859 lines
35 KiB
Python
859 lines
35 KiB
Python
# Copyright (c) 2025 Patrick Motsch
|
|
# All rights reserved.
|
|
import json
|
|
import logging
|
|
import re
|
|
import time
|
|
import base64
|
|
from typing import Dict, Any, List, Optional, Tuple
|
|
from modules.datamodels.datamodelChat import PromptPlaceholder, ChatDocument
|
|
from modules.services.serviceExtraction.mainServiceExtraction import ExtractionService
|
|
from modules.datamodels.datamodelAi import AiCallRequest, AiCallOptions, OperationTypeEnum, PriorityEnum, ProcessingModeEnum
|
|
from modules.datamodels.datamodelExtraction import ContentPart, DocumentIntent
|
|
from modules.datamodels.datamodelWorkflow import AiResponse, AiResponseMetadata, DocumentData
|
|
from modules.datamodels.datamodelDocument import RenderedDocument
|
|
from modules.interfaces.interfaceAiObjects import AiObjects
|
|
from modules.shared.jsonUtils import (
|
|
parseJsonWithModel
|
|
)
|
|
from modules.services.serviceAi.subJsonResponseHandling import JsonResponseHandler
|
|
from modules.datamodels.datamodelAi import JsonAccumulationState
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
# Rebuild the model to resolve forward references
|
|
AiCallRequest.model_rebuild()
|
|
|
|
class AiService:
|
|
"""AI service with core operations integrated."""
|
|
|
|
def __init__(self, serviceCenter=None) -> None:
|
|
"""Initialize AI service with service center access.
|
|
|
|
Args:
|
|
serviceCenter: Service center instance for accessing other services
|
|
"""
|
|
self.services = serviceCenter
|
|
# Only depend on interfaces
|
|
self.aiObjects = None # Will be initialized in create() or ensureAiObjectsInitialized()
|
|
# Submodules initialized as None - will be set in _initializeSubmodules() after aiObjects is ready
|
|
self.extractionService = None
|
|
|
|
def _initializeSubmodules(self):
|
|
"""Initialize all submodules after aiObjects is ready."""
|
|
if self.aiObjects is None:
|
|
raise RuntimeError("aiObjects must be initialized before initializing submodules")
|
|
|
|
if self.extractionService is None:
|
|
logger.info("Initializing ExtractionService...")
|
|
self.extractionService = ExtractionService(self.services)
|
|
|
|
# Initialize new submodules
|
|
from modules.services.serviceAi.subResponseParsing import ResponseParser
|
|
from modules.services.serviceAi.subDocumentIntents import DocumentIntentAnalyzer
|
|
from modules.services.serviceAi.subContentExtraction import ContentExtractor
|
|
from modules.services.serviceAi.subStructureGeneration import StructureGenerator
|
|
from modules.services.serviceAi.subStructureFilling import StructureFiller
|
|
from modules.services.serviceAi.subAiCallLooping import AiCallLooper
|
|
|
|
if not hasattr(self, 'responseParser'):
|
|
logger.info("Initializing ResponseParser...")
|
|
self.responseParser = ResponseParser(self.services)
|
|
|
|
if not hasattr(self, 'intentAnalyzer'):
|
|
logger.info("Initializing DocumentIntentAnalyzer...")
|
|
self.intentAnalyzer = DocumentIntentAnalyzer(self.services, self)
|
|
|
|
if not hasattr(self, 'contentExtractor'):
|
|
logger.info("Initializing ContentExtractor...")
|
|
self.contentExtractor = ContentExtractor(self.services, self, self.intentAnalyzer)
|
|
|
|
if not hasattr(self, 'structureGenerator'):
|
|
logger.info("Initializing StructureGenerator...")
|
|
self.structureGenerator = StructureGenerator(self.services, self)
|
|
|
|
if not hasattr(self, 'structureFiller'):
|
|
logger.info("Initializing StructureFiller...")
|
|
self.structureFiller = StructureFiller(self.services, self)
|
|
|
|
if not hasattr(self, 'aiCallLooper'):
|
|
logger.info("Initializing AiCallLooper...")
|
|
self.aiCallLooper = AiCallLooper(self.services, self, self.responseParser)
|
|
|
|
async def callAi(self, request: AiCallRequest, progressCallback=None):
|
|
"""Router: handles content parts via extractionService, text context via interface.
|
|
|
|
Replaces direct calls to self.aiObjects.call() to route content parts processing
|
|
through serviceExtraction layer.
