustgraph/query/triples/cassandra/service.py:
- Quoted triple serialization for query matching
trustgraph-cli/trustgraph/cli/invoke_graph_rag.py:
- Full explainability display with label resolution and source tracing
2026-03-10 10:00:01 +00:00
|
|
|
# Assert - is_final is based on end_of_session, not end_of_stream
|
2026-01-06 21:48:05 +00:00
|
|
|
assert is_final is True
|
|
|
|
|
assert result["response"] == ""
|
GraphRAG Query-Time Explainability (#677)
Implements full explainability pipeline for GraphRAG queries, enabling
traceability from answers back to source documents.
Renamed throughout for clarity:
- provenance_callback → explain_callback
- provenance_id → explain_id
- provenance_collection → explain_collection
- message_type "provenance" → "explain"
- Queue name "provenance" → "explainability"
GraphRAG queries now emit explainability events as they execute:
1. Session - query text and timestamp
2. Retrieval - edges retrieved from subgraph
3. Selection - selected edges with LLM reasoning (JSONL with id +
reasoning)
4. Answer - reference to synthesized response
Events stream via explain_callback during query(), enabling
real-time UX.
- Answers stored in librarian service (not inline in graph - too large)
- Document ID as URN: urn:trustgraph:answer:{session_id}
- Graph stores tg:document reference (IRI) to librarian document
- Added librarian producer/consumer to graph-rag service
- get_labelgraph() now returns (labeled_edges, uri_map)
- uri_map maps edge_id(label_s, label_p, label_o) →
(uri_s, uri_p, uri_o)
- Explainability data stores original URIs, not labels
- Enables tracing edges back to reifying statements via tg:reifies
- Added serialize_triple() to query service (matches storage format)
- get_term_value() now handles TRIPLE type terms
- Enables querying by quoted triple in object position:
?stmt tg:reifies <<s p o>>
- Displays real-time explainability events during query
- Resolves rdfs:label for edge components (s, p, o)
- Traces source chain via prov:wasDerivedFrom to root document
- Output: "Source: Chunk 1 → Page 2 → Document Title"
- Label caching to avoid repeated queries
GraphRagResponse:
- explain_id: str | None
- explain_collection: str | None
- message_type: str ("chunk" or "explain")
- end_of_session: bool
trustgraph-base/trustgraph/provenance/:
- namespaces.py - Added TG_DOCUMENT predicate
- triples.py - answer_triples() supports document_id reference
- uris.py - Added edge_selection_uri()
trustgraph-base/trustgraph/schema/services/retrieval.py:
- GraphRagResponse with explain_id, explain_collection, end_of_session
trustgraph-flow/trustgraph/retrieval/graph_rag/:
- graph_rag.py - URI preservation, streaming answer accumulation
- rag.py - Librarian integration, real-time explain emission
trustgraph-flow/trustgraph/query/triples/cassandra/service.py:
- Quoted triple serialization for query matching
trustgraph-cli/trustgraph/cli/invoke_graph_rag.py:
- Full explainability display with label resolution and source tracing
2026-03-10 10:00:01 +00:00
|
|
|
assert result["end_of_session"] is True
|
2026-01-06 21:48:05 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
class TestDocumentRagResponseTranslator:
|
|
|
|
|
"""Test DocumentRagResponseTranslator streaming behavior"""
|
|
|
|
|
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
def test_encode_with_empty_response(self):
|
2026-01-06 21:48:05 +00:00
|
|
|
"""Test that empty response strings are preserved"""
|
|
|
|
|
# Arrange
|
|
|
|
|
translator = DocumentRagResponseTranslator()
|
|
|
|
|
response = DocumentRagResponse(
|
|
|
|
|
response="",
|
|
|
|
|
end_of_stream=True,
|
|
|
|
|
error=None
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
# Act
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
result = translator.encode(response)
|
2026-01-06 21:48:05 +00:00
|
|
|
|
|
|
|
|
# Assert
|
|
|
|
|
assert "response" in result
|
|
|
|
|
assert result["response"] == ""
|
|
|
|
|
assert result["end_of_stream"] is True
|
|
|
|
|
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
def test_encode_with_non_empty_response(self):
|
2026-01-06 21:48:05 +00:00
|
|
|
"""Test that non-empty responses work correctly"""
|
|
|
|
|
# Arrange
|
|
|
|
|
translator = DocumentRagResponseTranslator()
|
|
|
|
|
response = DocumentRagResponse(
|
|
|
|
|
response="Document content",
|
|
|
|
|
end_of_stream=False,
|
|
|
|
|
error=None
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
# Act
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
result = translator.encode(response)
|
