apunkt/trustgraph
1
0
Fork 0
mirror of https://github.com/trustgraph-ai/trustgraph.git synced 2026-04-26 17:06:22 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

Release/v1.2 (#457)

Browse source 
* Bump setup.py versions for 1.1

* PoC MCP server (#419)

* Very initial MCP server PoC for TrustGraph

* Put service on port 8000

* Add MCP container and packages to buildout

* Update docs for API/CLI changes in 1.0 (#421)

* Update some API basics for the 0.23/1.0 API change

* Add MCP container push (#425)

* Add command args to the MCP server (#426)

* Host and port parameters

* Added websocket arg

* More docs

* MCP client support (#427)

- MCP client service
- Tool request/response schema
- API gateway support for mcp-tool
- Message translation for tool request & response
- Make mcp-tool using configuration service for information
  about where the MCP services are.

* Feature/react call mcp (#428)

Key Features

  - MCP Tool Integration: Added core MCP tool support with ToolClientSpec and ToolClient classes
  - API Enhancement: New mcp_tool method for flow-specific tool invocation
  - CLI Tooling: New tg-invoke-mcp-tool command for testing MCP integration
  - React Agent Enhancement: Fixed and improved multi-tool invocation capabilities
  - Tool Management: Enhanced CLI for tool configuration and management

Changes

  - Added MCP tool invocation to API with flow-specific integration
  - Implemented ToolClientSpec and ToolClient for tool call handling
  - Updated agent-manager-react to invoke MCP tools with configurable types
  - Enhanced CLI with new commands and improved help text
  - Added comprehensive documentation for new CLI commands
  - Improved tool configuration management

Testing

  - Added tg-invoke-mcp-tool CLI command for isolated MCP integration testing
  - Enhanced agent capability to invoke multiple tools simultaneously

* Test suite executed from CI pipeline (#433)

* Test strategy & test cases

* Unit tests

* Integration tests

* Extending test coverage (#434)

* Contract tests

* Testing embeedings

* Agent unit tests

* Knowledge pipeline tests

* Turn on contract tests

* Increase storage test coverage (#435)

* Fixing storage and adding tests

* PR pipeline only runs quick tests

* Empty configuration is returned as empty list, previously was not in response (#436)

* Update config util to take files as well as command-line text (#437)

* Updated CLI invocation and config model for tools and mcp (#438)

* Updated CLI invocation and config model for tools and mcp

* CLI anomalies

* Tweaked the MCP tool implementation for new model

* Update agent implementation to match the new model

* Fix agent tools, now all tested

* Fixed integration tests

* Fix MCP delete tool params

* Update Python deps to 1.2

* Update to enable knowledge extraction using the agent framework (#439)

* Implement KG extraction agent (kg-extract-agent)

* Using ReAct framework (agent-manager-react)
 
* ReAct manager had an issue when emitting JSON, which conflicts which ReAct manager's own JSON messages, so refactored ReAct manager to use traditional ReAct messages, non-JSON structure.
 
* Minor refactor to take the prompt template client out of prompt-template so it can be more readily used by other modules. kg-extract-agent uses this framework.

* Migrate from setup.py to pyproject.toml (#440)

* Converted setup.py to pyproject.toml

* Modern package infrastructure as recommended by py docs

* Install missing build deps (#441)

* Install missing build deps (#442)

* Implement logging strategy (#444)

* Logging strategy and convert all prints() to logging invocations

* Fix/startup failure (#445)

* Fix loggin startup problems

* Fix logging startup problems (#446)

* Fix logging startup problems (#447)

* Fixed Mistral OCR to use current API (#448)

* Fixed Mistral OCR to use current API

* Added PDF decoder tests

* Fix Mistral OCR ident to be standard pdf-decoder (#450)

* Fix Mistral OCR ident to be standard pdf-decoder

* Correct test

* Schema structure refactor (#451)

* Write schema refactor spec

* Implemented schema refactor spec

* Structure data mvp (#452)

* Structured data tech spec

* Architecture principles

* New schemas

* Updated schemas and specs

* Object extractor

* Add .coveragerc

* New tests

* Cassandra object storage

* Trying to object extraction working, issues exist

* Validate librarian collection (#453)

* Fix token chunker, broken API invocation (#454)

