# Copyright (c) 2025 Patrick Motsch # All rights reserved. """ Structure Filling Module Handles filling document structure with content, including: - Filling sections with content parts - Building section generation prompts - Aggregation logic """ import json import logging import copy import asyncio from typing import Dict, Any, List, Optional from modules.datamodels.datamodelExtraction import ContentPart from modules.datamodels.datamodelAi import AiCallRequest, AiCallOptions, OperationTypeEnum, PriorityEnum, ProcessingModeEnum logger = logging.getLogger(__name__) class StructureFiller: """Handles filling document structure with content.""" def __init__(self, services, aiService): """Initialize StructureFiller with service center and AI service access.""" self.services = services self.aiService = aiService def _getUserLanguage(self) -> str: """Get user language for document generation""" try: if self.services: # Prefer detected language if available (from user intention analysis) if hasattr(self.services, 'currentUserLanguage') and self.services.currentUserLanguage: return self.services.currentUserLanguage # Fallback to user's preferred language elif hasattr(self.services, 'user') and self.services.user and hasattr(self.services.user, 'language'): return self.services.user.language except Exception: pass return 'en' # Default fallback async def fillStructure( self, structure: Dict[str, Any], contentParts: List[ContentPart], userPrompt: str, parentOperationId: str, language: Optional[str] = None ) -> Dict[str, Any]: """ Phase 5D: Chapter-Content-Generierung (Zwei-Phasen-Ansatz). Phase 5D.1: Generiert Sections-Struktur für jedes Chapter Phase 5D.2: Füllt Sections mit ContentParts Args: structure: Struktur-Dict mit documents und chapters (nicht sections!) contentParts: Alle vorbereiteten ContentParts userPrompt: User-Anfrage parentOperationId: Parent Operation-ID für ChatLog-Hierarchie language: Language identified from user intention analysis (e.g., "de", "en", "fr") Returns: Gefüllte Struktur mit elements in jeder Section (nach Flattening) """ # Erstelle Operation-ID für Struktur-Abfüllen fillOperationId = f"{parentOperationId}_structure_filling" # Validate structure has chapters hasChapters = False for doc in structure.get("documents", []): if "chapters" in doc: hasChapters = True break if not hasChapters: error_msg = "Structure must have chapters. Legacy section-based structure is not supported." logger.error(error_msg) raise ValueError(error_msg) # Get language from services (user intention analysis) or parameter if language is None: language = self._getUserLanguage() logger.debug(f"Using language from services (user intention analysis): {language}") else: logger.debug(f"Using provided language parameter: {language}") # Starte ChatLog mit Parent-Referenz chapterCount = sum(len(doc.get("chapters", [])) for doc in structure.get("documents", [])) self.services.chat.progressLogStart( fillOperationId, "Chapter Content Generation", "Filling", f"Processing {chapterCount} chapters", parentOperationId=parentOperationId ) try: filledStructure = copy.deepcopy(structure) # Phase 5D.1: Sections-Struktur für jedes Chapter generieren filledStructure = await self._generateChapterSectionsStructure( filledStructure, contentParts, userPrompt, fillOperationId, language ) # Phase 5D.2: Sections mit ContentParts füllen filledStructure = await self._fillChapterSections( filledStructure, contentParts, userPrompt, fillOperationId, language ) # Flattening: Chapters zu Sections konvertieren flattenedStructure = self._flattenChaptersToSections(filledStructure) # Füge ContentParts-Metadaten zur Struktur hinzu (für Validierung) flattenedStructure = self._addContentPartsMetadata(flattenedStructure, contentParts) # ChatLog abschließen self.services.chat.progressLogFinish(fillOperationId, True) return flattenedStructure except Exception as e: self.services.chat.progressLogFinish(fillOperationId, False) logger.error(f"Error in fillStructure: {str(e)}") raise async def _generateSingleChapterSectionsStructure( self, chapter: Dict[str, Any], chapterIndex: int, chapterId: str, chapterLevel: int, chapterTitle: str, generationHint: str, contentPartIds: List[str], contentPartInstructions: Dict[str, Any], contentParts: List[ContentPart], userPrompt: str, language: str, parentOperationId: str, totalChapters: int ) -> None: """ Generate sections structure for a single chapter (used for parallel processing). Modifies chapter dict in place. """ try: # Update progress for chapter structure generation progress = chapterIndex / totalChapters if totalChapters > 0 else 1.0 self.services.chat.progressLogUpdate( parentOperationId, progress, f"Generating sections for Chapter {chapterIndex}/{totalChapters}: {chapterTitle}" ) chapterPrompt = self._buildChapterSectionsStructurePrompt( chapterId=chapterId, chapterLevel=chapterLevel, chapterTitle=chapterTitle, generationHint=generationHint, contentPartIds=contentPartIds, contentPartInstructions=contentPartInstructions, contentParts=contentParts, userPrompt=userPrompt, language=language ) # AI-Call für Chapter-Struktur-Generierung # Note: Debug logging is handled by callAiPlanning aiResponse = await self.aiService.callAiPlanning( prompt=chapterPrompt, debugType=f"chapter_structure_{chapterId}" ) sectionsStructure = json.loads( self.services.utils.jsonExtractString(aiResponse) ) chapter["sections"] = sectionsStructure.get("sections", []) # Setze useAiCall Flag (falls nicht von AI gesetzt) # WICHTIG: useAiCall kann nur true sein, wenn mindestens ein ContentPart Format "extracted" hat! # "object" und "reference" Formate werden direkt als Elemente hinzugefügt, benötigen kein AI. for section in chapter["sections"]: if "useAiCall" not in section: contentType = section.get("content_type", "paragraph") sectionContentPartIds = section.get("contentPartIds", []) # Prüfe ob mindestens ein ContentPart Format "extracted" hat hasExtractedPart = False for partId in sectionContentPartIds: part = self._findContentPartById(partId, contentParts) if part: contentFormat = part.metadata.get("contentFormat", "unknown") if contentFormat == "extracted": hasExtractedPart = True break # useAiCall kann nur true sein, wenn extracted Parts vorhanden sind useAiCall = False if hasExtractedPart: # Prüfe ob Transformation nötig ist useAiCall = contentType != "paragraph" # Prüfe contentPartInstructions für Transformation if not useAiCall: for partId in sectionContentPartIds: instruction = contentPartInstructions.get(partId, {}).get("instruction", "") if instruction and instruction.lower() not in ["include full text", "include all content", "use full extracted text"]: useAiCall = True break section["useAiCall"] = useAiCall logger.debug(f"Section {section.get('id')}: useAiCall={useAiCall} (hasExtractedPart={hasExtractedPart}, contentType={contentType})") # Update progress after chapter completion progress = chapterIndex / totalChapters if totalChapters > 0 else 1.0 self.services.chat.progressLogUpdate( parentOperationId, progress, f"Chapter {chapterIndex}/{totalChapters} completed: {chapterTitle}" ) except Exception as e: logger.error(f"Error generating sections structure for chapter {chapterId}: {str(e)}") # Set empty sections on error chapter["sections"] = [] # Update progress even on error progress = chapterIndex / totalChapters if totalChapters > 0 else 1.0 self.services.chat.progressLogUpdate( parentOperationId, progress, f"Chapter {chapterIndex}/{totalChapters} error: {chapterTitle}" ) raise async def _generateChapterSectionsStructure( self, chapterStructure: Dict[str, Any], contentParts: List[ContentPart], userPrompt: str, parentOperationId: str, language: str ) -> Dict[str, Any]: """ Phase 5D.1: Generiert Sections-Struktur für jedes Chapter (ohne Content) in parallel. Sections enthalten: content_type, contentPartIds, generationHint, useAiCall """ # Count total chapters for progress tracking totalChapters = sum(len(doc.get("chapters", [])) for doc in chapterStructure.get("documents", [])) # Collect all chapters with their indices for parallel processing chapterTasks = [] chapterIndex = 0 for doc in chapterStructure.get("documents", []): for chapter in doc.get("chapters", []): chapterIndex += 1 chapterId = chapter.get("id", "unknown") chapterLevel = chapter.get("level", 1) chapterTitle = chapter.get("title", "Untitled Chapter") generationHint = chapter.get("generationHint", "") contentPartIds = chapter.get("contentPartIds", []) contentPartInstructions = chapter.get("contentPartInstructions", {}) # Create task for parallel processing task = self._generateSingleChapterSectionsStructure( chapter=chapter, chapterIndex=chapterIndex, chapterId=chapterId, chapterLevel=chapterLevel, chapterTitle=chapterTitle, generationHint=generationHint, contentPartIds=contentPartIds, contentPartInstructions=contentPartInstructions, contentParts=contentParts, userPrompt=userPrompt, language=language, parentOperationId=parentOperationId, totalChapters=totalChapters ) chapterTasks.append((chapterIndex, chapter, task)) # Execute all chapter tasks in parallel if chapterTasks: # Create list of tasks (without indices for gather) tasks = [task for _, _, task in chapterTasks] # Execute in parallel with error handling results = await asyncio.gather(*tasks, return_exceptions=True) # Process results in order and handle errors for (originalIndex, originalChapter, _), result in zip(chapterTasks, results): if isinstance(result, Exception): logger.error(f"Error processing chapter {originalChapter.get('id')}: {str(result)}") # Chapter already has empty sections set by _generateSingleChapterSectionsStructure # Continue processing other chapters return chapterStructure async def _processAiResponseForSection( self, aiResponse: Any, contentType: str, operationType: OperationTypeEnum, sectionId: str, generationHint: str, generatedElements: List[Dict[str, Any]] ) -> List[Dict[str, Any]]: """ Helper method to process AI response and extract elements. Handles both IMAGE_GENERATE and DATA_ANALYSE operation types. """ elements = [] # Handle IMAGE_GENERATE differently - returns image data directly if contentType == "image" and operationType == OperationTypeEnum.IMAGE_GENERATE: import base64 base64Data = "" # Convert image data to base64 string if needed if isinstance(aiResponse.content, bytes): base64Data = base64.b64encode(aiResponse.content).decode('utf-8') elif isinstance(aiResponse.content, str): # Check if it's already a JSON structure try: jsonContent = json.loads(self.services.utils.jsonExtractString(aiResponse.content)) if isinstance(jsonContent, dict) and jsonContent.get("type") == "image": elements.append(jsonContent) logger.debug("AI returned proper JSON image structure") base64Data = None # Signal that image was already processed elif isinstance(jsonContent, list) and len(jsonContent) > 0: if isinstance(jsonContent[0], dict) and jsonContent[0].get("type") == "image": elements.extend(jsonContent) logger.debug("AI returned proper JSON image structure in list") base64Data = None # Signal that image was already processed else: base64Data = "" # Continue with normal processing else: base64Data = "" # Continue with normal processing except (json.JSONDecodeError, ValueError, AttributeError): base64Data = "" # Will be processed below # Process base64 if not already handled above if base64Data is None: # Already processed as JSON, skip base64 processing pass elif aiResponse.content.startswith("data:image/"): # Extract base64 from data URI base64Data = aiResponse.content.split(",", 1)[1] else: content_stripped = aiResponse.content.strip() if len(content_stripped) > 100 and all(c in "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/=\n\r\t " for c in content_stripped[:200]): base64Data = content_stripped.replace("\n", "").replace("\r", "").replace("\t", "").replace(" ", "") else: base64Data = aiResponse.content else: base64Data = "" # Always create proper JSON structure for images (if not already processed) if base64Data is None: # Image already processed as JSON, skip pass elif base64Data: elements.append({ "type": "image", "content": { "base64Data": base64Data, "altText": generationHint or "Generated image", "caption": "" } }) logger.debug(f"Created proper JSON image structure with base64Data length: {len(base64Data)}") else: logger.warning(f"IMAGE_GENERATE returned empty or invalid content for section {sectionId}") elements.append({ "type": "error", "message": f"Image generation returned empty or invalid content", "sectionId": sectionId }) else: # For non-image content: Use already parsed elements from _callAiWithLooping if generatedElements: elements.extend(generatedElements) else: # Fallback: Try to parse JSON response directly try: fallbackElements = json.loads( self.services.utils.jsonExtractString(aiResponse.content) ) if isinstance(fallbackElements, list): elements.extend(fallbackElements) elif isinstance(fallbackElements, dict) and "elements" in fallbackElements: elements.extend(fallbackElements["elements"]) elif isinstance(fallbackElements, dict) and fallbackElements.get("type"): elements.append(fallbackElements) except (json.JSONDecodeError, ValueError) as json_error: logger.error(f"Error parsing JSON response for section {sectionId}: {str(json_error)}") elements.append({ "type": "error", "message": f"Failed to parse JSON response: {str(json_error)}", "sectionId": sectionId }) return elements async def _processSingleSection( self, section: Dict[str, Any], sectionIndex: int, totalSections: int, chapterIndex: int, totalChapters: int, chapterId: str, chapterOperationId: str, fillOperationId: str, contentParts: List[ContentPart], userPrompt: str, all_sections_list: List[Dict[str, Any]], language: str, calculateOverallProgress: callable ) -> List[Dict[str, Any]]: """ Process a single section and return its elements. Used for parallel processing of sections within a chapter. """ sectionId = section.get("id") sectionTitle = section.get("title", sectionId) contentPartIds = section.get("contentPartIds", []) contentFormats = section.get("contentFormats", {}) generationHint = section.get("generationHint") or section.get("generation_hint") contentType = section.get("content_type", "paragraph") useAiCall = section.get("useAiCall", False) # Update overall progress at start of section overallProgress = calculateOverallProgress(chapterIndex - 1, totalChapters, sectionIndex, totalSections) self.services.chat.progressLogUpdate( fillOperationId, overallProgress, f"Chapter {chapterIndex}/{totalChapters}, Section {sectionIndex + 1}/{totalSections}: {sectionTitle}" ) # WICHTIG: Wenn keine ContentParts vorhanden sind UND kein generationHint, kann kein AI-Call gemacht werden if len(contentPartIds) == 0 and not generationHint: useAiCall = False logger.debug(f"Section {sectionId}: No content parts and no generation hint, setting useAiCall=False") elif len(contentPartIds) == 0 and generationHint and not useAiCall: useAiCall = True logger.info(f"Section {sectionId}: Overriding useAiCall=True (has generationHint but no content parts)") elements = [] # Prüfe ob Aggregation nötig ist needsAggregation = self._needsAggregation( contentType=contentType, contentPartCount=len(contentPartIds) ) logger.info(f"Processing section {sectionId}: contentType={contentType}, contentPartCount={len(contentPartIds)}, useAiCall={useAiCall}, needsAggregation={needsAggregation}, hasGenerationHint={bool(generationHint)}") try: if needsAggregation and useAiCall: # Aggregation: Alle Parts zusammen verarbeiten sectionParts = [ self._findContentPartById(pid, contentParts) for pid in contentPartIds ] sectionParts = [p for p in sectionParts if p is not None] if sectionParts: # Filtere nur extracted Parts für Aggregation (reference/object werden separat behandelt) extractedParts = [ p for p in sectionParts if contentFormats.get(p.id, p.metadata.get("contentFormat")) == "extracted" ] nonExtractedParts = [ p for p in sectionParts if contentFormats.get(p.id, p.metadata.get("contentFormat")) != "extracted" ] # Verarbeite non-extracted Parts separat (reference, object) for part in nonExtractedParts: contentFormat = contentFormats.get(part.id, part.metadata.get("contentFormat")) if contentFormat == "reference": elements.append({ "type": "reference", "documentReference": part.metadata.get("documentReference"), "label": part.metadata.get("usageHint", part.label) }) elif contentFormat == "object": if part.typeGroup == "image": elements.append({ "type": "image", "content": { "base64Data": part.data, "altText": part.metadata.get("usageHint", part.label), "caption": part.metadata.get("caption", "") } }) else: elements.append({ "type": part.typeGroup, "content": { "data": part.data, "mimeType": part.mimeType, "label": part.metadata.get("usageHint", part.label) } }) # Aggregiere extracted Parts mit AI if extractedParts: logger.debug(f"Section {sectionId}: Aggregating {len(extractedParts)} extracted parts with AI") isAggregation = True generationPrompt = self._buildSectionGenerationPrompt( section=section, contentParts=extractedParts, userPrompt=userPrompt, generationHint=generationHint, allSections=all_sections_list, sectionIndex=sectionIndex, isAggregation=isAggregation, language=language ) sectionOperationId = f"{fillOperationId}_section_{sectionId}" self.services.chat.progressLogStart( sectionOperationId, "Section Generation (Aggregation)", f"Section {sectionIndex + 1}/{totalSections}", f"{sectionTitle} ({len(extractedParts)} parts)", parentOperationId=chapterOperationId ) try: self.services.chat.progressLogUpdate(sectionOperationId, 0.2, "Building generation prompt") self.services.chat.progressLogUpdate(sectionOperationId, 0.4, "Calling AI for content generation") operationType = OperationTypeEnum.IMAGE_GENERATE if contentType == "image" else OperationTypeEnum.DATA_ANALYSE if operationType == OperationTypeEnum.IMAGE_GENERATE: maxPromptLength = 4000 if len(generationPrompt) > maxPromptLength: logger.warning(f"Truncating DALL-E prompt from {len(generationPrompt)} to {maxPromptLength} characters") generationPrompt = generationPrompt[:maxPromptLength].rsplit('\n', 1)[0] # Write debug file for IMAGE_GENERATE (direct callAi, no _callAiWithLooping) self.services.utils.writeDebugFile( generationPrompt, f"{chapterId}_section_{sectionId}_prompt" ) request = AiCallRequest( prompt=generationPrompt, contentParts=[], options=AiCallOptions( operationType=operationType, priority=PriorityEnum.BALANCED, processingMode=ProcessingModeEnum.DETAILED ) ) aiResponse = await self.aiService.callAi(request) generatedElements = [] # Write debug file for IMAGE_GENERATE response (direct callAi, no _callAiWithLooping) self.services.utils.writeDebugFile( aiResponse.content if hasattr(aiResponse, 'content') else str(aiResponse), f"{chapterId}_section_{sectionId}_response" ) else: async def buildSectionPromptWithContinuation( section: Dict[str, Any], contentParts: List[ContentPart], userPrompt: str, generationHint: str, allSections: List[Dict[str, Any]], sectionIndex: int, isAggregation: bool, continuationContext: Dict[str, Any], services: Any ) -> str: basePrompt = self._buildSectionGenerationPrompt( section=section, contentParts=contentParts, userPrompt=userPrompt, generationHint=generationHint, allSections=allSections, sectionIndex=sectionIndex, isAggregation=isAggregation, language=language ) continuationInfo = continuationContext.get("delivered_summary", "") cutOffElement = continuationContext.get("cut_off_element", "") continuationPrompt = f"""{basePrompt} --- CONTINUATION REQUEST --- The previous JSON response was incomplete. Please continue from where it stopped. PREVIOUSLY DELIVERED SUMMARY: {continuationInfo} LAST INCOMPLETE ELEMENT: {cutOffElement} TASK: Continue generating the JSON elements array from where it was cut off. Complete the incomplete element and continue with remaining elements. Return ONLY the continuation JSON (starting from the incomplete element). The JSON should be a fragment that can be merged with the previous response.""" return continuationPrompt options = AiCallOptions( operationType=operationType, priority=PriorityEnum.BALANCED, processingMode=ProcessingModeEnum.DETAILED ) aiResponseJson = await self.aiService._callAiWithLooping( prompt=generationPrompt, options=options, debugPrefix=f"{chapterId}_section_{sectionId}", promptBuilder=buildSectionPromptWithContinuation, promptArgs={ "section": section, "contentParts": extractedParts, "userPrompt": userPrompt, "generationHint": generationHint, "allSections": all_sections_list, "sectionIndex": sectionIndex, "isAggregation": isAggregation, "services": self.services }, operationId=sectionOperationId, userPrompt=userPrompt, contentParts=extractedParts ) try: parsedResponse = json.loads(self.services.utils.jsonExtractString(aiResponseJson)) if isinstance(parsedResponse, list): generatedElements = parsedResponse elif isinstance(parsedResponse, dict): if "elements" in parsedResponse: generatedElements = parsedResponse["elements"] elif "sections" in parsedResponse and len(parsedResponse["sections"]) > 0: firstSection = parsedResponse["sections"][0] generatedElements = firstSection.get("elements", []) elif parsedResponse.get("type"): generatedElements = [parsedResponse] else: generatedElements = [] else: generatedElements = [] class AiResponse: def __init__(self, content): self.content = content aiResponse = AiResponse(aiResponseJson) except Exception as parseError: logger.error(f"Error parsing response from _callAiWithLooping for section {sectionId}: {str(parseError)}") class AiResponse: def __init__(self, content): self.content = content aiResponse = AiResponse(aiResponseJson) generatedElements = [] self.services.chat.progressLogUpdate(sectionOperationId, 0.6, "Processing AI response") # Note: Debug files are written by _callAiWithLooping using debugPrefix self.services.chat.progressLogUpdate(sectionOperationId, 0.8, "Validating generated content") # Process AI response responseElements = await self._processAiResponseForSection( aiResponse=aiResponse, contentType=contentType, operationType=operationType, sectionId=sectionId, generationHint=generationHint, generatedElements=generatedElements ) elements.extend(responseElements) self.services.chat.progressLogFinish(sectionOperationId, True) chapterProgress = (sectionIndex + 1) / totalSections if totalSections > 0 else 1.0 self.services.chat.progressLogUpdate( chapterOperationId, chapterProgress, f"Section {sectionIndex + 1}/{totalSections} completed" ) except Exception as e: self.services.chat.progressLogFinish(sectionOperationId, False) elements.append({ "type": "error", "message": f"Error generating section {sectionId}: {str(e)}", "sectionId": sectionId }) logger.error(f"Error generating section {sectionId}: {str(e)}") chapterProgress = (sectionIndex + 1) / totalSections if totalSections > 0 else 1.0 self.services.chat.progressLogUpdate( chapterOperationId, chapterProgress, f"Section {sectionIndex + 1}/{totalSections} completed (with errors)" ) else: # Einzelverarbeitung: Jeder Part einzeln ODER Generation ohne ContentParts if len(contentPartIds) == 0 and useAiCall and generationHint: # Generate content from scratch using only generationHint logger.debug(f"Processing section {sectionId}: No content parts, generating from generationHint only") generationPrompt = self._buildSectionGenerationPrompt( section=section, contentParts=[], userPrompt=userPrompt, generationHint=generationHint, allSections=all_sections_list, sectionIndex=sectionIndex, isAggregation=False, language=language ) sectionOperationId = f"{fillOperationId}_section_{sectionId}" self.services.chat.progressLogStart( sectionOperationId, "Section Generation", f"Section {sectionIndex + 1}/{totalSections}", f"{sectionTitle} (from generationHint)", parentOperationId=chapterOperationId ) try: self.services.chat.progressLogUpdate(sectionOperationId, 0.2, "Building generation prompt") self.services.chat.progressLogUpdate(sectionOperationId, 0.4, "Calling AI for content generation") operationType = OperationTypeEnum.IMAGE_GENERATE if contentType == "image" else OperationTypeEnum.DATA_ANALYSE if operationType == OperationTypeEnum.IMAGE_GENERATE: maxPromptLength = 4000 if len(generationPrompt) > maxPromptLength: logger.warning(f"Truncating DALL-E prompt from {len(generationPrompt)} to {maxPromptLength} characters") generationPrompt = generationPrompt[:maxPromptLength].rsplit('\n', 1)[0] # Write debug file for IMAGE_GENERATE (direct callAi, no _callAiWithLooping) self.services.utils.writeDebugFile( generationPrompt, f"{chapterId}_section_{sectionId}_prompt" ) request = AiCallRequest( prompt=generationPrompt, contentParts=[], options=AiCallOptions( operationType=operationType, priority=PriorityEnum.BALANCED, processingMode=ProcessingModeEnum.DETAILED ) ) aiResponse = await self.aiService.callAi(request) generatedElements = [] # Write debug file for IMAGE_GENERATE response (direct callAi, no _callAiWithLooping) self.services.utils.writeDebugFile( aiResponse.content if hasattr(aiResponse, 'content') else str(aiResponse), f"{chapterId}_section_{sectionId}_response" ) else: isAggregation = False async def buildSectionPromptWithContinuation( section: Dict[str, Any], contentParts: List[ContentPart], userPrompt: str, generationHint: str, allSections: List[Dict[str, Any]], sectionIndex: int, isAggregation: bool, continuationContext: Dict[str, Any], services: Any ) -> str: basePrompt = self._buildSectionGenerationPrompt( section=section, contentParts=contentParts, userPrompt=userPrompt, generationHint=generationHint, allSections=allSections, sectionIndex=sectionIndex, isAggregation=isAggregation, language=language ) continuationInfo = continuationContext.get("delivered_summary", "") cutOffElement = continuationContext.get("cut_off_element", "") continuationPrompt = f"""{basePrompt} --- CONTINUATION REQUEST --- The previous JSON response was incomplete. Please continue from where it stopped. PREVIOUSLY DELIVERED SUMMARY: {continuationInfo} LAST INCOMPLETE ELEMENT: {cutOffElement} TASK: Continue generating the JSON elements array from where it was cut off. Complete the incomplete element and continue with remaining elements. Return ONLY the continuation JSON (starting from the incomplete element). The JSON should be a fragment that can be merged with the previous response.""" return continuationPrompt options = AiCallOptions( operationType=operationType, priority=PriorityEnum.BALANCED, processingMode=ProcessingModeEnum.DETAILED ) aiResponseJson = await self.aiService._callAiWithLooping( prompt=generationPrompt, options=options, debugPrefix=f"{chapterId}_section_{sectionId}", promptBuilder=buildSectionPromptWithContinuation, promptArgs={ "section": section, "contentParts": [], "userPrompt": userPrompt, "generationHint": generationHint, "allSections": all_sections_list, "sectionIndex": sectionIndex, "isAggregation": isAggregation, "services": self.services }, operationId=sectionOperationId, userPrompt=userPrompt, contentParts=[] ) try: parsedResponse = json.loads(self.services.utils.jsonExtractString(aiResponseJson)) if isinstance(parsedResponse, list): generatedElements = parsedResponse elif isinstance(parsedResponse, dict): if "elements" in parsedResponse: generatedElements = parsedResponse["elements"] elif "sections" in parsedResponse and len(parsedResponse["sections"]) > 0: firstSection = parsedResponse["sections"][0] generatedElements = firstSection.get("elements", []) elif parsedResponse.get("type"): generatedElements = [parsedResponse] else: generatedElements = [] else: generatedElements = [] class AiResponse: def __init__(self, content): self.content = content aiResponse = AiResponse(aiResponseJson) except Exception as parseError: logger.error(f"Error parsing response from _callAiWithLooping for section {sectionId}: {str(parseError)}") class AiResponse: def __init__(self, content): self.content = content aiResponse = AiResponse(aiResponseJson) generatedElements = [] self.services.chat.progressLogUpdate(sectionOperationId, 0.6, "Processing AI response") # Note: Debug files are written by _callAiWithLooping using debugPrefix self.services.chat.progressLogUpdate(sectionOperationId, 0.8, "Validating generated content") responseElements = await self._processAiResponseForSection( aiResponse=aiResponse, contentType=contentType, operationType=operationType, sectionId=sectionId, generationHint=generationHint, generatedElements=generatedElements ) elements.extend(responseElements) self.services.chat.progressLogFinish(sectionOperationId, True) chapterProgress = (sectionIndex + 1) / totalSections if totalSections > 0 else 1.0 self.services.chat.progressLogUpdate( chapterOperationId, chapterProgress, f"Section {sectionIndex + 1}/{totalSections} completed" ) except Exception as e: self.services.chat.progressLogFinish(sectionOperationId, False) elements.append({ "type": "error", "message": f"Error