2026-01-06 21:48:05 +00:00
|
|
|
|
|
|
|
|
# Assert
|
|
|
|
|
assert result["response"] == "Document content"
|
|
|
|
|
assert result["end_of_stream"] is False
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
class TestPromptResponseTranslator:
|
|
|
|
|
"""Test PromptResponseTranslator streaming behavior"""
|
|
|
|
|
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
def test_encode_with_empty_text(self):
|
2026-01-06 21:48:05 +00:00
|
|
|
"""Test that empty text strings are preserved"""
|
|
|
|
|
# Arrange
|
|
|
|
|
translator = PromptResponseTranslator()
|
|
|
|
|
response = PromptResponse(
|
|
|
|
|
text="",
|
|
|
|
|
object=None,
|
|
|
|
|
end_of_stream=True,
|
|
|
|
|
error=None
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
# Act
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
result = translator.encode(response)
|
2026-01-06 21:48:05 +00:00
|
|
|
|
|
|
|
|
# Assert
|
|
|
|
|
assert "text" in result
|
|
|
|
|
assert result["text"] == ""
|
|
|
|
|
assert result["end_of_stream"] is True
|
|
|
|
|
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
def test_encode_with_non_empty_text(self):
|
2026-01-06 21:48:05 +00:00
|
|
|
"""Test that non-empty text works correctly"""
|
|
|
|
|
# Arrange
|
|
|
|
|
translator = PromptResponseTranslator()
|
|
|
|
|
response = PromptResponse(
|
|
|
|
|
text="Some prompt response",
|
|
|
|
|
object=None,
|
|
|
|
|
end_of_stream=False,
|
|
|
|
|
error=None
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
# Act
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
result = translator.encode(response)
|
2026-01-06 21:48:05 +00:00
|
|
|
|
|
|
|
|
# Assert
|
|
|
|
|
assert result["text"] == "Some prompt response"
|
|
|
|
|
assert result["end_of_stream"] is False
|
|
|
|
|
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
def test_encode_with_none_text(self):
|
2026-01-06 21:48:05 +00:00
|
|
|
"""Test that None text is handled correctly"""
|
|
|
|
|
# Arrange
|
|
|
|
|
translator = PromptResponseTranslator()
|
|
|
|
|
response = PromptResponse(
|
|
|
|
|
text=None,
|
|
|
|
|
object='{"result": "data"}',
|
|
|
|
|
end_of_stream=True,
|
|
|
|
|
error=None
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
# Act
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
result = translator.encode(response)
|
2026-01-06 21:48:05 +00:00
|
|
|
|
|
|
|
|
# Assert
|
|
|
|
|
assert "text" not in result
|
|
|
|
|
assert "object" in result
|
|
|
|
|
assert result["end_of_stream"] is True
|
|
|
|
|
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
def test_encode_includes_end_of_stream(self):
|
2026-01-06 21:48:05 +00:00
|
|
|
"""Test that end_of_stream flag is always included"""
|
|
|
|
|
# Arrange
|
|
|
|
|
translator = PromptResponseTranslator()
|
|
|
|
|
|
|
|
|
|
# Test with end_of_stream=False
|
|
|
|
|
response = PromptResponse(
|
|
|
|
|
text="chunk",
|
|
|
|
|
object=None,
|
|
|
|
|
end_of_stream=False,
|
|
|
|
|
error=None
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
# Act
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
result = translator.encode(response)
|
2026-01-06 21:48:05 +00:00
|
|
|
|
|
|
|
|
# Assert
|
|
|
|
|
assert "end_of_stream" in result
|
|
|
|
|
assert result["end_of_stream"] is False
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
class TestTextCompletionResponseTranslator:
|
|
|
|
|
"""Test TextCompletionResponseTranslator streaming behavior"""
|
|
|
|
|
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
def test_encode_always_includes_response(self):
|
2026-01-06 21:48:05 +00:00
|
|
|
"""Test that response field is always included, even if empty"""
|
|
|
|
|
# Arrange
|
|
|
|
|
translator = TextCompletionResponseTranslator()
|
|
|
|
|
response = TextCompletionResponse(
|
|
|
|
|
response="",
|
|
|
|
|
end_of_stream=True,
|
|
|
|
|
error=None,
|
|
|
|
|
in_token=100,
|
|
|
|
|
out_token=5,
|
|
|
|
|
model="test-model"
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
# Act
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
result = translator.encode(response)
|
2026-01-06 21:48:05 +00:00
|
|
|
|
|
|
|
|
# Assert - Response should always be present
|
|
|
|
|
assert "response" in result
|
|
|
|
|
assert result["response"] == ""
|
|
|
|
|
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
def test_encode_with_completion_with_empty_final(self):
|
2026-01-06 21:48:05 +00:00
|
|
|
"""Test that empty final response is handled correctly"""
|
|
|
|
|
# Arrange
|
|
|
|
|
translator = TextCompletionResponseTranslator()
|
|
|
|
|
response = TextCompletionResponse(
