wx44wx/pipecat
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Deprecate expect_stripped_words option from LLMAssistantAggregatorParams, when used with the newer LLMAssistantAggregator, which now handles word spacing automatically.

Browse Source 
This commit does not change how it works in the older `LLMAssistantContextAggregator`.
This commit is contained in:
Paul Kompfner
2025-10-30 15:48:15 -04:00
parent 5e00133e64
commit ac5734d0ed
18 changed files with 130 additions and 133 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
CHANGELOG.md
View File

@@ -33,13 +33,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
```python ```python
context = LLMContext(messages, tools) context = LLMContext(messages, tools)
context_aggregator = LLMContextAggregatorPair( context_aggregator = LLMContextAggregatorPair(context)
context,
# This part is `OpenAIRealtimeLLMService`-specific.
# `expect_stripped_words=False` needed when OpenAI Realtime used with
# "audio" modality (the default).
assistant_params=LLMAssistantAggregatorParams(expect_stripped_words=False),
)
``` ```
(Note that even though `OpenAIRealtimeLLMService` now supports the universal (Note that even though `OpenAIRealtimeLLMService` now supports the universal
@@ -116,13 +110,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
```python ```python
context = LLMContext(messages, tools) context = LLMContext(messages, tools)
context_aggregator = LLMContextAggregatorPair( context_aggregator = LLMContextAggregatorPair(context)
context,
# This part is `GeminiLiveLLMService`-specific.
# `expect_stripped_words=False` needed when Gemini Live used with AUDIO
# modality (the default).
assistant_params=LLMAssistantAggregatorParams(expect_stripped_words=False),
)
``` ```
(Note that even though `GeminiLiveLLMService` now supports the universal (Note that even though `GeminiLiveLLMService` now supports the universal
@@ -202,6 +190,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
### Deprecated ### Deprecated
- The `expect_stripped_words` parameter of `LLMAssistantAggregatorParams` is
ignored when used with the newer `LLMAssistantAggregator`, which now handles
word spacing automatically.
- `LLMService.request_image_frame()` is deprecated, push a - `LLMService.request_image_frame()` is deprecated, push a
`UserImageRequestFrame` instead. `UserImageRequestFrame` instead.

7
examples/foundational/19-openai-realtime.py
View File

@@ -187,12 +187,7 @@ Remember, your responses should be short. Just one or two sentences, usually. Re
tools, tools,
) )
context_aggregator = LLMContextAggregatorPair( context_aggregator = LLMContextAggregatorPair(context)
context,
# `expect_stripped_words=False` needed when OpenAI Realtime used with
# "audio" modality (the default)
assistant_params=LLMAssistantAggregatorParams(expect_stripped_words=False),
)
pipeline = Pipeline( pipeline = Pipeline(
[ [

7
examples/foundational/19a-azure-realtime.py
View File

@@ -175,12 +175,7 @@ Remember, your responses should be short. Just one or two sentences, usually. Re
tools, tools,
) )
context_aggregator = LLMContextAggregatorPair( context_aggregator = LLMContextAggregatorPair(context)
context,
# `expect_stripped_words=False` needed when OpenAI Realtime used with
# "audio" modality (the default)
assistant_params=LLMAssistantAggregatorParams(expect_stripped_words=False),
)
pipeline = Pipeline( pipeline = Pipeline(
[ [

7
examples/foundational/26a-gemini-live-transcription.py
View File

@@ -92,12 +92,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
# }, # },
], ],
) )
context_aggregator = LLMContextAggregatorPair( context_aggregator = LLMContextAggregatorPair(context)
context,
# `expect_stripped_words=False` needed when Gemini Live used with AUDIO
# modality (the default)
assistant_params=LLMAssistantAggregatorParams(expect_stripped_words=False),
)
transcript = TranscriptProcessor() transcript = TranscriptProcessor()