* Fix token chunker, broken API invocation (#455)

* Knowledge load utility CLI (#456)

* Knowledge loader

* More tests
This commit is contained in:
cybermaggedon 2025-08-18 20:56:09 +01:00 • committed by GitHub
parent c85ba197be
commit 89be656990
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
509 changed files with 49632 additions and 5159 deletions
Download patch file Download diff file Expand all files Collapse all files

0
tests/unit/test_chunking/__init__.py Normal file
View file

153
tests/unit/test_chunking/conftest.py Normal file
View file

@ -0,0 +1,153 @@
import pytest
from unittest.mock import AsyncMock, Mock, patch
from trustgraph.schema import TextDocument, Metadata
from trustgraph.chunking.recursive.chunker import Processor as RecursiveChunker
from trustgraph.chunking.token.chunker import Processor as TokenChunker
from prometheus_client import REGISTRY
@pytest.fixture
def mock_flow():
"""Mock flow function that returns a mock output producer."""
output_mock = AsyncMock()
flow_mock = Mock(return_value=output_mock)
return flow_mock, output_mock
@pytest.fixture
def mock_consumer():
"""Mock consumer with test attributes."""
consumer = Mock()
consumer.id = "test-consumer"
consumer.flow = "test-flow"
return consumer
@pytest.fixture
def sample_text_document():
"""Sample document with moderate length text."""
metadata = Metadata(
id="test-doc-1",
metadata=[],
user="test-user",
collection="test-collection"
)
text = "The quick brown fox jumps over the lazy dog. " * 20
return TextDocument(
metadata=metadata,
text=text.encode("utf-8")
)
@pytest.fixture
def long_text_document():
"""Long document for testing multiple chunks."""
metadata = Metadata(
id="test-doc-long",
metadata=[],
user="test-user",
collection="test-collection"
)
# Create a long text that will definitely be chunked
text = " ".join([f"Sentence number {i}. This is part of a long document." for i in range(200)])
return TextDocument(
metadata=metadata,
text=text.encode("utf-8")
)
@pytest.fixture
def unicode_text_document():
"""Document with various unicode characters."""
metadata = Metadata(
id="test-doc-unicode",
metadata=[],
user="test-user",
collection="test-collection"
)
text = """
English: Hello World!
Chinese: 你好世界
Japanese: こんにちは世界
Korean: 안녕하세요 세계
Arabic: مرحبا بالعالم
Russian: Привет мир
Emoji: 🌍 🌎 🌏 😀 🎉
Math: ∑ ∏ ∫ ∞ √ π
Symbols: © ® ™ € £ ¥
"""
return TextDocument(
metadata=metadata,
text=text.encode("utf-8")
)
@pytest.fixture
def empty_text_document():
"""Empty document for edge case testing."""
metadata = Metadata(
id="test-doc-empty",
metadata=[],
user="test-user",
collection="test-collection"
)
return TextDocument(
metadata=metadata,
text=b""
)
@pytest.fixture
def mock_message(sample_text_document):
"""Mock message containing a document."""
msg = Mock()
msg.value.return_value = sample_text_document
return msg
@pytest.fixture(autouse=True)
def clear_metrics():
"""Clear metrics before each test to avoid duplicates."""
# Clear the chunk_metric class attribute if it exists
if hasattr(RecursiveChunker, 'chunk_metric'):
# Unregister from Prometheus registry first
try:
REGISTRY.unregister(RecursiveChunker.chunk_metric)
except KeyError:
pass # Already unregistered
delattr(RecursiveChunker, 'chunk_metric')
if hasattr(TokenChunker, 'chunk_metric'):
try:
REGISTRY.unregister(TokenChunker.chunk_metric)
except KeyError:
pass # Already unregistered
delattr(TokenChunker, 'chunk_metric')
yield
# Clean up after test as well
if hasattr(RecursiveChunker, 'chunk_metric'):
try:
REGISTRY.unregister(RecursiveChunker.chunk_metric)
except KeyError:
pass
delattr(RecursiveChunker, 'chunk_metric')
if hasattr(TokenChunker, 'chunk_metric'):
try:
REGISTRY.unregister(TokenChunker.chunk_metric)
except KeyError:
pass
delattr(TokenChunker, 'chunk_metric')
@pytest.fixture
def mock_async_processor_init():
"""Mock AsyncProcessor.__init__ to avoid taskgroup requirement."""
def init_mock(self, **kwargs):
# Set attributes that AsyncProcessor would normally set
self.config_handlers = []
self.specifications = []
self.flows = {}
self.id = kwargs.get('id', 'test-processor')
# Don't call the real __init__
with patch('trustgraph.base.async_processor.AsyncProcessor.__init__', init_mock):
yield