|
|
"""
|
|
if hasattr(request, 'contentParts') and request.contentParts:
|
|
return await self.extractionService.processContentPartsWithAi(
|
|
request, self.aiObjects, progressCallback
|
|
)
|
|
return await self.aiObjects.callWithTextContext(request)
|
|
|
|
async def ensureAiObjectsInitialized(self):
|
|
"""Ensure aiObjects is initialized and submodules are ready."""
|
|
if self.aiObjects is None:
|
|
logger.info("Lazy initializing AiObjects...")
|
|
self.aiObjects = await AiObjects.create()
|
|
logger.info("AiObjects initialization completed")
|
|
# Initialize submodules after aiObjects is ready
|
|
self._initializeSubmodules()
|
|
|
|
@classmethod
|
|
async def create(cls, serviceCenter=None) -> "AiService":
|
|
"""Create AiService instance with all connectors and submodules initialized."""
|
|
logger.info("AiService.create() called")
|
|
instance = cls(serviceCenter)
|
|
logger.info("AiService created, about to call AiObjects.create()...")
|
|
instance.aiObjects = await AiObjects.create()
|
|
logger.info("AiObjects.create() completed")
|
|
# Initialize all submodules after aiObjects is ready
|
|
instance._initializeSubmodules()
|
|
logger.info("AiService submodules initialized")
|
|
return instance
|
|
|
|
# Helper methods
|
|
|
|
def _buildPromptWithPlaceholders(self, prompt: str, placeholders: Optional[Dict[str, str]]) -> str:
|
|
"""
|
|
Build full prompt by replacing placeholders with their content.
|
|
Uses the new {{KEY:placeholder}} format.
|
|
|
|
Args:
|
|
prompt: The base prompt template
|
|
placeholders: Dictionary of placeholder key-value pairs
|
|
|
|
Returns:
|
|
Prompt with placeholders replaced
|
|
"""
|
|
if not placeholders:
|
|
return prompt
|
|
|
|
full_prompt = prompt
|
|
for placeholder, content in placeholders.items():
|
|
# Skip if content is None or empty
|
|
if content is None:
|
|
continue
|
|
# Replace {{KEY:placeholder}}
|
|
full_prompt = full_prompt.replace(f"{{{{KEY:{placeholder}}}}}", str(content))
|
|
|
|
return full_prompt
|
|
|
|
async def _analyzePromptAndCreateOptions(self, prompt: str) -> AiCallOptions:
|
|
"""Analyze prompt to determine appropriate AiCallOptions parameters."""
|
|
try:
|
|
# Get dynamic enum values from Pydantic models
|
|
operationTypes = [e.value for e in OperationTypeEnum]
|
|
priorities = [e.value for e in PriorityEnum]
|
|
processingModes = [e.value for e in ProcessingModeEnum]
|
|
|
|
# Create analysis prompt for AI to determine operation type and parameters
|
|
analysisPrompt = f"""
|
|
You are an AI operation analyzer. Analyze the following prompt and determine the most appropriate operation type and parameters.
|
|
|
|
PROMPT TO ANALYZE:
|
|
{self.services.utils.sanitizePromptContent(prompt, 'userinput')}
|
|
|
|
Based on the prompt content, determine:
|
|
1. operationType: Choose the most appropriate from: {', '.join(operationTypes)}
|
|
2. priority: Choose from: {', '.join(priorities)}
|
|
3. processingMode: Choose from: {', '.join(processingModes)}
|
|
4. compressPrompt: true/false (true for story-like prompts, false for structured prompts with JSON/schemas)
|
|
5. compressContext: true/false (true to summarize context, false to process fully)
|
|
|
|
Respond with ONLY a JSON object in this exact format:
|
|
{{
|
|
"operationType": "dataAnalyse",
|
|
"priority": "balanced",
|
|
"processingMode": "basic",
|
|
"compressPrompt": true,
|
|
"compressContext": true
|
|
}}
|
|
"""
|
|
|
|
# Use AI to analyze the prompt
|
|
request = AiCallRequest(
|
|
prompt=analysisPrompt,
|
|
options=AiCallOptions(
|
|
operationType=OperationTypeEnum.DATA_ANALYSE,
|
|
priority=PriorityEnum.SPEED,
|
|
processingMode=ProcessingModeEnum.BASIC,
|
|
compressPrompt=True,
|
|
compressContext=False
|
|
)
|
|
)
|
|
|
|
response = await self.callAi(request)
|
|
|
|
# Parse AI response using structured parsing with AiCallOptions model
|
|
try:
|
|
# Use parseJsonWithModel to parse response into AiCallOptions (handles enum conversion automatically)
|
|
analysis = parseJsonWithModel(response.content, AiCallOptions)
|
|
return analysis
|
|
except Exception as e:
|
|
logger.warning(f"Failed to parse AI analysis response: {e}")
|
|
|
|
except Exception as e:
|
|
logger.warning(f"Prompt analysis failed: {e}")
|
|
|
|
# Fallback to default options
|
|
return AiCallOptions(
|
|
operationType=OperationTypeEnum.DATA_ANALYSE,
|
|
priority=PriorityEnum.BALANCED,
|
|
processingMode=ProcessingModeEnum.BASIC
|
|
)
|
|
|
|
async def callAiWithLooping(
|
|
self,
|
|
prompt: str,
|
|
options: AiCallOptions,
|
|
debugPrefix: str = "ai_call",
|
|
promptBuilder: Optional[callable] = None,
|
|
promptArgs: Optional[Dict[str, Any]] = None,
|
|
operationId: Optional[str] = None,
|
|
userPrompt: Optional[str] = None,
|
|
contentParts: Optional[List[ContentPart]] = None, # ARCHITECTURE: Support ContentParts for large content
|
|
useCaseId: Optional[str] = None # REQUIRED: Explicit use case ID for generic looping system
|
|
) -> str:
|
|
"""Public method: Delegate to AiCallLooper for AI calls with looping support."""