7
examples/foundational/26b-gemini-live-function-calling.py
View File

@@ -144,12 +144,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
context = LLMContext( context = LLMContext(
[{"role": "user", "content": "Say hello."}], [{"role": "user", "content": "Say hello."}],
) )
context_aggregator = LLMContextAggregatorPair( context_aggregator = LLMContextAggregatorPair(context)
context,
# `expect_stripped_words=False` needed when Gemini Live used with AUDIO
# modality (the default)
assistant_params=LLMAssistantAggregatorParams(expect_stripped_words=False),
)
pipeline = Pipeline( pipeline = Pipeline(
[ [

7
examples/foundational/26c-gemini-live-video.py
View File

@@ -75,12 +75,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
}, },
], ],
) )
context_aggregator = LLMContextAggregatorPair( context_aggregator = LLMContextAggregatorPair(context)
context,
# `expect_stripped_words=False` needed when Gemini Live used with AUDIO
# modality (the default)
assistant_params=LLMAssistantAggregatorParams(expect_stripped_words=False),
)
pipeline = Pipeline( pipeline = Pipeline(
[ [

7
examples/foundational/26e-gemini-live-google-search.py
View File

@@ -100,12 +100,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
} }
], ],
) )
context_aggregator = LLMContextAggregatorPair( context_aggregator = LLMContextAggregatorPair(context)
context,
# `expect_stripped_words=False` needed when Gemini Live used with AUDIO
# modality (the default)
assistant_params=LLMAssistantAggregatorParams(expect_stripped_words=False),
)
pipeline = Pipeline( pipeline = Pipeline(
[ [

7
examples/foundational/26f-gemini-live-files-api.py
View File

@@ -164,12 +164,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
) )
# Create context aggregator # Create context aggregator
context_aggregator = LLMContextAggregatorPair( context_aggregator = LLMContextAggregatorPair(context)
context,
# `expect_stripped_words=False` needed when Gemini Live used with AUDIO
# modality (the default)
assistant_params=LLMAssistantAggregatorParams(expect_stripped_words=False),
)
# Build the pipeline # Build the pipeline
pipeline = Pipeline( pipeline = Pipeline(

7
examples/foundational/26g-gemini-live-groundingMetadata.py
View File

@@ -127,12 +127,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
# Set up conversation context and management # Set up conversation context and management
context = LLMContext(messages) context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair( context_aggregator = LLMContextAggregatorPair(context)
context,
# `expect_stripped_words=False` needed when Gemini Live used with AUDIO
# modality (the default)
assistant_params=LLMAssistantAggregatorParams(expect_stripped_words=False),
)
pipeline = Pipeline( pipeline = Pipeline(
[ [

7
examples/foundational/26h-gemini-live-vertex-function-calling.py
View File

@@ -140,12 +140,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
llm.register_function("get_restaurant_recommendation", fetch_restaurant_recommendation) llm.register_function("get_restaurant_recommendation", fetch_restaurant_recommendation)
context = LLMContext([{"role": "user", "content": "Say hello."}]) context = LLMContext([{"role": "user", "content": "Say hello."}])
context_aggregator = LLMContextAggregatorPair( context_aggregator = LLMContextAggregatorPair(context)
context,
# `expect_stripped_words=False` needed when Gemini Live used with AUDIO
# modality (the default)
assistant_params=LLMAssistantAggregatorParams(expect_stripped_words=False),
)
pipeline = Pipeline( pipeline = Pipeline(
[ [

7
examples/foundational/26i-gemini-live-graceful-end.py
View File

@@ -157,12 +157,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
context = LLMContext( context = LLMContext(
[{"role": "user", "content": "Say hello."}], [{"role": "user", "content": "Say hello."}],
) )
context_aggregator = LLMContextAggregatorPair( context_aggregator = LLMContextAggregatorPair(context)
context,
# `expect_stripped_words=False` needed when Gemini Live used with AUDIO
# modality (the default)
assistant_params=LLMAssistantAggregatorParams(expect_stripped_words=False),
)
pipeline = Pipeline( pipeline = Pipeline(
[ [