211
tests/unit/test_chunking/test_recursive_chunker.py Normal file
View file

@ -0,0 +1,211 @@
import pytest
import asyncio
from unittest.mock import AsyncMock, Mock, patch, MagicMock
from trustgraph.schema import TextDocument, Chunk, Metadata
from trustgraph.chunking.recursive.chunker import Processor as RecursiveChunker
@pytest.fixture
def mock_flow():
output_mock = AsyncMock()
flow_mock = Mock(return_value=output_mock)
return flow_mock, output_mock
@pytest.fixture
def mock_consumer():
consumer = Mock()
consumer.id = "test-consumer"
consumer.flow = "test-flow"
return consumer
@pytest.fixture
def sample_document():
metadata = Metadata(
id="test-doc-1",
metadata=[],
user="test-user",
collection="test-collection"
)
text = "This is a test document. " * 100 # Create text long enough to be chunked
return TextDocument(
metadata=metadata,
text=text.encode("utf-8")
)
@pytest.fixture
def short_document():
metadata = Metadata(
id="test-doc-2",
metadata=[],
user="test-user",
collection="test-collection"
)
text = "This is a very short document."
return TextDocument(
metadata=metadata,
text=text.encode("utf-8")
)
class TestRecursiveChunker:
def test_init_default_params(self, mock_async_processor_init):
processor = RecursiveChunker()
assert processor.text_splitter._chunk_size == 2000
assert processor.text_splitter._chunk_overlap == 100
def test_init_custom_params(self, mock_async_processor_init):
processor = RecursiveChunker(chunk_size=500, chunk_overlap=50)
assert processor.text_splitter._chunk_size == 500
assert processor.text_splitter._chunk_overlap == 50
def test_init_with_id(self, mock_async_processor_init):
processor = RecursiveChunker(id="custom-chunker")
assert processor.id == "custom-chunker"
@pytest.mark.asyncio
async def test_on_message_single_chunk(self, mock_async_processor_init, mock_flow, mock_consumer, short_document):
flow_mock, output_mock = mock_flow
processor = RecursiveChunker(chunk_size=2000, chunk_overlap=100)
msg = Mock()
msg.value.return_value = short_document
await processor.on_message(msg, mock_consumer, flow_mock)
# Should produce exactly one chunk for short text
assert output_mock.send.call_count == 1
# Verify the chunk was created correctly
chunk_call = output_mock.send.call_args[0][0]
assert isinstance(chunk_call, Chunk)
assert chunk_call.metadata == short_document.metadata
assert chunk_call.chunk.decode("utf-8") == short_document.text.decode("utf-8")
@pytest.mark.asyncio
async def test_on_message_multiple_chunks(self, mock_async_processor_init, mock_flow, mock_consumer, sample_document):
flow_mock, output_mock = mock_flow
processor = RecursiveChunker(chunk_size=100, chunk_overlap=20)
msg = Mock()
msg.value.return_value = sample_document
await processor.on_message(msg, mock_consumer, flow_mock)
# Should produce multiple chunks
assert output_mock.send.call_count > 1
# Verify all chunks have correct metadata
for call in output_mock.send.call_args_list:
chunk = call[0][0]
assert isinstance(chunk, Chunk)
assert chunk.metadata == sample_document.metadata
assert len(chunk.chunk) > 0
@pytest.mark.asyncio
async def test_on_message_chunk_overlap(self, mock_async_processor_init, mock_flow, mock_consumer):
flow_mock, output_mock = mock_flow
processor = RecursiveChunker(chunk_size=50, chunk_overlap=10)
