493 lines
21 KiB
Python
493 lines
21 KiB
Python
import logging
|
|
import httpx
|
|
from typing import List
|
|
from fastapi import HTTPException
|
|
from modules.shared.configuration import APP_CONFIG
|
|
from modules.aicore.aicoreBase import BaseConnectorAi
|
|
from modules.datamodels.datamodelAi import AiModel, PriorityEnum, ProcessingModeEnum, OperationTypeEnum, AiModelCall, AiModelResponse, createOperationTypeRatings
|
|
|
|
# Configure logger
|
|
logger = logging.getLogger(__name__)
|
|
|
|
def loadConfigData():
|
|
"""Load configuration data for Perplexity connector"""
|
|
return {
|
|
"apiKey": APP_CONFIG.get('Connector_AiPerplexity_API_SECRET'),
|
|
}
|
|
|
|
class AiPerplexity(BaseConnectorAi):
|
|
"""Connector for communication with the Perplexity API."""
|
|
|
|
def __init__(self):
|
|
super().__init__()
|
|
# Load configuration
|
|
self.config = loadConfigData()
|
|
self.apiKey = self.config["apiKey"]
|
|
|
|
# HttpClient for API calls
|
|
self.httpClient = httpx.AsyncClient(
|
|
timeout=120.0, # Longer timeout for complex requests
|
|
headers={
|
|
"Authorization": f"Bearer {self.apiKey}",
|
|
"Content-Type": "application/json",
|
|
"Accept": "application/json"
|
|
}
|
|
)
|
|
|
|
logger.info("Perplexity Connector initialized")
|
|
|
|
def getConnectorType(self) -> str:
|
|
"""Get the connector type identifier."""
|
|
return "perplexity"
|
|
|
|
def getModels(self) -> List[AiModel]:
|
|
"""Get all available Perplexity models."""
|
|
return [
|
|
AiModel(
|
|
name="llama-3.1-sonar-large-128k-online",
|
|
displayName="Perplexity Llama 3.1 Sonar Large 128k",
|
|
connectorType="perplexity",
|
|
apiUrl="https://api.perplexity.ai/chat/completions",
|
|
temperature=0.2,
|
|
maxTokens=128000,
|
|
contextLength=128000,
|
|
costPer1kTokensInput=0.005,
|
|
costPer1kTokensOutput=0.005,
|
|
speedRating=8,
|
|
qualityRating=8,
|
|
# capabilities removed (not used in business logic)
|
|
functionCall=self.callAiBasic,
|
|
priority=PriorityEnum.BALANCED,
|
|
processingMode=ProcessingModeEnum.ADVANCED,
|
|
operationTypes=createOperationTypeRatings(
|
|
(OperationTypeEnum.PLAN, 7),
|
|
(OperationTypeEnum.DATA_ANALYSE, 8),
|
|
(OperationTypeEnum.DATA_GENERATE, 7)
|
|
),
|
|
version="llama-3.1-sonar-large-128k-online",
|
|
calculatePriceUsd=lambda processingTime, bytesSent, bytesReceived: (bytesSent / 4 / 1000) * 0.005 + (bytesReceived / 4 / 1000) * 0.005
|
|
),
|
|
AiModel(
|
|
name="sonar-pro",
|
|
displayName="Perplexity Sonar Pro",
|
|
connectorType="perplexity",
|
|
apiUrl="https://api.perplexity.ai/chat/completions",
|
|
temperature=0.2,
|
|
maxTokens=128000,
|
|
contextLength=128000,
|
|
costPer1kTokensInput=0.01,
|
|
costPer1kTokensOutput=0.01,
|
|
speedRating=6, # Slower due to AI analysis
|
|
qualityRating=10, # Best AI analysis quality
|
|
# capabilities removed (not used in business logic)
|
|
functionCall=self.callAiWithWebSearch,
|
|
priority=PriorityEnum.QUALITY,
|
|
processingMode=ProcessingModeEnum.DETAILED,
|
|
operationTypes=createOperationTypeRatings(
|
|
(OperationTypeEnum.WEB_RESEARCH, 10),
|
|
(OperationTypeEnum.WEB_SEARCH, 9),
|
|
(OperationTypeEnum.WEB_CRAWL, 8),
|
|
(OperationTypeEnum.WEB_NEWS, 8),
|
|
(OperationTypeEnum.WEB_QUESTIONS, 9)
|
|
),
|
|
version="sonar-pro",
|
|