|
|
return await self.aiCallLooper.callAiWithLooping(
|
|
prompt, options, debugPrefix, promptBuilder, promptArgs, operationId, userPrompt, contentParts, useCaseId
|
|
)
|
|
|
|
async def _defineKpisFromPrompt(
|
|
self,
|
|
userPrompt: str,
|
|
rawJsonString: Optional[str],
|
|
continuationContext: Dict[str, Any],
|
|
debugPrefix: str = "kpi"
|
|
) -> List[Dict[str, Any]]:
|
|
"""Delegate to AiCallLooper."""
|
|
return await self.aiCallLooper._defineKpisFromPrompt(
|
|
userPrompt, rawJsonString, continuationContext, debugPrefix
|
|
)
|
|
|
|
# JSON merging logic moved to subJsonResponseHandling.py
|
|
|
|
def _extractSectionsFromResponse(
|
|
self,
|
|
result: str,
|
|
iteration: int,
|
|
debugPrefix: str,
|
|
allSections: List[Dict[str, Any]] = None,
|
|
accumulationState: Optional[JsonAccumulationState] = None
|
|
) -> Tuple[List[Dict[str, Any]], bool, Optional[Dict[str, Any]], Optional[JsonAccumulationState]]:
|
|
"""Delegate to ResponseParser."""
|
|
return self.responseParser.extractSectionsFromResponse(
|
|
result, iteration, debugPrefix, allSections, accumulationState
|
|
)
|
|
|
|
def _shouldContinueGeneration(
|
|
self,
|
|
allSections: List[Dict[str, Any]],
|
|
iteration: int,
|
|
wasJsonComplete: bool,
|
|
rawResponse: str = None
|
|
) -> bool:
|
|
"""Delegate to ResponseParser."""
|
|
return self.responseParser.shouldContinueGeneration(
|
|
allSections, iteration, wasJsonComplete, rawResponse
|
|
)
|
|
|
|
def _extractDocumentMetadata(
|
|
self,
|
|
parsedResult: Dict[str, Any]
|
|
) -> Optional[Dict[str, Any]]:
|
|
"""Delegate to ResponseParser."""
|
|
return self.responseParser.extractDocumentMetadata(parsedResult)
|
|
|
|
def _buildFinalResultFromSections(
|
|
self,
|
|
allSections: List[Dict[str, Any]],
|
|
documentMetadata: Optional[Dict[str, Any]] = None
|
|
) -> str:
|
|
"""Delegate to ResponseParser."""
|
|
return self.responseParser.buildFinalResultFromSections(allSections, documentMetadata)
|
|
|
|
# Public API Methods
|
|
|
|
# Planning AI Call
|
|
async def callAiPlanning(
|
|
self,
|
|
prompt: str,
|
|
placeholders: Optional[List[PromptPlaceholder]] = None,
|
|
debugType: Optional[str] = None
|
|
) -> str:
|
|
"""
|
|
Planning AI call for task planning, action planning, action selection, etc.
|
|
Always uses static parameters optimized for planning tasks.