7
examples/foundational/46-video-processing.py
View File

@@ -111,12 +111,7 @@ async def run_bot(pipecat_transport):
] ]
context = LLMContext(messages) context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair( context_aggregator = LLMContextAggregatorPair(context)
context,
# `expect_stripped_words=False` needed when Gemini Live used with AUDIO
# modality (the default)
assistant_params=LLMAssistantAggregatorParams(expect_stripped_words=False),
)
# RTVI events for Pipecat client UI # RTVI events for Pipecat client UI
rtvi = RTVIProcessor() rtvi = RTVIProcessor()

4
src/pipecat/processors/aggregators/llm_response.py
View File

@@ -89,7 +89,9 @@ class LLMAssistantAggregatorParams:
Parameters: Parameters:
expect_stripped_words: Whether to expect and handle stripped words expect_stripped_words: Whether to expect and handle stripped words
in text frames by adding spaces between tokens. in text frames by adding spaces between tokens. This parameter is
ignored when used with the newer LLMAssistantAggregator, which
handles word spacing automatically.
""" """
expect_stripped_words: bool = True expect_stripped_words: bool = True

49
src/pipecat/processors/aggregators/llm_response_universal.py
View File

@@ -13,6 +13,7 @@ LLM processing, and text-to-speech components in conversational AI pipelines.
import asyncio import asyncio
import json import json
import warnings
from abc import abstractmethod from abc import abstractmethod
from typing import Any, Dict, List, Literal, Optional, Set from typing import Any, Dict, List, Literal, Optional, Set
@@ -65,6 +66,7 @@ from pipecat.processors.aggregators.llm_response import (
LLMUserAggregatorParams, LLMUserAggregatorParams,
) )
from pipecat.processors.frame_processor import FrameDirection, FrameProcessor from pipecat.processors.frame_processor import FrameDirection, FrameProcessor
from pipecat.utils.string import concatenate_aggregated_text
from pipecat.utils.time import time_now_iso8601 from pipecat.utils.time import time_now_iso8601
@@ -88,7 +90,7 @@ class LLMContextAggregator(FrameProcessor):
self._context = context self._context = context
self._role = role self._role = role
self._aggregation: str = "" self._aggregation: List[str] = []
@property @property
def messages(self) -> List[LLMContextMessage]: def messages(self) -> List[LLMContextMessage]:
@@ -168,13 +170,21 @@ class LLMContextAggregator(FrameProcessor):
async def reset(self): async def reset(self):
"""Reset the aggregation state.""" """Reset the aggregation state."""
self._aggregation = "" self._aggregation = []
@abstractmethod @abstractmethod
async def push_aggregation(self): async def push_aggregation(self):
"""Push the current aggregation downstream.""" """Push the current aggregation downstream."""
pass pass
def aggregation_string(self) -> str:
"""Get the current aggregation as a string.
Returns:
The concatenated aggregation string.
"""
return concatenate_aggregated_text(self._aggregation)
class LLMUserAggregator(LLMContextAggregator): class LLMUserAggregator(LLMContextAggregator):
"""User LLM aggregator that processes speech-to-text transcriptions. """User LLM aggregator that processes speech-to-text transcriptions.
@@ -212,8 +222,6 @@ class LLMUserAggregator(LLMContextAggregator):
self._turn_params: Optional[SmartTurnParams] = None self._turn_params: Optional[SmartTurnParams] = None
if "aggregation_timeout" in kwargs: if "aggregation_timeout" in kwargs:
import warnings
with warnings.catch_warnings(): with warnings.catch_warnings():
warnings.simplefilter("always") warnings.simplefilter("always")
warnings.warn( warnings.warn(
@@ -307,7 +315,7 @@ class LLMUserAggregator(LLMContextAggregator):
async def _process_aggregation(self): async def _process_aggregation(self):
"""Process the current aggregation and push it downstream.""" """Process the current aggregation and push it downstream."""
aggregation = self._aggregation aggregation = self.aggregation_string()
await self.reset() await self.reset()
