- New version of RAG_SPECIALIST and RAG_AGENT, including definition of conversation_purpose and response_depth.
This commit is contained in:
Josako
@@ -0,0 +1,30 @@
|
||||
CHANNEL_ADAPTATION = [
|
||||
{
|
||||
"name": "Mobile Chat/Text",
|
||||
"description": "Short, scannable messages. Uses bullet points, emojis, and concise language.",
|
||||
"when_to_use": "Instant messaging, SMS, or chatbots."
|
||||
},
|
||||
{
|
||||
"name": "Voice/Spoken",
|
||||
"description": "Natural, conversational language. Includes pauses, intonation, and simple sentences.",
|
||||
"when_to_use": "Voice assistants, phone calls, or voice-enabled apps."
|
||||
},
|
||||
{
|
||||
"name": "Email/Formal Text",
|
||||
"description": "Structured, polished, and professional. Uses paragraphs and clear formatting.",
|
||||
"when_to_use": "Emails, reports, or official documentation."
|
||||
},
|
||||
{
|
||||
"name": "Multimodal",
|
||||
"description": "Combines text with visuals, buttons, or interactive elements for clarity and engagement.",
|
||||
"when_to_use": "Websites, apps, or platforms supporting rich media."
|
||||
}
|
||||
]
|
||||
|
||||
def get_channel_adaptation_context(channel_adaptation:str) -> str:
|
||||
selected_channel_adaptation = next(
|
||||
(item for item in CHANNEL_ADAPTATION if item["name"] == channel_adaptation),
|
||||
None
|
||||
)
|
||||
channel_adaptation_context = f"{selected_channel_adaptation['description']}"
|
||||
return channel_adaptation_context
|
||||
@@ -0,0 +1,30 @@
|
||||
CONVERSATION_PURPOSE = [
|
||||
{
|
||||
"name": "Informative",
|
||||
"description": "Focus on sharing facts, explanations, or instructions.",
|
||||
"when_to_use": "User seeks knowledge, clarification, or guidance."
|
||||
},
|
||||
{
|
||||
"name": "Persuasive",
|
||||
"description": "Aim to convince, motivate, or drive action. Highlights benefits and calls to action.",
|
||||
"when_to_use": "Sales, marketing, or when encouraging the user to take a specific step."
|
||||
},
|
||||
{
|
||||
"name": "Supportive",
|
||||
"description": "Empathetic, solution-oriented, and reassuring. Prioritizes the user's needs and emotions.",
|
||||
"when_to_use": "Customer support, healthcare, or emotionally sensitive situations."
|
||||
},
|
||||
{
|
||||
"name": "Collaborative",
|
||||
"description": "Encourages dialogue, brainstorming, and co-creation. Invites user input and ideas.",
|
||||
"when_to_use": "Teamwork, creative processes, or open-ended discussions."
|
||||
}
|
||||
]
|
||||
|
||||
def get_conversation_purpose_context(conversation_purpose:str) -> str:
|
||||
selected_conversation_purpose = next(
|
||||
(item for item in CONVERSATION_PURPOSE if item["name"] == conversation_purpose),
|
||||
None
|
||||
)
|
||||
conversation_purpose_context = f"{selected_conversation_purpose['description']}"
|
||||
return conversation_purpose_context
|
||||
@@ -0,0 +1,25 @@
|
||||
RESPONSE_DEPTH = [
|
||||
{
|
||||
"name": "Concise",
|
||||
"description": "Short, direct answers. No extra explanation or context.",
|
||||
"when_to_use": "Quick queries, urgent requests, or when the user prefers brevity."
|
||||
},
|
||||
{
|
||||
"name": "Balanced",
|
||||
"description": "Clear answers with minimal context or next steps. Not too short, not too detailed.",
|
||||
"when_to_use": "General interactions, FAQs, or when the user needs a mix of speed and clarity."
|
||||
},
|
||||
{
|
||||
"name": "Detailed",
|
||||
"description": "Comprehensive answers with background, examples, or step-by-step guidance.",
|
||||
"when_to_use": "Complex topics, tutorials, or when the user requests in-depth information."