|
|
|
|
Args:
|
|
prompt: The planning prompt
|
|
placeholders: Optional list of placeholder replacements
|
|
debugType: Optional debug file type identifier (e.g., 'taskplan', 'dynamic', 'intentanalysis')
|
|
If not provided, defaults to 'plan'
|
|
|
|
Returns:
|
|
Planning JSON response
|
|
"""
|
|
await self.ensureAiObjectsInitialized()
|
|
|
|
# Planning calls always use static parameters
|
|
options = AiCallOptions(
|
|
operationType=OperationTypeEnum.PLAN,
|
|
priority=PriorityEnum.QUALITY,
|
|
processingMode=ProcessingModeEnum.DETAILED,
|
|
compressPrompt=False,
|
|
compressContext=False
|
|
)
|
|
|
|
# Build full prompt with placeholders
|
|
if placeholders:
|
|
placeholdersDict = {p.label: p.content for p in placeholders}
|
|
fullPrompt = self._buildPromptWithPlaceholders(prompt, placeholdersDict)
|
|
else:
|
|
fullPrompt = prompt
|
|
|
|
# Root-cause fix: planning must return raw single-shot JSON, not section-based output
|
|
request = AiCallRequest(
|
|
prompt=fullPrompt,
|
|
context="",
|
|
options=options
|
|
)
|
|
|
|
# Debug: persist prompt/response for analysis with context-specific naming
|
|
debugPrefix = debugType if debugType else "plan"
|
|
self.services.utils.writeDebugFile(fullPrompt, f"{debugPrefix}_prompt")
|
|
response = await self.aiObjects.callWithTextContext(request)
|
|
result = response.content or ""
|
|
self.services.utils.writeDebugFile(result, f"{debugPrefix}_response")
|
|
return result
|
|
|
|
# Helper methods for callAiContent refactoring
|
|
|
|
async def _handleImageGeneration(
|
|
self,
|
|
prompt: str,
|
|
options: AiCallOptions,
|
|
title: Optional[str],
|
|
parentOperationId: Optional[str]
|
|
) -> AiResponse:
|
|
"""Handle IMAGE_GENERATE operation type using image generation path."""
|
|
from modules.services.serviceGeneration.paths.imagePath import ImageGenerationPath
|
|
|
|
imagePath = ImageGenerationPath(self.services)
|
|
|
|
# Extract format from options
|
|
format = options.resultFormat or "png"
|
|
|
|
return await imagePath.generateImages(
|
|
userPrompt=prompt,
|
|
format=format,
|
|
title=title,
|
|
parentOperationId=parentOperationId
|
|
)
|
|
|
|
async def _handleWebOperation(
|
|
self,
|
|
prompt: str,
|
|
options: AiCallOptions,
|
|
opType: OperationTypeEnum,
|
|
aiOperationId: str
|
|
) -> AiResponse:
|
|
"""Handle WEB_SEARCH_DATA and WEB_CRAWL operation types."""
|
|
self.services.chat.progressLogUpdate(aiOperationId, 0.4, f"Calling AI for {opType.name}")
|
|
|
|
request = AiCallRequest(
|
|
prompt=prompt, # Raw JSON prompt - connector will parse it
|
|
context="",
|
|
options=options
|
|
)
|
|
|
|
response = await self.callAi(request)
|
|
|
|
if not response.content:
|
|
errorMsg = f"No content returned from {opType.name}: {response.content}"
|
|
logger.error(f"Error in {opType.name}: {errorMsg}")
|
|
self.services.chat.progressLogFinish(aiOperationId, False)
|
|
raise ValueError(errorMsg)
|
|
|
|
metadata = AiResponseMetadata(
|
|
operationType=opType.value
|
|
)
|
|
|
|
self.services.chat.storeWorkflowStat(
|
|
self.services.workflow,
|
|
response,
|
|
f"ai.{opType.name.lower()}"
|
|
)
|
|
|
|
self.services.chat.progressLogUpdate(aiOperationId, 0.9, f"{opType.name} completed")
|
|
self.services.chat.progressLogFinish(aiOperationId, True)
|
|
|
|
return AiResponse(
|
|
content=response.content,
|
|
metadata=metadata
|
|
)
|
|
|
|
def _getIntentForDocument(
|
|
self,
|
|
docId: str,
|
|
intents: Optional[List[DocumentIntent]]
|
|
) -> Optional[DocumentIntent]:
|
|
"""Find DocumentIntent for given documentId."""
|
|
if not intents:
|
|
return None
|
|
for intent in intents:
|
|
if intent.documentId == docId:
|
|
return intent
|
|
return None
|
|
|
|
async def clarifyDocumentIntents(
|
|
self,
|
|
documents: List[ChatDocument],
|
|
userPrompt: str,
|
|
actionParameters: Dict[str, Any],
|
|
parentOperationId: str
|
|
) -> List[DocumentIntent]:
|
|
"""Public method: Delegate to DocumentIntentAnalyzer."""