calculatePriceUsd=lambda processingTime, bytesSent, bytesReceived: (bytesSent / 4 / 1000) * 0.01 + (bytesReceived / 4 / 1000) * 0.01
|
|
),
|
|
AiModel(
|
|
name="mistral-7b-instruct",
|
|
displayName="Perplexity Mistral 7B Instruct",
|
|
connectorType="perplexity",
|
|
apiUrl="https://api.perplexity.ai/chat/completions",
|
|
temperature=0.2,
|
|
maxTokens=32000,
|
|
contextLength=32000,
|
|
costPer1kTokensInput=0.002,
|
|
costPer1kTokensOutput=0.002,
|
|
speedRating=9, # Fast for basic AI tasks
|
|
qualityRating=7, # Good but not premium quality
|
|
# capabilities removed (not used in business logic)
|
|
functionCall=self.researchTopic,
|
|
priority=PriorityEnum.COST,
|
|
processingMode=ProcessingModeEnum.BASIC,
|
|
operationTypes=createOperationTypeRatings(
|
|
(OperationTypeEnum.WEB_RESEARCH, 7),
|
|
(OperationTypeEnum.WEB_SEARCH, 6),
|
|
(OperationTypeEnum.WEB_CRAWL, 5),
|
|
(OperationTypeEnum.WEB_NEWS, 5),
|
|
(OperationTypeEnum.WEB_QUESTIONS, 6)
|
|
),
|
|
version="mistral-7b-instruct",
|
|
calculatePriceUsd=lambda processingTime, bytesSent, bytesReceived: (bytesSent / 4 / 1000) * 0.002 + (bytesReceived / 4 / 1000) * 0.002
|
|
),
|
|
AiModel(
|
|
name="mistral-7b-instruct-qa",
|
|
displayName="Perplexity Mistral 7B Instruct QA",
|
|
connectorType="perplexity",
|
|
apiUrl="https://api.perplexity.ai/chat/completions",
|
|
temperature=0.2,
|
|
maxTokens=32000,
|
|
contextLength=32000,
|
|
costPer1kTokensInput=0.002,
|
|
costPer1kTokensOutput=0.002,
|
|
speedRating=9, # Fast for Q&A tasks
|
|
qualityRating=7, # Good but not premium quality
|
|
# capabilities removed (not used in business logic)
|
|
functionCall=self.answerQuestion,
|
|
priority=PriorityEnum.COST,
|
|
processingMode=ProcessingModeEnum.BASIC,
|
|
operationTypes=createOperationTypeRatings(
|
|
(OperationTypeEnum.WEB_RESEARCH, 6),
|
|
(OperationTypeEnum.WEB_SEARCH, 5),
|
|
(OperationTypeEnum.WEB_CRAWL, 4),
|
|
(OperationTypeEnum.WEB_NEWS, 4),
|
|
(OperationTypeEnum.WEB_QUESTIONS, 10)
|
|
),
|
|
version="mistral-7b-instruct",
|
|
calculatePriceUsd=lambda processingTime, bytesSent, bytesReceived: (bytesSent / 4 / 1000) * 0.002 + (bytesReceived / 4 / 1000) * 0.002
|
|
),
|
|
AiModel(
|
|
name="mistral-7b-instruct-news",
|
|
displayName="Perplexity Mistral 7B Instruct News",
|
|
connectorType="perplexity",
|
|
apiUrl="https://api.perplexity.ai/chat/completions",
|
|
temperature=0.2,
|
|
maxTokens=32000,
|
|
contextLength=32000,
|
|
costPer1kTokensInput=0.002,
|
|
costPer1kTokensOutput=0.002,
|
|
speedRating=9, # Fast for news tasks
|
|
qualityRating=7, # Good but not premium quality
|
|
# capabilities removed (not used in business logic)
|
|
functionCall=self.getCurrentNews,
|
|
priority=PriorityEnum.COST,
|
|
processingMode=ProcessingModeEnum.BASIC,
|
|
operationTypes=createOperationTypeRatings(
|
|
(OperationTypeEnum.WEB_RESEARCH, 6),
|
|
(OperationTypeEnum.WEB_SEARCH, 5),
|
|
(OperationTypeEnum.WEB_CRAWL, 4),
|
|
(OperationTypeEnum.WEB_NEWS, 10),
|
|
(OperationTypeEnum.WEB_QUESTIONS, 4)
|
|
),
|
|
version="mistral-7b-instruct",
|
|
calculatePriceUsd=lambda processingTime, bytesSent, bytesReceived: (bytesSent / 4 / 1000) * 0.002 + (bytesReceived / 4 / 1000) * 0.002
|
|
)
|
|
]
|
|
|
|
async def callAiBasic(self, modelCall: AiModelCall) -> AiModelResponse:
|
|
"""
|
|
Calls the Perplexity API with the given messages using standardized pattern.