|
||||
}
|
||||
]
|
||||
|
||||
def get_response_depth_context(response_depth:str) -> str:
|
||||
selected_response_depth = next(
|
||||
(item for item in RESPONSE_DEPTH if item["name"] == response_depth),
|
||||
None
|
||||
)
|
||||
response_depth_context = f"{selected_response_depth['description']}"
|
||||
return response_depth_context
|
||||
225
eveai_chat_workers/specialists/globals/RAG_SPECIALIST/1_2.py
Normal file
225
eveai_chat_workers/specialists/globals/RAG_SPECIALIST/1_2.py
Normal file
@@ -0,0 +1,225 @@
|
||||
import json
|
||||
from os import wait
|
||||
from typing import Optional, List, Dict, Any
|
||||
|
||||
from crewai.flow.flow import start, listen, and_
|
||||
from flask import current_app
|
||||
from pydantic import BaseModel, Field
|
||||
|
||||
from common.services.utils.translation_services import TranslationServices
|
||||
from common.utils.business_event_context import current_event
|
||||
from eveai_chat_workers.definitions.conversation_purpose.conversation_purpose_v1_0 import \
|
||||
get_conversation_purpose_context
|
||||
from eveai_chat_workers.definitions.language_level.language_level_v1_0 import get_language_level_context
|
||||
from eveai_chat_workers.definitions.response_depth.response_depth_v1_0 import get_response_depth_context
|
||||
from eveai_chat_workers.definitions.tone_of_voice.tone_of_voice_v1_0 import get_tone_of_voice_context
|
||||
from eveai_chat_workers.retrievers.retriever_typing import RetrieverArguments
|
||||
from eveai_chat_workers.specialists.crewai_base_specialist import CrewAIBaseSpecialistExecutor
|
||||
from eveai_chat_workers.specialists.specialist_typing import SpecialistResult, SpecialistArguments
|
||||
from eveai_chat_workers.outputs.globals.rag.rag_v1_0 import RAGOutput
|
||||
from eveai_chat_workers.specialists.crewai_base_classes import EveAICrewAICrew, EveAICrewAIFlow, EveAIFlowState
|
||||
|
||||
INSUFFICIENT_INFORMATION_MESSAGES = [
|
||||
"I'm afraid I don't have enough information to answer that properly. Feel free to ask something else!",
|
||||
"There isn’t enough data available right now to give you a clear answer. You're welcome to rephrase or ask a different question.",
|
||||
"Sorry, I can't provide a complete answer based on the current information. Would you like to try asking something else?",
|
||||
"I don’t have enough details to give you a confident answer. You can always ask another question if you’d like.",
|
||||
"Unfortunately, I can’t answer that accurately with the information at hand. Please feel free to ask something else.",
|
||||
"That’s a great question, but I currently lack the necessary information to respond properly. Want to ask something different?",
|
||||
"I wish I could help more, but the data I have isn't sufficient to answer this. You’re welcome to explore other questions.",
|
||||
"There’s not enough context for me to provide a good answer. Don’t hesitate to ask another question if you'd like!",
|
||||
"I'm not able to give a definitive answer to that. Perhaps try a different question or angle?",
|
||||
"Thanks for your question. At the moment, I can’t give a solid answer — but I'm here if you want to ask something else!"