self._context.add_message({"role": self.role, "content": aggregation}) self._context.add_message({"role": self.role, "content": aggregation})
frame = LLMContextFrame(self._context) frame = LLMContextFrame(self._context)
@@ -355,7 +363,7 @@ class LLMUserAggregator(LLMContextAggregator):
""" """
async def should_interrupt(strategy: BaseInterruptionStrategy): async def should_interrupt(strategy: BaseInterruptionStrategy):
await strategy.append_text(self._aggregation) await strategy.append_text(self.aggregation_string())
return await strategy.should_interrupt() return await strategy.should_interrupt()
return any([await should_interrupt(s) for s in self._interruption_strategies]) return any([await should_interrupt(s) for s in self._interruption_strategies])
@@ -425,7 +433,7 @@ class LLMUserAggregator(LLMContextAggregator):
if not text.strip(): if not text.strip():
return return
self._aggregation += f" {text}" if self._aggregation else text self._aggregation.append(text)
# We just got a final result, so let's reset interim results. # We just got a final result, so let's reset interim results.
self._seen_interim_results = False self._seen_interim_results = False
# Reset aggregation timer. # Reset aggregation timer.
@@ -550,23 +558,31 @@ class LLMAssistantAggregator(LLMContextAggregator):
Args: Args:
context: The OpenAI LLM context for conversation storage. context: The OpenAI LLM context for conversation storage.
params: Configuration parameters for aggregation behavior. params: Configuration parameters for aggregation behavior.
**kwargs: Additional arguments. Supports deprecated 'expect_stripped_words'. **kwargs: Additional arguments.
""" """
super().__init__(context=context, role="assistant", **kwargs) super().__init__(context=context, role="assistant", **kwargs)
self._params = params or LLMAssistantAggregatorParams() self._params = params or LLMAssistantAggregatorParams()
if "expect_stripped_words" in kwargs: if "expect_stripped_words" in kwargs:
import warnings
with warnings.catch_warnings(): with warnings.catch_warnings():
warnings.simplefilter("always") warnings.simplefilter("always")
warnings.warn( warnings.warn(
"Parameter 'expect_stripped_words' is deprecated, use 'params' instead.", "Parameter 'expect_stripped_words' is deprecated. "
"LLMAssistantAggregator now handles word spacing automatically.",
DeprecationWarning, DeprecationWarning,
) )
self._params.expect_stripped_words = kwargs["expect_stripped_words"] self._params.expect_stripped_words = kwargs["expect_stripped_words"]
if params and not params.expect_stripped_words:
with warnings.catch_warnings():
warnings.simplefilter("always")
warnings.warn(
"params.expect_stripped_words is deprecated. "
"LLMAssistantAggregator now handles word spacing automatically.",
DeprecationWarning,
)
self._started = 0 self._started = 0
self._function_calls_in_progress: Dict[str, Optional[FunctionCallInProgressFrame]] = {} self._function_calls_in_progress: Dict[str, Optional[FunctionCallInProgressFrame]] = {}
self._context_updated_tasks: Set[asyncio.Task] = set() self._context_updated_tasks: Set[asyncio.Task] = set()
@@ -629,7 +645,7 @@ class LLMAssistantAggregator(LLMContextAggregator):
if not self._aggregation: if not self._aggregation:
return return
aggregation = self._aggregation.strip() aggregation = self.aggregation_string()
await self.reset() await self.reset()
if aggregation: if aggregation:
@@ -793,10 +809,11 @@ class LLMAssistantAggregator(LLMContextAggregator):
if not self._started: if not self._started:
return return
if self._params.expect_stripped_words: # Make sure we really have text (spaces count, too!)
self._aggregation += f" {frame.text}" if self._aggregation else frame.text if len(frame.text) == 0:
else: return
self._aggregation += frame.text
self._aggregation.append(frame.text)
def _context_updated_task_finished(self, task: asyncio.Task): def _context_updated_task_finished(self, task: asyncio.Task):
self._context_updated_tasks.discard(task) self._context_updated_tasks.discard(task)