# Create a document with predictable content
metadata = Metadata(id="test", metadata=[], user="test-user", collection="test-collection")
text = "ABCDEFGHIJ" * 10 # 100 characters
document = TextDocument(metadata=metadata, text=text.encode("utf-8"))
msg = Mock()
msg.value.return_value = document
await processor.on_message(msg, mock_consumer, flow_mock)
# Collect all chunks
chunks = []
for call in output_mock.send.call_args_list:
chunk_text = call[0][0].chunk.decode("utf-8")
chunks.append(chunk_text)
# Verify chunks have expected overlap
for i in range(len(chunks) - 1):
# The end of chunk i should overlap with the beginning of chunk i+1
# Check if there's some overlap (exact overlap depends on text splitter logic)
assert len(chunks[i]) <= 50 + 10 # chunk_size + some tolerance
@pytest.mark.asyncio
async def test_on_message_empty_document(self, mock_async_processor_init, mock_flow, mock_consumer):
flow_mock, output_mock = mock_flow
processor = RecursiveChunker()
metadata = Metadata(id="empty", metadata=[], user="test-user", collection="test-collection")
document = TextDocument(metadata=metadata, text=b"")
msg = Mock()
msg.value.return_value = document
await processor.on_message(msg, mock_consumer, flow_mock)
# Empty documents typically don't produce chunks with langchain splitters
# This behavior is expected - no chunks should be produced
assert output_mock.send.call_count == 0
@pytest.mark.asyncio
async def test_on_message_unicode_handling(self, mock_async_processor_init, mock_flow, mock_consumer):
flow_mock, output_mock = mock_flow
processor = RecursiveChunker(chunk_size=500, chunk_overlap=20) # Fixed overlap < chunk_size
metadata = Metadata(id="unicode", metadata=[], user="test-user", collection="test-collection")
text = "Hello 世界! 🌍 This is a test with émojis and spëcial characters."
document = TextDocument(metadata=metadata, text=text.encode("utf-8"))
msg = Mock()
msg.value.return_value = document
await processor.on_message(msg, mock_consumer, flow_mock)
# Verify unicode is preserved correctly
all_chunks = []
for call in output_mock.send.call_args_list:
chunk_text = call[0][0].chunk.decode("utf-8")
all_chunks.append(chunk_text)
# Reconstruct text (approximately, due to overlap)
reconstructed = "".join(all_chunks)
assert "世界" in reconstructed
assert "🌍" in reconstructed
assert "émojis" in reconstructed
@pytest.mark.asyncio
async def test_metrics_recorded(self, mock_async_processor_init, mock_flow, mock_consumer, sample_document):
flow_mock, output_mock = mock_flow
processor = RecursiveChunker(chunk_size=100)
msg = Mock()
msg.value.return_value = sample_document
# Mock the metric
with patch.object(RecursiveChunker.chunk_metric, 'labels') as mock_labels:
mock_observe = Mock()
mock_labels.return_value.observe = mock_observe
await processor.on_message(msg, mock_consumer, flow_mock)
# Verify metrics were recorded
mock_labels.assert_called_with(id="test-consumer", flow="test-flow")
assert mock_observe.call_count > 0
# Verify chunk sizes were observed
for call in mock_observe.call_args_list:
chunk_size = call[0][0]
assert chunk_size > 0
def test_add_args(self):
parser = Mock()
RecursiveChunker.add_args(parser)
# Verify arguments were added
calls = parser.add_argument.call_args_list
arg_names = [call[0][0] for call in calls]
assert '-z' in arg_names or '--chunk-size' in arg_names
assert '-v' in arg_names or '--chunk-overlap' in arg_names