|
|
return await self.intentAnalyzer.clarifyDocumentIntents(
|
|
documents, userPrompt, actionParameters, parentOperationId
|
|
)
|
|
|
|
async def extractAndPrepareContent(
|
|
self,
|
|
documents: List[ChatDocument],
|
|
documentIntents: List[DocumentIntent],
|
|
parentOperationId: str
|
|
) -> List[ContentPart]:
|
|
"""Public method: Delegate to ContentExtractor."""
|
|
return await self.contentExtractor.extractAndPrepareContent(
|
|
documents, documentIntents, parentOperationId, self._getIntentForDocument
|
|
)
|
|
|
|
async def generateStructure(
|
|
self,
|
|
userPrompt: str,
|
|
contentParts: List[ContentPart],
|
|
outputFormat: str,
|
|
parentOperationId: str
|
|
) -> Dict[str, Any]:
|
|
"""Public method: Delegate to StructureGenerator."""
|
|
return await self.structureGenerator.generateStructure(
|
|
userPrompt, contentParts, outputFormat, parentOperationId
|
|
)
|
|
|
|
async def fillStructure(
|
|
self,
|
|
structure: Dict[str, Any],
|
|
contentParts: List[ContentPart],
|
|
userPrompt: str,
|
|
parentOperationId: str
|
|
) -> Dict[str, Any]:
|
|
"""Public method: Delegate to StructureFiller."""
|
|
return await self.structureFiller.fillStructure(
|
|
structure, contentParts, userPrompt, parentOperationId
|
|
)
|
|
|
|
async def renderResult(
|
|
self,
|
|
filledStructure: Dict[str, Any],
|
|
outputFormat: str,
|
|
title: str,
|
|
userPrompt: str,
|
|
parentOperationId: str
|
|
) -> List[RenderedDocument]:
|
|
"""
|
|
Phase 5E: Rendert gefüllte Struktur zum Ziel-Format.
|
|
Jedes Dokument wird einzeln gerendert, jeder Renderer kann 1..n Dokumente zurückgeben.
|
|
|
|
Args:
|
|
filledStructure: Gefüllte Struktur mit elements
|
|
outputFormat: Ziel-Format (pdf, docx, html, etc.) - wird für alle Dokumente verwendet
|
|
title: Dokument-Titel
|
|
userPrompt: User-Anfrage
|
|
parentOperationId: Parent Operation-ID für ChatLog-Hierarchie
|
|
|
|
Returns:
|
|
List of RenderedDocument objects.
|
|
Jedes RenderedDocument repräsentiert ein gerendertes Dokument (Hauptdokument oder unterstützende Datei)
|
|
"""
|
|
# Erstelle Operation-ID für Rendering
|
|
renderOperationId = f"{parentOperationId}_rendering"
|
|
|
|
# Starte ChatLog mit Parent-Referenz
|
|
self.services.chat.progressLogStart(
|
|
renderOperationId,
|
|
"Content Rendering",
|
|
"Rendering",
|
|
f"Rendering to {outputFormat} format",
|
|
parentOperationId=parentOperationId
|
|
)
|
|
|
|
try:
|
|
from modules.services.serviceGeneration.mainServiceGeneration import GenerationService
|
|
|
|
generationService = GenerationService(self.services)
|
|
|
|
# renderReport verarbeitet jetzt jedes Dokument einzeln
|
|
# und gibt Liste von (documentData, mimeType, filename) zurück
|
|
renderedDocuments = await generationService.renderReport(
|
|
filledStructure,
|
|
outputFormat,
|
|
title,
|
|
userPrompt,
|
|
self,
|
|
parentOperationId=renderOperationId # Parent-Referenz für ChatLog-Hierarchie
|
|
)
|
|
|
|
# ChatLog abschließen
|
|
self.services.chat.progressLogFinish(renderOperationId, True)
|
|
|
|
return renderedDocuments
|
|
|
|
except Exception as e:
|
|
self.services.chat.progressLogFinish(renderOperationId, False)
|
|
logger.error(f"Error in _renderResult: {str(e)}")
|
|
raise
|
|
|
|
def _shouldSkipContentPart(
|
|
self,
|
|
part: ContentPart
|
|
) -> bool:
|
|
"""Check if ContentPart should be skipped (already structured JSON)."""