|
|
|
|
Args:
|
|
modelCall: AiModelCall with messages and options
|
|
|
|
Returns:
|
|
AiModelResponse with content and metadata
|
|
|
|
Raises:
|
|
HTTPException: For errors in API communication
|
|
"""
|
|
try:
|
|
# Extract parameters from modelCall
|
|
messages = modelCall.messages
|
|
model = modelCall.model
|
|
options = modelCall.options
|
|
temperature = options.get("temperature", model.temperature)
|
|
maxTokens = model.maxTokens
|
|
|
|
payload = {
|
|
"model": model.name,
|
|
"messages": messages,
|
|
"temperature": temperature,
|
|
"max_tokens": maxTokens
|
|
}
|
|
|
|
response = await self.httpClient.post(
|
|
model.apiUrl,
|
|
json=payload
|
|
)
|
|
|
|
if response.status_code != 200:
|
|
error_detail = f"Perplexity API error: {response.status_code} - {response.text}"
|
|
logger.error(error_detail)
|
|
|
|
# Provide more specific error messages based on status code
|
|
if response.status_code == 429:
|
|
error_message = "Rate limit exceeded. Please wait before making another request."
|
|
elif response.status_code == 401:
|
|
error_message = "Invalid API key. Please check your Perplexity API configuration."
|
|
elif response.status_code == 400:
|
|
error_message = f"Invalid request to Perplexity API: {response.text}"
|
|
else:
|
|
error_message = f"Perplexity API error ({response.status_code}): {response.text}"
|
|
|
|
raise HTTPException(status_code=500, detail=error_message)
|
|
|
|
responseJson = response.json()
|
|
content = responseJson["choices"][0]["message"]["content"]
|
|
|
|
return AiModelResponse(
|
|
content=content,
|
|
success=True,
|
|
modelId=model.name,
|
|
metadata={"response_id": responseJson.get("id", "")}
|
|
)
|
|
|
|
except Exception as e:
|
|
logger.error(f"Error calling Perplexity API: {str(e)}")
|
|
raise HTTPException(status_code=500, detail=f"Error calling Perplexity API: {str(e)}")
|
|
|
|
async def callAiWithWebSearch(self, modelCall: AiModelCall) -> AiModelResponse:
|
|
"""
|
|
Calls Perplexity API with web search capabilities for research using standardized pattern.
|
|
|
|
Args:
|
|
modelCall: AiModelCall with messages and options
|
|
|
|
Returns:
|
|
AiModelResponse with content and metadata
|
|
"""
|
|
try:
|
|
# Extract parameters from modelCall
|
|
messages = modelCall.messages
|
|
model = modelCall.model
|
|
options = modelCall.options
|
|
temperature = options.get("temperature", model.temperature)
|
|
maxTokens = model.maxTokens
|
|
|
|
payload = {
|
|
"model": model.name,
|
|
"messages": messages,
|
|
"temperature": temperature,
|
|
"max_tokens": maxTokens
|
|
}
|
|
|
|
response = await self.httpClient.post(
|
|
model.apiUrl,
|
|
json=payload
|
|
)
|
|
|
|
if response.status_code != 200:
|
|
error_detail = f"Perplexity Web Search API error: {response.status_code} - {response.text}"
|
|
logger.error(error_detail)
|
|
|
|
if response.status_code == 429:
|
|
error_message = "Rate limit exceeded for web search. Please wait before making another request."