275
tests/unit/test_chunking/test_token_chunker.py Normal file
View file

@ -0,0 +1,275 @@
import pytest
import asyncio
from unittest.mock import AsyncMock, Mock, patch
from trustgraph.schema import TextDocument, Chunk, Metadata
from trustgraph.chunking.token.chunker import Processor as TokenChunker
@pytest.fixture
def mock_flow():
output_mock = AsyncMock()
flow_mock = Mock(return_value=output_mock)
return flow_mock, output_mock
@pytest.fixture
def mock_consumer():
consumer = Mock()
consumer.id = "test-consumer"
consumer.flow = "test-flow"
return consumer
@pytest.fixture
def sample_document():
metadata = Metadata(
id="test-doc-1",
metadata=[],
user="test-user",
collection="test-collection"
)
# Create text that will result in multiple token chunks
text = "The quick brown fox jumps over the lazy dog. " * 50
return TextDocument(
metadata=metadata,
text=text.encode("utf-8")
)
@pytest.fixture
def short_document():
metadata = Metadata(
id="test-doc-2",
metadata=[],
user="test-user",
collection="test-collection"
)
text = "Short text."
return TextDocument(
metadata=metadata,
text=text.encode("utf-8")
)
class TestTokenChunker:
def test_init_default_params(self, mock_async_processor_init):
processor = TokenChunker()
assert processor.text_splitter._chunk_size == 250
assert processor.text_splitter._chunk_overlap == 15
# Just verify the text splitter was created (encoding verification is complex)
assert processor.text_splitter is not None
assert hasattr(processor.text_splitter, 'split_text')
def test_init_custom_params(self, mock_async_processor_init):
processor = TokenChunker(chunk_size=100, chunk_overlap=10)
assert processor.text_splitter._chunk_size == 100
assert processor.text_splitter._chunk_overlap == 10
def test_init_with_id(self, mock_async_processor_init):
processor = TokenChunker(id="custom-token-chunker")
assert processor.id == "custom-token-chunker"
@pytest.mark.asyncio
async def test_on_message_single_chunk(self, mock_async_processor_init, mock_flow, mock_consumer, short_document):
flow_mock, output_mock = mock_flow
processor = TokenChunker(chunk_size=250, chunk_overlap=15)
msg = Mock()
msg.value.return_value = short_document
await processor.on_message(msg, mock_consumer, flow_mock)
# Short text should produce exactly one chunk
assert output_mock.send.call_count == 1
# Verify the chunk was created correctly
chunk_call = output_mock.send.call_args[0][0]
assert isinstance(chunk_call, Chunk)
assert chunk_call.metadata == short_document.metadata
assert chunk_call.chunk.decode("utf-8") == short_document.text.decode("utf-8")
@pytest.mark.asyncio
async def test_on_message_multiple_chunks(self, mock_async_processor_init, mock_flow, mock_consumer, sample_document):
flow_mock, output_mock = mock_flow
processor = TokenChunker(chunk_size=50, chunk_overlap=5)
msg = Mock()
msg.value.return_value = sample_document
await processor.on_message(msg, mock_consumer, flow_mock)
# Should produce multiple chunks
assert output_mock.send.call_count > 1
# Verify all chunks have correct metadata
for call in output_mock.send.call_args_list:
chunk = call[0][0]
assert isinstance(chunk, Chunk)
assert chunk.metadata == sample_document.metadata
assert len(chunk.chunk) > 0
@pytest.mark.asyncio
async def test_on_message_token_overlap(self, mock_async_processor_init, mock_flow, mock_consumer):
flow_mock, output_mock = mock_flow
processor = TokenChunker(chunk_size=20, chunk_overlap=5)
# Create a document with repeated pattern
metadata = Metadata(id="test", metadata=[], user="test-user", collection="test-collection")
text = "one two three four five six seven eight nine ten " * 5
document = TextDocument(metadata=metadata, text=text.encode("utf-8"))
msg = Mock()
msg.value.return_value = document
await processor.on_message(msg, mock_consumer, flow_mock)
# Collect all chunks
chunks = []
for call in output_mock.send.call_args_list:
chunk_text = call[0][0].chunk.decode("utf-8")
chunks.append(chunk_text)
# Should have multiple chunks
assert len(chunks) > 1
# Verify chunks are not empty
for chunk in chunks:
assert len(chunk) > 0
@pytest.mark.asyncio
async def test_on_message_empty_document(self, mock_async_processor_init, mock_flow, mock_consumer):
flow_mock, output_mock = mock_flow
processor = TokenChunker()