|
|
if part.typeGroup == "structure" and part.mimeType == "application/json":
|
|
if part.metadata.get("skipExtraction", False):
|
|
logger.debug(f"Skipping already-structured JSON ContentPart {part.id} (skipExtraction=True)")
|
|
return True
|
|
try:
|
|
if isinstance(part.data, str):
|
|
jsonData = json.loads(part.data)
|
|
if isinstance(jsonData, dict) and ("documents" in jsonData or "sections" in jsonData):
|
|
logger.debug(f"Skipping already-structured JSON ContentPart {part.id} (contains documents/sections)")
|
|
return True
|
|
except Exception:
|
|
pass # Not JSON, continue processing
|
|
return False
|
|
|
|
async def callAiContent(
|
|
self,
|
|
prompt: str,
|
|
options: AiCallOptions,
|
|
contentParts: Optional[List[ContentPart]] = None,
|
|
documentList: Optional[Any] = None, # DocumentReferenceList
|
|
documentIntents: Optional[List[DocumentIntent]] = None,
|
|
outputFormat: Optional[str] = None,
|
|
title: Optional[str] = None,
|
|
parentOperationId: Optional[str] = None,
|
|
generationIntent: Optional[str] = None # NEW: Explicit intent from action (skips detection)
|
|
) -> AiResponse:
|
|
"""
|
|
Unified AI content generation with explicit intent requirement.
|
|
|
|
All AI-Actions (ai.process, ai.generateDocument, etc.) route through here.
|
|
They differ only in parameters, not in logic.
|
|
|
|
Args:
|
|
prompt: The main prompt for the AI call
|
|
options: AI call configuration options (REQUIRED - operationType must be set)
|
|
contentParts: Optional list of already-extracted content parts (preferred)
|
|
documentList: Optional DocumentReferenceList (wird zu ChatDocuments konvertiert)
|
|
documentIntents: Optional list of DocumentIntent objects (wird erstellt wenn nicht vorhanden)
|
|
outputFormat: Optional output format for document generation (e.g., 'pdf', 'docx', 'xlsx')
|
|
title: Optional title for generated documents
|
|
parentOperationId: Optional parent operation ID for hierarchical logging
|
|
generationIntent: REQUIRED explicit intent ("document" | "code" | "image") from action.
|
|
NO auto-detection - actions must explicitly specify intent.
|
|
|
|
Returns:
|
|
AiResponse with content, metadata, and optional documents
|
|
"""
|
|
await self.ensureAiObjectsInitialized()
|
|
|
|
# Erstelle Operation-ID
|
|
workflowId = self.services.workflow.id if self.services.workflow else f"no-workflow-{int(time.time())}"
|
|
aiOperationId = f"ai_content_{workflowId}_{int(time.time())}"
|
|
|
|
# Starte Progress-Tracking mit Parent-Referenz
|
|
self.services.chat.progressLogStart(
|
|
aiOperationId,
|
|
"AI content processing",
|
|
"Content Processing",
|
|
f"Format: {outputFormat or 'text'}",
|
|
parentOperationId=parentOperationId
|
|
)
|
|
|
|
try:
|
|
# Initialisiere Defaults
|
|
if not outputFormat:
|
|
outputFormat = "txt"
|
|
|
|
opType = getattr(options, "operationType", None)
|
|
if not opType:
|
|
options.operationType = OperationTypeEnum.DATA_GENERATE
|
|
opType = OperationTypeEnum.DATA_GENERATE
|
|
|
|
# Route zu Operation-spezifischen Handlern
|
|
if opType == OperationTypeEnum.IMAGE_GENERATE:
|
|
# Image generation - route to image path
|
|
return await self._handleImageGeneration(prompt, options, title, parentOperationId)
|
|
|
|
if opType == OperationTypeEnum.WEB_SEARCH_DATA or opType == OperationTypeEnum.WEB_CRAWL:
|
|
return await self._handleWebOperation(prompt, options, opType, aiOperationId)
|
|
|
|
# Data generation - REQUIRES explicit generationIntent
|
|
if opType == OperationTypeEnum.DATA_GENERATE:
|
|
if not generationIntent:
|
|
errorMsg = (
|
|
"generationIntent is required for DATA_GENERATE operation. "
|
|
"Actions must explicitly specify 'document' or 'code' intent. "
|
|
"No auto-detection - use qualified actions (ai.generateDocument, ai.generateCode)."