|
||||
]
|
||||
|
||||
class SpecialistExecutor(CrewAIBaseSpecialistExecutor):
|
||||
"""
|
||||
type: RAG_SPECIALIST
|
||||
type_version: 1.0
|
||||
RAG Specialist Executor class
|
||||
"""
|
||||
|
||||
def __init__(self, tenant_id, specialist_id, session_id, task_id, **kwargs):
|
||||
self.rag_crew = None
|
||||
|
||||
super().__init__(tenant_id, specialist_id, session_id, task_id)
|
||||
|
||||
@property
|
||||
def type(self) -> str:
|
||||
return "RAG_SPECIALIST"
|
||||
|
||||
@property
|
||||
def type_version(self) -> str:
|
||||
return "1.2"
|
||||
|
||||
def _config_task_agents(self):
|
||||
self._add_task_agent("rag_task", "rag_agent")
|
||||
|
||||
def _config_pydantic_outputs(self):
|
||||
self._add_pydantic_output("rag_task", RAGOutput, "rag_output")
|
||||
|
||||
def _config_state_result_relations(self):
|
||||
self._add_state_result_relation("rag_output")
|
||||
self._add_state_result_relation("citations")
|
||||
|
||||
def _instantiate_specialist(self):
|
||||
verbose = self.tuning
|
||||
|
||||
rag_agents = [self.rag_agent]
|
||||
rag_tasks = [self.rag_task]
|
||||
self.rag_crew = EveAICrewAICrew(
|
||||
self,
|
||||
"Rag Crew",
|
||||
agents=rag_agents,
|
||||
tasks=rag_tasks,
|
||||
verbose=verbose,
|
||||
)
|
||||
|
||||
self.flow = RAGFlow(
|
||||
self,
|
||||
self.rag_crew,
|
||||
)
|
||||
|
||||
def execute(self, arguments: SpecialistArguments, formatted_context, citations) -> SpecialistResult:
|
||||
self.log_tuning("RAG Specialist execution started", {})
|
||||
|
||||
if not self._cached_session.interactions:
|
||||
specialist_phase = "initial"
|
||||
else:
|
||||
specialist_phase = self._cached_session.interactions[-1].specialist_results.get('phase', 'initial')
|
||||
|
||||
results = None
|
||||
|
||||
match specialist_phase:
|
||||
case "initial":
|
||||
results = self.execute_initial_state(arguments, formatted_context, citations)
|
||||
case "rag":
|
||||
results = self.execute_rag_state(arguments, formatted_context, citations)
|
||||
|
||||
self.log_tuning(f"RAG Specialist execution ended", {"Results": results.model_dump()})
|
||||
|
||||
return results
|
||||
|
||||
def execute_initial_state(self, arguments: SpecialistArguments, formatted_context, citations) -> SpecialistResult:
|
||||
self.log_tuning("RAG Specialist initial_state execution started", {})
|
||||
|
||||
welcome_message = self.specialist.configuration.get('welcome_message', 'Welcome! You can start asking questions')
|
||||
welcome_message = TranslationServices.translate(self.tenant_id, welcome_message, arguments.language)
|
||||
|
||||
self.flow.state.answer = welcome_message
|
||||
self.flow.state.phase = "rag"
|
||||
|
||||
results = RAGSpecialistResult.create_for_type(self.type, self.type_version)
|
||||
|
||||
return results
|
||||
|
||||
def execute_rag_state(self, arguments: SpecialistArguments, formatted_context, citations) -> SpecialistResult:
|
||||
self.log_tuning("RAG Specialist rag_state execution started", {})
|
||||
|
||||
insufficient_info_message = TranslationServices.translate(self.tenant_id,
|
||||
INSUFFICIENT_INFORMATION_MESSAGE,
|
||||
arguments.language)
|
||||
|
||||
formatted_context, citations = self._retrieve_context(arguments)
|
||||
self.flow.state.citations = citations
|
||||
tone_of_voice = self.specialist.configuration.get('tone_of_voice', 'Professional & Neutral')
|
||||
tone_of_voice_context = get_tone_of_voice_context(tone_of_voice)
|
||||
language_level = self.specialist.configuration.get('language_level', 'Standard')
|
||||
language_level_context = get_language_level_context(language_level)
|
||||
response_depth = self.specialist.configuration.get('response_depth', 'Balanced')
|
||||
response_depth_context = get_response_depth_context(response_depth)
|
||||
conversation_purpose = self.specialist.configuration.get('conversation_purpose', 'Informative')