metadata = Metadata(id="empty", metadata=[], user="test-user", collection="test-collection")
document = TextDocument(metadata=metadata, text=b"")
msg = Mock()
msg.value.return_value = document
await processor.on_message(msg, mock_consumer, flow_mock)
# Empty documents typically don't produce chunks with langchain splitters
# This behavior is expected - no chunks should be produced
assert output_mock.send.call_count == 0
@pytest.mark.asyncio
async def test_on_message_unicode_handling(self, mock_async_processor_init, mock_flow, mock_consumer):
flow_mock, output_mock = mock_flow
processor = TokenChunker(chunk_size=50)
metadata = Metadata(id="unicode", metadata=[], user="test-user", collection="test-collection")
# Test with various unicode characters
text = "Hello 世界! 🌍 Test émojis café naïve résumé. Greek: αβγδε Hebrew: אבגדה"
document = TextDocument(metadata=metadata, text=text.encode("utf-8"))
msg = Mock()
msg.value.return_value = document
await processor.on_message(msg, mock_consumer, flow_mock)
# Verify unicode is preserved correctly
all_chunks = []
for call in output_mock.send.call_args_list:
chunk_text = call[0][0].chunk.decode("utf-8")
all_chunks.append(chunk_text)
# Reconstruct text
reconstructed = "".join(all_chunks)
assert "世界" in reconstructed
assert "🌍" in reconstructed
assert "émojis" in reconstructed
assert "αβγδε" in reconstructed
assert "אבגדה" in reconstructed
@pytest.mark.asyncio
async def test_on_message_token_boundary_preservation(self, mock_async_processor_init, mock_flow, mock_consumer):
flow_mock, output_mock = mock_flow
processor = TokenChunker(chunk_size=10, chunk_overlap=2)
metadata = Metadata(id="boundary", metadata=[], user="test-user", collection="test-collection")
# Text with clear word boundaries
text = "This is a test of token boundaries and proper splitting."
document = TextDocument(metadata=metadata, text=text.encode("utf-8"))
msg = Mock()
msg.value.return_value = document
await processor.on_message(msg, mock_consumer, flow_mock)
# Collect all chunks
chunks = []
for call in output_mock.send.call_args_list:
chunk_text = call[0][0].chunk.decode("utf-8")
chunks.append(chunk_text)
# Token chunker should respect token boundaries
for chunk in chunks:
# Chunks should not start or end with partial words (in most cases)
assert len(chunk.strip()) > 0
@pytest.mark.asyncio
async def test_metrics_recorded(self, mock_async_processor_init, mock_flow, mock_consumer, sample_document):
flow_mock, output_mock = mock_flow
processor = TokenChunker(chunk_size=50)
msg = Mock()
msg.value.return_value = sample_document
# Mock the metric
with patch.object(TokenChunker.chunk_metric, 'labels') as mock_labels:
mock_observe = Mock()
mock_labels.return_value.observe = mock_observe
await processor.on_message(msg, mock_consumer, flow_mock)
# Verify metrics were recorded
mock_labels.assert_called_with(id="test-consumer", flow="test-flow")
assert mock_observe.call_count > 0
# Verify chunk sizes were observed
for call in mock_observe.call_args_list:
chunk_size = call[0][0]
assert chunk_size > 0
def test_add_args(self):
parser = Mock()
TokenChunker.add_args(parser)
# Verify arguments were added
calls = parser.add_argument.call_args_list
arg_names = [call[0][0] for call in calls]
assert '-z' in arg_names or '--chunk-size' in arg_names
assert '-v' in arg_names or '--chunk-overlap' in arg_names
@pytest.mark.asyncio
async def test_encoding_specific_behavior(self, mock_async_processor_init, mock_flow, mock_consumer):
flow_mock, output_mock = mock_flow
processor = TokenChunker(chunk_size=10, chunk_overlap=0)
metadata = Metadata(id="encoding", metadata=[], user="test-user", collection="test-collection")
# Test text that might tokenize differently with cl100k_base encoding
text = "GPT-4 is an AI model. It uses tokens."
document = TextDocument(metadata=metadata, text=text.encode("utf-8"))
msg = Mock()
msg.value.return_value = document
await processor.on_message(msg, mock_consumer, flow_mock)
# Verify chunking happened
assert output_mock.send.call_count >= 1
# Collect all chunks
chunks = []
for call in output_mock.send.call_args_list:
chunk_text = call[0][0].chunk.decode("utf-8")
chunks.append(chunk_text)
# Verify all text is preserved (allowing for overlap)
all_text = " ".join(chunks)
assert "GPT-4" in all_text
assert "AI model" in all_text
assert "tokens" in all_text