|
|
|
|
|
response="",
|
|
|
|
|
end_of_stream=True,
|
|
|
|
|
error=None,
|
|
|
|
|
in_token=100,
|
|
|
|
|
out_token=5,
|
|
|
|
|
model="test-model"
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
# Act
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
result, is_final = translator.encode_with_completion(response)
|
2026-01-06 21:48:05 +00:00
|
|
|
|
|
|
|
|
# Assert
|
|
|
|
|
assert is_final is True
|
|
|
|
|
assert result["response"] == ""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
class TestStreamingProtocolCompliance:
|
|
|
|
|
"""Test that all translators follow streaming protocol conventions"""
|
|
|
|
|
|
|
|
|
|
@pytest.mark.parametrize("translator_class,response_class,field_name", [
|
|
|
|
|
(GraphRagResponseTranslator, GraphRagResponse, "response"),
|
|
|
|
|
(DocumentRagResponseTranslator, DocumentRagResponse, "response"),
|
|
|
|
|
(PromptResponseTranslator, PromptResponse, "text"),
|
|
|
|
|
(TextCompletionResponseTranslator, TextCompletionResponse, "response"),
|
|
|
|
|
])
|
|
|
|
|
def test_empty_final_chunk_preserved(self, translator_class, response_class, field_name):
|
|
|
|
|
"""Test that all translators preserve empty final chunks"""
|
|
|
|
|
# Arrange
|
|
|
|
|
translator = translator_class()
|
|
|
|
|
kwargs = {
|
|
|
|
|
field_name: "",
|
|
|
|
|
"end_of_stream": True,
|
|
|
|
|
"error": None,
|
|
|
|
|
}
|
|
|
|
|
response = response_class(**kwargs)
|
|
|
|
|
|
|
|
|
|
# Act
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
result = translator.encode(response)
|
2026-01-06 21:48:05 +00:00
|
|
|
|
|
|
|
|
# Assert
|
|
|
|
|
assert field_name in result, f"{translator_class.__name__} should include '{field_name}' field even when empty"
|
|
|
|
|
assert result[field_name] == "", f"{translator_class.__name__} should preserve empty string"
|
|
|
|
|
|
|
|
|
|
@pytest.mark.parametrize("translator_class,response_class,field_name", [
|
|
|
|
|
(GraphRagResponseTranslator, GraphRagResponse, "response"),
|
|
|
|
|
(DocumentRagResponseTranslator, DocumentRagResponse, "response"),
|
|
|
|
|
(TextCompletionResponseTranslator, TextCompletionResponse, "response"),
|
|
|
|
|
])
|
|
|
|
|
def test_end_of_stream_flag_included(self, translator_class, response_class, field_name):
|
|
|
|
|
"""Test that end_of_stream flag is included in all response translators"""
|
|
|
|
|
# Arrange
|
|
|
|
|
translator = translator_class()
|
|
|
|
|
kwargs = {
|
|
|
|
|
field_name: "test content",
|
|
|
|
|
"end_of_stream": True,
|
|
|
|
|
"error": None,
|
|
|
|
|
}
|
|
|
|
|
response = response_class(**kwargs)
|
|
|
|
|
|
|
|
|
|
# Act
|
Pub/sub abstraction: decouple from Pulsar (#751)
Remove Pulsar-specific concepts from application code so that
the pub/sub backend is swappable via configuration.
Rename translators:
- to_pulsar/from_pulsar → decode/encode across all translator
classes, dispatch handlers, and tests (55+ files)
- from_response_with_completion → encode_with_completion
- Remove pulsar.schema.Record from translator base class
Queue naming (CLASS:TOPICSPACE:TOPIC):
- Replace topic() helper with queue() using new format:
flow:tg:name, request:tg:name, response:tg:name, state:tg:name
- Queue class implies persistence/TTL (no QoS in names)
- Update Pulsar backend map_topic() to parse new format
- Librarian queues use flow class (persistent, for chunking)
- Config push uses state class (persistent, last-value)
- Remove 15 dead topic imports from schema files
- Update init_trustgraph.py namespace: config → state
Confine Pulsar to pulsar_backend.py:
- Delete legacy PulsarClient class from pubsub.py
- Move add_args to add_pubsub_args() with standalone flag
for CLI tools (defaults to localhost)
- PulsarBackendConsumer.receive() catches _pulsar.Timeout,
raises standard TimeoutError
- Remove Pulsar imports from: async_processor, flow_processor,
log_level, all 11 client files, 4 storage writers, gateway
service, gateway config receiver
- Remove log_level/LoggerLevel from client API
- Rewrite tg-monitor-prompts to use backend abstraction
- Update tg-dump-queues to use add_pubsub_args
Also: pubsub-abstraction.md tech spec covering problem statement,
design goals, as-is requirements, candidate broker assessment,
approach, and implementation order.
2026-04-01 20:16:53 +01:00
|
|
|
result = translator.encode(response)
|
2026-01-06 21:48:05 +00:00
|
|
|
|
|
|
|
|
# Assert
|
|
|
|
|
assert "end_of_stream" in result, f"{translator_class.__name__} should include 'end_of_stream' flag"
|
|
|
|
|
assert result["end_of_stream"] is True
|