|
|
elif response.status_code == 401:
|
|
error_message = "Invalid API key for web search. Please check your Perplexity API configuration."
|
|
elif response.status_code == 400:
|
|
error_message = f"Invalid request to Perplexity Web Search API: {response.text}"
|
|
else:
|
|
error_message = f"Perplexity Web Search API error ({response.status_code}): {response.text}"
|
|
|
|
raise HTTPException(status_code=500, detail=error_message)
|
|
|
|
responseJson = response.json()
|
|
content = responseJson["choices"][0]["message"]["content"]
|
|
|
|
return AiModelResponse(
|
|
content=content,
|
|
success=True,
|
|
modelId=model.name,
|
|
metadata={"response_id": responseJson.get("id", "")}
|
|
)
|
|
|
|
except Exception as e:
|
|
logger.error(f"Error calling Perplexity Web Search API: {str(e)}")
|
|
raise HTTPException(status_code=500, detail=f"Error calling Perplexity Web Search API: {str(e)}")
|
|
|
|
async def researchTopic(self, modelCall: AiModelCall) -> AiModelResponse:
|
|
"""
|
|
Research a topic using Perplexity's web search capabilities using standardized pattern.
|
|
|
|
Args:
|
|
modelCall: AiModelCall with messages and options
|
|
|
|
Returns:
|
|
AiModelResponse with research content
|
|
"""
|
|
try:
|
|
# Extract parameters from modelCall
|
|
messages = modelCall.messages
|
|
model = modelCall.model
|
|
options = modelCall.options
|
|
temperature = options.get("temperature", model.temperature)
|
|
maxTokens = model.maxTokens
|
|
|
|
payload = {
|
|
"model": model.name,
|
|
"messages": messages,
|
|
"temperature": temperature,
|
|
"max_tokens": maxTokens
|
|
}
|
|
|
|
response = await self.httpClient.post(
|
|
model.apiUrl,
|
|
json=payload
|
|
)
|
|
|
|
if response.status_code != 200:
|
|
error_detail = f"Perplexity Research API error: {response.status_code} - {response.text}"
|
|
logger.error(error_detail)
|
|
|
|
if response.status_code == 429:
|
|
error_message = "Rate limit exceeded for research. Please wait before making another request."
|
|
elif response.status_code == 401:
|
|
error_message = "Invalid API key for research. Please check your Perplexity API configuration."
|
|
elif response.status_code == 400:
|
|
error_message = f"Invalid request to Perplexity Research API: {response.text}"
|
|
else:
|
|
error_message = f"Perplexity Research API error ({response.status_code}): {response.text}"
|
|
|
|
raise HTTPException(status_code=500, detail=error_message)
|
|
|
|
responseJson = response.json()
|
|
content = responseJson["choices"][0]["message"]["content"]
|
|
|
|
return AiModelResponse(
|
|
content=content,
|
|
success=True,
|
|
modelId=model.name,
|
|
metadata={"response_id": responseJson.get("id", "")}
|
|
)
|
|
|
|
except Exception as e:
|
|
logger.error(f"Error researching topic: {str(e)}")
|
|
raise HTTPException(status_code=500, detail=f"Error researching topic: {str(e)}")
|
|
|
|
async def answerQuestion(self, modelCall: AiModelCall) -> AiModelResponse:
|
|
"""
|
|
Answer a question using web search for current information using standardized pattern.
|
|
|
|
Args:
|
|
modelCall: AiModelCall with messages and options
|
|
|
|
Returns:
|
|
AiModelResponse with answer content
|
|
"""
|
|
try:
|
|
# Extract parameters from modelCall
|
|
messages = modelCall.messages
|
|
model = modelCall.model
|
|
options = modelCall.options
|
|
temperature = options.get("temperature", model.temperature)
|
|
maxTokens = model.maxTokens
|
|
|
|
payload = {
|
|
"model": model.name,
|
|
"messages": messages,
|
|
"temperature": temperature,
|
|
"max_tokens": maxTokens
|
|
}
|
|
|
|
response = await self.httpClient.post(
|
|
model.apiUrl,
|
|
json=payload
|
|
)
|
|
|
|
if response.status_code != 200:
|
|
error_detail = f"Perplexity Q&A API error: {response.status_code} - {response.text}"
|
|
logger.error(error_detail)
|
|
|
|
if response.status_code == 429:
|
|
error_message = "Rate limit exceeded for Q&A. Please wait before making another request."
