ce9ef3ca79
* Add support for non-ASCII characters in LLM prompts - Add ensure_ascii parameter to Graphiti class (default: True) - Create to_prompt_json helper function for consistent JSON serialization - Update all prompt files to use new helper function - Preserve Korean/Japanese/Chinese characters when ensure_ascii=False - Maintain backward compatibility with existing behavior Fixes issue where non-ASCII characters were escaped as unicode sequences in prompts, making them unreadable in LLM logs and potentially affecting model understanding. * Remove unused json imports after replacing with to_prompt_json helper - Fix ruff lint errors (F401) for unused json imports - All prompt files now use to_prompt_json helper instead of json.dumps - Maintains clean code style and passes lint checks * Fix ensure_ascii propagation to all LLM calls - Add ensure_ascii parameter to maintenance operation functions that were missing it - Update function signatures in node_operations, community_operations, temporal_operations, and edge_operations - Ensure all llm_client.generate_response calls receive proper ensure_ascii context - Fix hardcoded ensure_ascii: True values that prevented non-ASCII character preservation - Maintain backward compatibility with default ensure_ascii=True - Complete the fix for issue #804 ensuring Korean/Japanese/Chinese characters are properly handled in LLM prompts
133 lines
4.3 KiB
Python
133 lines
4.3 KiB
Python
"""
|
|
Copyright 2024, Zep Software, Inc.
|
|
|
|
Licensed under the Apache License, Version 2.0 (the "License");
|
|
you may not use this file except in compliance with the License.
|
|
You may obtain a copy of the License at
|
|
|
|
http://www.apache.org/licenses/LICENSE-2.0
|
|
|
|
Unless required by applicable law or agreed to in writing, software
|
|
distributed under the License is distributed on an "AS IS" BASIS,
|
|
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
See the License for the specific language governing permissions and
|
|
limitations under the License.
|
|
"""
|
|
|
|
from typing import Any, Protocol, TypedDict
|
|
|
|
from pydantic import BaseModel, Field
|
|
|
|
from .models import Message, PromptFunction, PromptVersion
|
|
from .prompt_helpers import to_prompt_json
|
|
|
|
|
|
class Summary(BaseModel):
|
|
summary: str = Field(
|
|
...,
|
|
description='Summary containing the important information about the entity. Under 250 words',
|
|
)
|
|
|
|
|
|
class SummaryDescription(BaseModel):
|
|
description: str = Field(..., description='One sentence description of the provided summary')
|
|
|
|
|
|
class Prompt(Protocol):
|
|
summarize_pair: PromptVersion
|
|
summarize_context: PromptVersion
|
|
summary_description: PromptVersion
|
|
|
|
|
|
class Versions(TypedDict):
|
|
summarize_pair: PromptFunction
|
|
summarize_context: PromptFunction
|
|
summary_description: PromptFunction
|
|
|
|
|
|
def summarize_pair(context: dict[str, Any]) -> list[Message]:
|
|
return [
|
|
Message(
|
|
role='system',
|
|
content='You are a helpful assistant that combines summaries.',
|
|
),
|
|
Message(
|
|
role='user',
|
|
content=f"""
|
|
Synthesize the information from the following two summaries into a single succinct summary.
|
|
|
|
Summaries must be under 250 words.
|
|
|
|
Summaries:
|
|
{to_prompt_json(context['node_summaries'], ensure_ascii=context.get('ensure_ascii', True), indent=2)}
|
|
""",
|
|
),
|
|
]
|
|
|
|
|
|
def summarize_context(context: dict[str, Any]) -> list[Message]:
|
|
return [
|
|
Message(
|
|
role='system',
|
|
content='You are a helpful assistant that extracts entity properties from the provided text.',
|
|
),
|
|
Message(
|
|
role='user',
|
|
content=f"""
|
|
|
|
<MESSAGES>
|
|
{to_prompt_json(context['previous_episodes'], ensure_ascii=context.get('ensure_ascii', True), indent=2)}
|
|
{to_prompt_json(context['episode_content'], ensure_ascii=context.get('ensure_ascii', True), indent=2)}
|
|
</MESSAGES>
|
|
|
|
Given the above MESSAGES and the following ENTITY name, create a summary for the ENTITY. Your summary must only use
|
|
information from the provided MESSAGES. Your summary should also only contain information relevant to the
|
|
provided ENTITY. Summaries must be under 250 words.
|
|
|
|
In addition, extract any values for the provided entity properties based on their descriptions.
|
|
If the value of the entity property cannot be found in the current context, set the value of the property to the Python value None.
|
|
|
|
Guidelines:
|
|
1. Do not hallucinate entity property values if they cannot be found in the current context.
|
|
2. Only use the provided messages, entity, and entity context to set attribute values.
|
|
|
|
<ENTITY>
|
|
{context['node_name']}
|
|
</ENTITY>
|
|
|
|
<ENTITY CONTEXT>
|
|
{context['node_summary']}
|
|
</ENTITY CONTEXT>
|
|
|
|
<ATTRIBUTES>
|
|
{to_prompt_json(context['attributes'], ensure_ascii=context.get('ensure_ascii', True), indent=2)}
|
|
</ATTRIBUTES>
|
|
""",
|
|
),
|
|
]
|
|
|
|
|
|
def summary_description(context: dict[str, Any]) -> list[Message]:
|
|
return [
|
|
Message(
|
|
role='system',
|
|
content='You are a helpful assistant that describes provided contents in a single sentence.',
|
|
),
|
|
Message(
|
|
role='user',
|
|
content=f"""
|
|
Create a short one sentence description of the summary that explains what kind of information is summarized.
|
|
Summaries must be under 250 words.
|
|
|
|
Summary:
|
|
{to_prompt_json(context['summary'], ensure_ascii=context.get('ensure_ascii', True), indent=2)}
|
|
""",
|
|
),
|
|
]
|
|
|
|
|
|
versions: Versions = {
|
|
'summarize_pair': summarize_pair,
|
|
'summarize_context': summarize_context,
|
|
'summary_description': summary_description,
|
|
}
|