|
|
)
|
|
logger.error(errorMsg)
|
|
self.services.chat.progressLogFinish(aiOperationId, False)
|
|
raise ValueError(errorMsg)
|
|
|
|
# Route based on explicit intent (no auto-detection, no fallback)
|
|
if generationIntent == "code":
|
|
# Route to code generation path
|
|
return await self._handleCodeGeneration(
|
|
prompt=prompt,
|
|
options=options,
|
|
contentParts=contentParts,
|
|
outputFormat=outputFormat,
|
|
title=title,
|
|
parentOperationId=parentOperationId
|
|
)
|
|
else:
|
|
# Route to document generation path (existing behavior)
|
|
return await self._handleDocumentGeneration(
|
|
prompt=prompt,
|
|
options=options,
|
|
documentList=documentList,
|
|
documentIntents=documentIntents,
|
|
contentParts=contentParts,
|
|
outputFormat=outputFormat,
|
|
title=title,
|
|
parentOperationId=parentOperationId
|
|
)
|
|
|
|
# DATA_EXTRACT: Extract content from documents and process with AI (no structure generation)
|
|
if opType == OperationTypeEnum.DATA_EXTRACT:
|
|
return await self._handleDataExtraction(
|
|
prompt=prompt,
|
|
options=options,
|
|
documentList=documentList,
|
|
documentIntents=documentIntents,
|
|
contentParts=contentParts,
|
|
outputFormat=outputFormat,
|
|
title=title,
|
|
parentOperationId=parentOperationId
|
|
)
|
|
|
|
# Other operation types (DATA_ANALYSE, etc.) - existing logic
|
|
# Fallback to document generation for backward compatibility (should not happen)
|
|
logger.warning(f"Unhandled operation type: {opType}, falling back to document generation")
|
|
return await self._handleDocumentGeneration(
|
|
prompt=prompt,
|
|
options=options,
|
|
documentList=documentList,
|
|
documentIntents=documentIntents,
|
|
contentParts=contentParts,
|
|
outputFormat=outputFormat,
|
|
title=title,
|
|
parentOperationId=parentOperationId
|
|
)
|
|
|
|
except Exception as e:
|
|
logger.error(f"Error in callAiContent: {str(e)}")
|
|
self.services.chat.progressLogFinish(aiOperationId, False)
|
|
raise
|
|
|
|
async def _handleDataExtraction(
|
|
self,
|
|
prompt: str,
|
|
options: AiCallOptions,
|
|
documentList: Optional[Any],
|
|
documentIntents: Optional[List[DocumentIntent]],
|
|
contentParts: Optional[List[ContentPart]],
|
|
outputFormat: str,
|
|
title: str,
|
|
parentOperationId: Optional[str]
|
|
) -> AiResponse:
|
|
"""
|
|
Handle DATA_EXTRACT: Extract content from documents (no AI), then process with AI.
|
|
This is the original flow: extract all documents first, then process contentParts with AI.
|
|
"""
|
|
import time
|
|
|
|
# Create operation ID
|
|
workflowId = self.services.workflow.id if self.services.workflow else f"no-workflow-{int(time.time())}"
|
|
extractOperationId = f"data_extract_{workflowId}_{int(time.time())}"
|
|
|
|
# Start progress tracking
|
|
self.services.chat.progressLogStart(
|
|
extractOperationId,
|
|
"Data Extraction",
|
|
"Extraction",
|
|
f"Format: {outputFormat}",
|
|
parentOperationId=parentOperationId
|
|
)
|
|
|
|
try:
|
|
# Step 1: Get documents from documentList
|
|
documents = []
|
|
if documentList:
|
|
documents = self.services.chat.getChatDocumentsFromDocumentList(documentList)
|
|
|
|
# Step 2: Clarify document intents (if not provided) - REQUIRED for all documents
|
|
if not documentIntents and documents:
|
|
documentIntents = await self.clarifyDocumentIntents(
|
|
documents,
|
|
prompt,
|
|
{"outputFormat": outputFormat},
|
|
extractOperationId
|
|
)
|
|
|
|
# Step 3: Extract and prepare content (NO AI - pure extraction) - REQUIRED for all documents
|
|
if documents:
|
|
preparedContentParts = await self.extractAndPrepareContent(
|
|
documents,
|
|
documentIntents or [],
|
|
extractOperationId
|
|
)
|
|
|
|
# Merge with provided contentParts (if any)
|
|
if contentParts:
|
|
for part in contentParts:
|
|
if part.metadata.get("skipExtraction", False):
|
|
part.metadata.setdefault("contentFormat", "extracted")
|
|
part.metadata.setdefault("isPreExtracted", True)
|
|
preparedContentParts.extend(contentParts)
|
|
|
|
contentParts = preparedContentParts
|
|
|
|
# Step 4: Process extracted contentParts with AI (simple text processing, no structure generation)
|