|
|
elif response.status_code == 401:
|
|
error_message = "Invalid API key for Q&A. Please check your Perplexity API configuration."
|
|
elif response.status_code == 400:
|
|
error_message = f"Invalid request to Perplexity Q&A API: {response.text}"
|
|
else:
|
|
error_message = f"Perplexity Q&A API error ({response.status_code}): {response.text}"
|
|
|
|
raise HTTPException(status_code=500, detail=error_message)
|
|
|
|
responseJson = response.json()
|
|
content = responseJson["choices"][0]["message"]["content"]
|
|
|
|
return AiModelResponse(
|
|
content=content,
|
|
success=True,
|
|
modelId=model.name,
|
|
metadata={"response_id": responseJson.get("id", "")}
|
|
)
|
|
|
|
except Exception as e:
|
|
logger.error(f"Error answering question: {str(e)}")
|
|
raise HTTPException(status_code=500, detail=f"Error answering question: {str(e)}")
|
|
|
|
async def getCurrentNews(self, modelCall: AiModelCall) -> AiModelResponse:
|
|
"""
|
|
Get current news on a specific topic using standardized pattern.
|
|
|
|
Args:
|
|
modelCall: AiModelCall with messages and options
|
|
|
|
Returns:
|
|
AiModelResponse with news content
|
|
"""
|
|
try:
|
|
# Extract parameters from modelCall
|
|
messages = modelCall.messages
|
|
model = modelCall.model
|
|
options = modelCall.options
|
|
temperature = options.get("temperature", model.temperature)
|
|
maxTokens = model.maxTokens
|
|
|
|
payload = {
|
|
"model": model.name,
|
|
"messages": messages,
|
|
"temperature": temperature,
|
|
"max_tokens": maxTokens
|
|
}
|
|
|
|
response = await self.httpClient.post(
|
|
model.apiUrl,
|
|
json=payload
|
|
)
|
|
|
|
if response.status_code != 200:
|
|
error_detail = f"Perplexity News API error: {response.status_code} - {response.text}"
|
|
logger.error(error_detail)
|
|
|
|
if response.status_code == 429:
|
|
error_message = "Rate limit exceeded for news. Please wait before making another request."
|
|
elif response.status_code == 401:
|
|
error_message = "Invalid API key for news. Please check your Perplexity API configuration."
|
|
elif response.status_code == 400:
|
|
error_message = f"Invalid request to Perplexity News API: {response.text}"
|
|
else:
|
|
error_message = f"Perplexity News API error ({response.status_code}): {response.text}"
|
|
|
|
raise HTTPException(status_code=500, detail=error_message)
|
|
|
|
responseJson = response.json()
|
|
content = responseJson["choices"][0]["message"]["content"]
|
|
|
|
return AiModelResponse(
|
|
content=content,
|
|
success=True,
|
|
modelId=model.name,
|
|
metadata={"response_id": responseJson.get("id", "")}
|
|
)
|
|
|
|
except Exception as e:
|
|
logger.error(f"Error getting current news: {str(e)}")
|
|
raise HTTPException(status_code=500, detail=f"Error getting current news: {str(e)}")
|
|
|
|
async def _testConnection(self) -> bool:
|
|
"""
|
|
Tests the connection to the Perplexity API.
|
|
|
|
Returns:
|
|
True if connection is successful, False otherwise
|
|
"""
|
|
try:
|
|
# Try a simple test message
|
|
testMessages = [
|
|
{"role": "user", "content": "Hello, please respond with just 'OK' to confirm the connection works."}
|
|
]
|
|
|
|
response = await self.callAiBasic(testMessages)
|
|
return response and len(response.strip()) > 0
|
|
|
|
except Exception as e:
|
|
logger.error(f"Perplexity connection test failed: {str(e)}")
|
|
return False
|