25
src/pipecat/processors/transcript_processor.py
View File

@@ -26,6 +26,7 @@ from pipecat.frames.frames import (
TTSTextFrame, TTSTextFrame,
) )
from pipecat.processors.frame_processor import FrameDirection, FrameProcessor from pipecat.processors.frame_processor import FrameDirection, FrameProcessor
from pipecat.utils.string import concatenate_aggregated_text
from pipecat.utils.time import time_now_iso8601 from pipecat.utils.time import time_now_iso8601
@@ -140,29 +141,7 @@ class AssistantTranscriptProcessor(BaseTranscriptProcessor):
Result: "Hello there how are you" Result: "Hello there how are you"
""" """
if self._current_text_parts and self._aggregation_start_time: if self._current_text_parts and self._aggregation_start_time:
# Check specifically for space characters, previously isspace() was used content = concatenate_aggregated_text(self._current_text_parts)
# but that includes all whitespace characters (e.g. \n), not just spaces.
has_leading_spaces = any(
part and part[0] == " " for part in self._current_text_parts[1:]
)
has_trailing_spaces = any(
part and part[-1] == " " for part in self._current_text_parts[:-1]
)
# If there are embedded spaces in the fragments, use direct concatenation
contains_spacing_between_fragments = has_leading_spaces or has_trailing_spaces
# Apply corresponding joining method
if contains_spacing_between_fragments:
# Fragments already have spacing - just concatenate
content = "".join(self._current_text_parts)
else:
# Word-by-word fragments - join with spaces
content = " ".join(self._current_text_parts)
# Clean up any excessive whitespace
content = content.strip()
if content: if content:
logger.trace(f"Emitting aggregated assistant message: {content}") logger.trace(f"Emitting aggregated assistant message: {content}")
message = TranscriptionMessage( message = TranscriptionMessage(

39
src/pipecat/utils/string.py
View File

@@ -18,7 +18,7 @@ Dependencies:
""" """
import re import re
from typing import FrozenSet, Optional, Sequence, Tuple from typing import FrozenSet, List, Optional, Sequence, Tuple
import nltk import nltk
from loguru import logger from loguru import logger
@@ -196,3 +196,40 @@ def parse_start_end_tags(
return (None, len(text)) return (None, len(text))
return (None, current_tag_index) return (None, current_tag_index)
def concatenate_aggregated_text(text_parts: List[str]) -> str:
"""Concatenate a list of text parts into a single string.
This function joins the provided list of text parts into a single string,
taking into account whether or not the parts already contain spacing.
This function is useful for aggregating text segments received from LLMs or
transcription services.
Args:
text_parts: A list of strings representing parts of text to concatenate.
Returns:
A single concatenated string.
"""
# Check specifically for space characters, previously isspace() was used
# but that includes all whitespace characters (e.g. \n), not just spaces.
has_leading_spaces = any(part and part[0] == " " for part in text_parts[1:])
has_trailing_spaces = any(part and part[-1] == " " for part in text_parts[:-1])
# If there are embedded spaces in the fragments, use direct concatenation
contains_spacing_between_fragments = has_leading_spaces or has_trailing_spaces
# Apply corresponding joining method
if contains_spacing_between_fragments:
# Fragments already have spacing - just concatenate
result = "".join(text_parts)
else:
# Word-by-word fragments - join with spaces
result = " ".join(text_parts)
# Clean up any excessive whitespace
result = result.strip()
return result