|
if not contentParts:
|
|
raise ValueError("No content extracted from documents")
|
|
|
|
# Use simple AI call to process extracted content
|
|
# Prepare content for AI processing
|
|
contentText = "\n\n".join([
|
|
f"[Document: {part.metadata.get('documentName', 'Unknown')}]\n{part.data}"
|
|
for part in contentParts
|
|
if part.data
|
|
])
|
|
|
|
# Call AI with extracted content
|
|
aiRequest = AiCallRequest(
|
|
prompt=f"{prompt}\n\nExtracted Content:\n{contentText}",
|
|
context="",
|
|
options=options
|
|
)
|
|
|
|
aiResponse = await self.callAi(aiRequest)
|
|
|
|
# Create response document
|
|
resultDocument = DocumentData(
|
|
documentName=f"{title or 'extracted_data'}.{outputFormat}",
|
|
documentData=aiResponse.content.encode('utf-8') if isinstance(aiResponse.content, str) else aiResponse.content,
|
|
mimeType=f"text/{outputFormat}" if outputFormat in ["txt", "json", "csv"] else "application/octet-stream"
|
|
)
|
|
|
|
metadata = AiResponseMetadata(
|
|
title=title or "Extracted Data",
|
|
operationType=OperationTypeEnum.DATA_EXTRACT.value
|
|
)
|
|
|
|
self.services.chat.progressLogFinish(extractOperationId, True)
|
|
|
|
return AiResponse(
|
|
content=aiResponse.content if isinstance(aiResponse.content, str) else aiResponse.content.decode('utf-8', errors='replace'),
|
|
metadata=metadata,
|
|
documents=[resultDocument]
|
|
)
|
|
|
|
except Exception as e:
|
|
logger.error(f"Error in data extraction: {str(e)}")
|
|
self.services.chat.progressLogFinish(extractOperationId, False)
|
|
raise
|
|
|
|
async def _handleCodeGeneration(
|
|
self,
|
|
prompt: str,
|
|
options: AiCallOptions,
|
|
contentParts: Optional[List[ContentPart]],
|
|
outputFormat: str,
|
|
title: str,
|
|
parentOperationId: Optional[str]
|
|
) -> AiResponse:
|
|
"""Handle code generation using code generation path."""
|
|
from modules.services.serviceGeneration.paths.codePath import CodeGenerationPath
|
|
|
|
codePath = CodeGenerationPath(self.services)
|
|
return await codePath.generateCode(
|
|
userPrompt=prompt,
|
|
outputFormat=outputFormat,
|
|
contentParts=contentParts,
|
|
title=title or "Generated Code",
|
|
parentOperationId=parentOperationId
|
|
)
|
|
|
|
async def _handleDocumentGeneration(
|
|
self,
|
|
prompt: str,
|
|
options: AiCallOptions,
|
|
documentList: Optional[Any],
|
|
documentIntents: Optional[List[DocumentIntent]],
|
|
contentParts: Optional[List[ContentPart]],
|
|
outputFormat: str,
|
|
title: str,
|
|
parentOperationId: Optional[str]
|
|
) -> AiResponse:
|
|
"""Handle document generation using document generation path."""
|
|
from modules.services.serviceGeneration.paths.documentPath import DocumentGenerationPath
|
|
|
|
# Set compression options for document generation
|
|
options.compressPrompt = False
|
|
options.compressContext = False
|
|
|
|
documentPath = DocumentGenerationPath(self.services)
|
|
return await documentPath.generateDocument(
|
|
userPrompt=prompt,
|
|
documentList=documentList,
|
|
documentIntents=documentIntents,
|
|
contentParts=contentParts,
|
|
outputFormat=outputFormat,
|
|
title=title or "Generated Document",
|
|
parentOperationId=parentOperationId
|
|
)
|
|
|
|
|
|
def _determineDocumentName(
|
|
self,
|
|
filledStructure: Dict[str, Any],
|
|
outputFormat: str,
|
|
title: Optional[str]
|
|
) -> str:
|
|
"""Bestimme Dokument-Namen aus Struktur oder Titel."""
|
|
# Versuche aus Struktur zu extrahieren
|
|
if isinstance(filledStructure, dict) and "documents" in filledStructure:
|
|
docs = filledStructure["documents"]
|
|
if isinstance(docs, list) and len(docs) > 0:
|
|
firstDoc = docs[0]
|
|
if isinstance(firstDoc, dict) and firstDoc.get("filename"):
|
|
return firstDoc["filename"]
|
|
|
|
# Fallback zu Titel
|
|
if title:
|
|
sanitized = re.sub(r"[^a-zA-Z0-9._-]", "_", title)
|
|
sanitized = re.sub(r"_+", "_", sanitized).strip("_")
|
|
if sanitized:
|
|
if not sanitized.lower().endswith(f".{outputFormat}"):
|
|
return f"{sanitized}.{outputFormat}"
|
|
return sanitized
|
|
|
|
return f"generated.{outputFormat}"
|
|
|