|
||||
conversation_purpose_context = get_conversation_purpose_context(conversation_purpose)
|
||||
|
||||
if formatted_context:
|
||||
flow_inputs = {
|
||||
"language": arguments.language,
|
||||
"question": arguments.question,
|
||||
"context": formatted_context,
|
||||
"history": self.formatted_history,
|
||||
"name": self.specialist.configuration.get('name', ''),
|
||||
"welcome_message": self.specialist.configuration.get('welcome_message', ''),
|
||||
"tone_of_voice": tone_of_voice,
|
||||
"tone_of_voice_context": tone_of_voice_context,
|
||||
"language_level": language_level,
|
||||
"language_level_context": language_level_context,
|
||||
"response_depth": response_depth,
|
||||
"response_depth_context": response_depth_context,
|
||||
"conversation_purpose": conversation_purpose,
|
||||
"conversation_purpose_context": conversation_purpose_context,
|
||||
}
|
||||
|
||||
flow_results = self.flow.kickoff(inputs=flow_inputs)
|
||||
|
||||
if flow_results.rag_output.insufficient_info:
|
||||
flow_results.rag_output.answer = insufficient_info_message
|
||||
|
||||
rag_output = flow_results.rag_output
|
||||
else:
|
||||
rag_output = RAGOutput(answer=insufficient_info_message, insufficient_info=True)
|
||||
|
||||
self.flow.state.rag_output = rag_output
|
||||
self.flow.state.citations = citations
|
||||
self.flow.state.answer = rag_output.answer
|
||||
self.flow.state.phase = "rag"
|
||||
|
||||
results = RAGSpecialistResult.create_for_type(self.type, self.type_version)
|
||||
|
||||
return results
|
||||
|
||||
|
||||
class RAGSpecialistInput(BaseModel):
|
||||
language: Optional[str] = Field(None, alias="language")
|
||||
question: Optional[str] = Field(None, alias="question")
|
||||
context: Optional[str] = Field(None, alias="context")
|
||||
history: Optional[str] = Field(None, alias="history")
|
||||
name: Optional[str] = Field(None, alias="name")
|
||||
welcome_message: Optional[str] = Field(None, alias="welcome_message")
|
||||
|
||||
|
||||
class RAGSpecialistResult(SpecialistResult):
|
||||
rag_output: Optional[RAGOutput] = Field(None, alias="Rag Output")
|
||||
|
||||
|
||||
class RAGFlowState(EveAIFlowState):
|
||||
"""Flow state for RAG specialist that automatically updates from task outputs"""
|
||||
input: Optional[RAGSpecialistInput] = None
|
||||
rag_output: Optional[RAGOutput] = None
|
||||
citations: Optional[List[Dict[str, Any]]] = None
|
||||
|
||||
|
||||
class RAGFlow(EveAICrewAIFlow[RAGFlowState]):
|
||||
def __init__(self,
|
||||
specialist_executor: CrewAIBaseSpecialistExecutor,
|
||||
rag_crew: EveAICrewAICrew,
|
||||
**kwargs):
|
||||
super().__init__(specialist_executor, "RAG Specialist Flow", **kwargs)
|
||||
self.specialist_executor = specialist_executor
|
||||
self.rag_crew = rag_crew
|
||||
self.exception_raised = False
|
||||
|
||||
@start()
|
||||
def process_inputs(self):
|
||||
return ""
|
||||
|
||||
@listen(process_inputs)
|
||||
async def execute_rag(self):
|
||||
inputs = self.state.input.model_dump()
|
||||
try:
|
||||
crew_output = await self.rag_crew.kickoff_async(inputs=inputs)
|
||||
self.specialist_executor.log_tuning("RAG Crew Output", crew_output.model_dump())
|
||||
output_pydantic = crew_output.pydantic
|
||||
if not output_pydantic:
|
||||
raw_json = json.loads(crew_output.raw)
|
||||
output_pydantic = RAGOutput.model_validate(raw_json)
|
||||
self.state.rag_output = output_pydantic
|
||||
return crew_output
|
||||
except Exception as e:
|
||||
current_app.logger.error(f"CREW rag_crew Kickoff Error: {str(e)}")
|
||||
self.exception_raised = True
|
||||
raise e
|
||||
|
||||
async def kickoff_async(self, inputs=None):
|
||||
self.state.input = RAGSpecialistInput.model_validate(inputs)
|
||||
result = await super().kickoff_async(inputs)
|
||||
return self.state
|
||||
Reference in New Issue
Block a user