45
tests/test_context_aggregators.py
View File

@@ -6,7 +6,7 @@
import json import json
import unittest import unittest
from typing import Any from typing import Any, Optional
from pipecat.audio.interruptions.min_words_interruption_strategy import MinWordsInterruptionStrategy from pipecat.audio.interruptions.min_words_interruption_strategy import MinWordsInterruptionStrategy
from pipecat.audio.turn.smart_turn.base_smart_turn import SmartTurnParams from pipecat.audio.turn.smart_turn.base_smart_turn import SmartTurnParams
@@ -22,6 +22,8 @@ from pipecat.frames.frames import (
InterimTranscriptionFrame, InterimTranscriptionFrame,
InterruptionFrame, InterruptionFrame,
InterruptionTaskFrame, InterruptionTaskFrame,
LLMContextAssistantTimestampFrame,
LLMContextFrame,
LLMFullResponseEndFrame, LLMFullResponseEndFrame,
LLMFullResponseStartFrame, LLMFullResponseStartFrame,
OpenAILLMContextAssistantTimestampFrame, OpenAILLMContextAssistantTimestampFrame,
@@ -38,6 +40,7 @@ from pipecat.processors.aggregators.llm_response import (
LLMUserAggregatorParams, LLMUserAggregatorParams,
LLMUserContextAggregator, LLMUserContextAggregator,
) )
from pipecat.processors.aggregators.llm_response_universal import LLMAssistantAggregator
from pipecat.processors.aggregators.openai_llm_context import ( from pipecat.processors.aggregators.openai_llm_context import (
OpenAILLMContext, OpenAILLMContext,
OpenAILLMContextFrame, OpenAILLMContextFrame,
@@ -586,11 +589,16 @@ class BaseTestUserContextAggregator:
assert context_processor.context_received assert context_processor.context_received
class BaseTestAssistantContextAggreagator: class BaseTestAssistantContextAggregator:
CONTEXT_CLASS = None # To be set in subclasses CONTEXT_CLASS = None # To be set in subclasses
AGGREGATOR_CLASS = None # To be set in subclasses AGGREGATOR_CLASS = None # To be set in subclasses
EXPECTED_CONTEXT_FRAMES = None # To be set in subclasses EXPECTED_CONTEXT_FRAMES = None # To be set in subclasses
def create_assistant_aggregator_params(
self, **kwargs
) -> Optional[LLMAssistantAggregatorParams]:
return LLMAssistantAggregatorParams(**kwargs)
def check_message_content(self, context: OpenAILLMContext, index: int, content: str): def check_message_content(self, context: OpenAILLMContext, index: int, content: str):
assert context.messages[index]["content"] == content assert context.messages[index]["content"] == content
@@ -641,7 +649,7 @@ class BaseTestAssistantContextAggreagator:
context = self.CONTEXT_CLASS() context = self.CONTEXT_CLASS()
aggregator = self.AGGREGATOR_CLASS( aggregator = self.AGGREGATOR_CLASS(
context, params=LLMAssistantAggregatorParams(expect_stripped_words=False) context, params=self.create_assistant_aggregator_params(expect_stripped_words=False)
) )
frames_to_send = [ frames_to_send = [
LLMFullResponseStartFrame(), LLMFullResponseStartFrame(),
@@ -687,7 +695,7 @@ class BaseTestAssistantContextAggreagator:
context = self.CONTEXT_CLASS() context = self.CONTEXT_CLASS()
aggregator = self.AGGREGATOR_CLASS( aggregator = self.AGGREGATOR_CLASS(
context, params=LLMAssistantAggregatorParams(expect_stripped_words=False) context, params=self.create_assistant_aggregator_params(expect_stripped_words=False)
) )
frames_to_send = [ frames_to_send = [
LLMFullResponseStartFrame(), LLMFullResponseStartFrame(),
@@ -714,7 +722,7 @@ class BaseTestAssistantContextAggreagator:
context = self.CONTEXT_CLASS() context = self.CONTEXT_CLASS()
aggregator = self.AGGREGATOR_CLASS( aggregator = self.AGGREGATOR_CLASS(
context, params=LLMAssistantAggregatorParams(expect_stripped_words=False) context, params=self.create_assistant_aggregator_params(expect_stripped_words=False)
) )
frames_to_send = [ frames_to_send = [
LLMFullResponseStartFrame(), LLMFullResponseStartFrame(),
@@ -838,7 +846,7 @@ class TestAnthropicUserContextAggregator(
class TestAnthropicAssistantContextAggregator( class TestAnthropicAssistantContextAggregator(
BaseTestAssistantContextAggreagator, unittest.IsolatedAsyncioTestCase BaseTestAssistantContextAggregator, unittest.IsolatedAsyncioTestCase
): ):
CONTEXT_CLASS = AnthropicLLMContext CONTEXT_CLASS = AnthropicLLMContext
AGGREGATOR_CLASS = AnthropicAssistantContextAggregator AGGREGATOR_CLASS = AnthropicAssistantContextAggregator
@@ -873,7 +881,7 @@ class TestAWSBedrockUserContextAggregator(
class TestAWSBedrockAssistantContextAggregator( class TestAWSBedrockAssistantContextAggregator(
BaseTestAssistantContextAggreagator, unittest.IsolatedAsyncioTestCase BaseTestAssistantContextAggregator, unittest.IsolatedAsyncioTestCase
): ):
CONTEXT_CLASS = AWSBedrockLLMContext CONTEXT_CLASS = AWSBedrockLLMContext
AGGREGATOR_CLASS = AWSBedrockAssistantContextAggregator AGGREGATOR_CLASS = AWSBedrockAssistantContextAggregator
@@ -914,7 +922,7 @@ class TestGoogleUserContextAggregator(
class TestGoogleAssistantContextAggregator( class TestGoogleAssistantContextAggregator(
BaseTestAssistantContextAggreagator, unittest.IsolatedAsyncioTestCase BaseTestAssistantContextAggregator, unittest.IsolatedAsyncioTestCase
): ):
CONTEXT_CLASS = GoogleLLMContext CONTEXT_CLASS = GoogleLLMContext
AGGREGATOR_CLASS = GoogleAssistantContextAggregator AGGREGATOR_CLASS = GoogleAssistantContextAggregator
@@ -948,8 +956,27 @@ class TestOpenAIUserContextAggregator(
class TestOpenAIAssistantContextAggregator( class TestOpenAIAssistantContextAggregator(
BaseTestAssistantContextAggreagator, unittest.IsolatedAsyncioTestCase BaseTestAssistantContextAggregator, unittest.IsolatedAsyncioTestCase
): ):
CONTEXT_CLASS = OpenAILLMContext CONTEXT_CLASS = OpenAILLMContext
AGGREGATOR_CLASS = OpenAIAssistantContextAggregator AGGREGATOR_CLASS = OpenAIAssistantContextAggregator
EXPECTED_CONTEXT_FRAMES = [OpenAILLMContextFrame, OpenAILLMContextAssistantTimestampFrame] EXPECTED_CONTEXT_FRAMES = [OpenAILLMContextFrame, OpenAILLMContextAssistantTimestampFrame]
#
# Universal
#
class TestLLMAssistantAggregator(
BaseTestAssistantContextAggregator, unittest.IsolatedAsyncioTestCase
):
CONTEXT_CLASS = OpenAILLMContext
AGGREGATOR_CLASS = LLMAssistantAggregator
EXPECTED_CONTEXT_FRAMES = [LLMContextFrame, LLMContextAssistantTimestampFrame]
# Override to remove 'expect_stripped_words' parameter, which is deprecated
# for LLMAssistantAggregator
def create_assistant_aggregator_params(
self, **kwargs
) -> Optional[LLMAssistantAggregatorParams]:
kwargs.pop("expect_stripped_words", None)
return LLMAssistantAggregatorParams(**kwargs) if kwargs else None

4
tests/test_langchain.py
View File

@@ -65,9 +65,7 @@ class TestLangchain(unittest.IsolatedAsyncioTestCase):
self.mock_proc = self.MockProcessor("token_collector") self.mock_proc = self.MockProcessor("token_collector")
context = LLMContext() context = LLMContext()
context_aggregator = LLMContextAggregatorPair( context_aggregator = LLMContextAggregatorPair(context)
context, assistant_params=LLMAssistantAggregatorParams(expect_stripped_words=False)
)
pipeline = Pipeline( pipeline = Pipeline(
[context_aggregator.user(), proc, self.mock_proc, context_aggregator.assistant()] [context_aggregator.user(), proc, self.mock_proc, context_aggregator.assistant()]