generating section {sectionId}: {str(e)}", "sectionId": sectionId }) logger.error(f"Error generating section {sectionId}: {str(e)}") chapterProgress = (sectionIndex + 1) / totalSections if totalSections > 0 else 1.0 self.services.chat.progressLogUpdate( chapterOperationId, chapterProgress, f"Section {sectionIndex + 1}/{totalSections} completed (with errors)" ) # Einzelverarbeitung: Jeder Part einzeln for partId in contentPartIds: part = self._findContentPartById(partId, contentParts) if not part: continue contentFormat = contentFormats.get(partId, part.metadata.get("contentFormat")) if contentFormat == "reference": elements.append({ "type": "reference", "documentReference": part.metadata.get("documentReference"), "label": part.metadata.get("usageHint", part.label) }) elif contentFormat == "object": if part.typeGroup == "image": elements.append({ "type": "image", "content": { "base64Data": part.data, "altText": part.metadata.get("usageHint", part.label), "caption": part.metadata.get("caption", "") } }) else: elements.append({ "type": part.typeGroup, "content": { "data": part.data, "mimeType": part.mimeType, "label": part.metadata.get("usageHint", part.label) } }) elif contentFormat == "extracted": if useAiCall and generationHint: # AI-Call mit einzelnen ContentPart logger.debug(f"Processing section {sectionId}: Single extracted part with AI call") generationPrompt = self._buildSectionGenerationPrompt( section=section, contentParts=[part], userPrompt=userPrompt, generationHint=generationHint, allSections=all_sections_list, sectionIndex=sectionIndex, isAggregation=False, language=language ) sectionOperationId = f"{fillOperationId}_section_{sectionId}" self.services.chat.progressLogStart( sectionOperationId, "Section Generation", f"Section {sectionIndex + 1}/{totalSections}", f"{sectionTitle} (single part)", parentOperationId=chapterOperationId ) try: self.services.chat.progressLogUpdate(sectionOperationId, 0.2, "Building generation prompt") self.services.chat.progressLogUpdate(sectionOperationId, 0.4, "Calling AI for content generation") operationType = OperationTypeEnum.IMAGE_GENERATE if contentType == "image" else OperationTypeEnum.DATA_ANALYSE if operationType == OperationTypeEnum.IMAGE_GENERATE: maxPromptLength = 4000 if len(generationPrompt) > maxPromptLength: logger.warning(f"Truncating DALL-E prompt from {len(generationPrompt)} to {maxPromptLength} characters") generationPrompt = generationPrompt[:maxPromptLength].rsplit('\n', 1)[0] # Write debug file for IMAGE_GENERATE (direct callAi, no _callAiWithLooping) self.services.utils.writeDebugFile( generationPrompt, f"{chapterId}_section_{sectionId}_prompt" ) request = AiCallRequest( prompt=generationPrompt, contentParts=[], options=AiCallOptions( operationType=operationType, priority=PriorityEnum.BALANCED, processingMode=ProcessingModeEnum.DETAILED ) ) aiResponse = await self.aiService.callAi(request) generatedElements = [] # Write debug file for IMAGE_GENERATE response (direct callAi, no _callAiWithLooping) self.services.utils.writeDebugFile( aiResponse.content if hasattr(aiResponse, 'content') else str(aiResponse), f"{chapterId}_section_{sectionId}_response" ) else: isAggregation = False async def buildSectionPromptWithContinuation( section: Dict[str, Any], contentParts: List[ContentPart], userPrompt: str, generationHint: str, allSections: List[Dict[str, Any]], sectionIndex: int, isAggregation: bool, continuationContext: Dict[str, Any], services: Any ) -> str: basePrompt = self._buildSectionGenerationPrompt( section=section, contentParts=contentParts, userPrompt=userPrompt, generationHint=generationHint, allSections=allSections, sectionIndex=sectionIndex, isAggregation=isAggregation, language=language ) continuationInfo = continuationContext.get("delivered_summary", "") cutOffElement = continuationContext.get("cut_off_element", "") continuationPrompt = f"""{basePrompt} --- CONTINUATION REQUEST --- The previous JSON response was incomplete. Please continue from where it stopped. PREVIOUSLY DELIVERED SUMMARY: {continuationInfo} LAST INCOMPLETE ELEMENT: {cutOffElement} TASK: Continue generating the JSON elements array from where it was cut off. Complete the incomplete element and continue with remaining elements. Return ONLY the continuation JSON (starting from the incomplete element). The JSON should be a fragment that can be merged with the previous response.""" return continuationPrompt options = AiCallOptions( operationType=operationType, priority=PriorityEnum.BALANCED, processingMode=ProcessingModeEnum.DETAILED ) aiResponseJson = await self.aiService._callAiWithLooping( prompt=generationPrompt, options=options, debugPrefix=f"{chapterId}_section_{sectionId}", promptBuilder=buildSectionPromptWithContinuation, promptArgs={ "section": section, "contentParts": [part], "userPrompt": userPrompt, "generationHint": generationHint, "allSections": all_sections_list, "sectionIndex": sectionIndex, "isAggregation": isAggregation, "services": self.services }, operationId=sectionOperationId, userPrompt=userPrompt, contentParts=[part] ) try: parsedResponse = json.loads(self.services.utils.jsonExtractString(aiResponseJson)) if isinstance(parsedResponse, list): generatedElements = parsedResponse elif isinstance(parsedResponse, dict): if "elements" in parsedResponse: generatedElements = parsedResponse["elements"] elif "sections" in parsedResponse and len(parsedResponse["sections"]) > 0: firstSection = parsedResponse["sections"][0] generatedElements = firstSection.get("elements", []) elif parsedResponse.get("type"): generatedElements = [parsedResponse] else: generatedElements = [] else: generatedElements = [] class AiResponse: def __init__(self, content): self.content = content aiResponse = AiResponse(aiResponseJson) except Exception as parseError: logger.error(f"Error parsing response from _callAiWithLooping for section {sectionId}: {str(parseError)}") class AiResponse: def __init__(self, content): self.content = content aiResponse = AiResponse(aiResponseJson) generatedElements = [] self.services.chat.progressLogUpdate(sectionOperationId, 0.6, "Processing AI response") # Note: Debug files are written by _callAiWithLooping using debugPrefix self.services.chat.progressLogUpdate(sectionOperationId, 0.8, "Validating generated content") responseElements = await self._processAiResponseForSection( aiResponse=aiResponse, contentType=contentType, operationType=operationType, sectionId=sectionId, generationHint=generationHint, generatedElements=generatedElements ) elements.extend(responseElements) self.services.chat.progressLogFinish(sectionOperationId, True) chapterProgress = (sectionIndex + 1) / totalSections if totalSections > 0 else 1.0 self.services.chat.progressLogUpdate( chapterOperationId, chapterProgress, f"Section {sectionIndex + 1}/{totalSections} completed" ) except Exception as e: self.services.chat.progressLogFinish(sectionOperationId, False) elements.append({ "type": "error", "message": f"Error generating section {sectionId}: {str(e)}", "sectionId": sectionId }) logger.error(f"Error generating section {sectionId}: {str(e)}") chapterProgress = (sectionIndex + 1) / totalSections if totalSections > 0 else 1.0 self.services.chat.progressLogUpdate( chapterOperationId, chapterProgress, f"Section {sectionIndex + 1}/{totalSections} completed (with errors)" ) else: # Füge extrahierten Content direkt hinzu (kein AI-Call) if part.typeGroup == "image": logger.debug(f"Processing section {sectionId}: Single extracted IMAGE part WITHOUT AI call") elements.append({ "type": "image", "content": { "base64Data": part.data, "altText": part.metadata.get("usageHint", part.label), "caption": part.metadata.get("caption", "") } }) else: logger.debug(f"Processing section {sectionId}: Single extracted TEXT part WITHOUT AI call") elements.append({ "type": "extracted_text", "content": part.data, "source": part.metadata.get("documentId"), "extractionPrompt": part.metadata.get("extractionPrompt") }) # Update progress after section completion chapterProgress = (sectionIndex + 1) / totalSections if totalSections > 0 else 1.0 self.services.chat.progressLogUpdate( chapterOperationId, chapterProgress, f"Section {sectionIndex + 1}/{totalSections} completed" ) overallProgress = calculateOverallProgress(chapterIndex - 1, totalChapters, sectionIndex + 1, totalSections) self.services.chat.progressLogUpdate( fillOperationId, overallProgress, f"Chapter {chapterIndex}/{totalChapters}, Section {sectionIndex + 1}/{totalSections} completed" ) except Exception as e: logger.error(f"Unexpected error processing section {sectionId}: {str(e)}") elements.append({ "type": "error", "message": f"Unexpected error processing section {sectionId}: {str(e)}", "sectionId": sectionId }) return elements async def _fillChapterSections( self, chapterStructure: Dict[str, Any], contentParts: List[ContentPart], userPrompt: str, parentOperationId: str, language: str ) -> Dict[str, Any]: """ Phase 5D.2: Füllt Sections mit ContentParts. """ # Sammle alle Sections für Kontext-Informationen (für alle Sections) all_sections_list = [] for doc in chapterStructure.get("documents", []): for chapter in doc.get("chapters", []): for section in chapter.get("sections", []): all_sections_list.append(section) # Berechne Gesamtanzahl Chapters für Progress-Tracking totalChapters = sum(len(doc.get("chapters", [])) for doc in chapterStructure.get("documents", [])) fillOperationId = parentOperationId # Helper function to calculate overall progress def calculateOverallProgress(chapterIndex, totalChapters, sectionIndex, totalSections): """Calculate overall progress: 0.0 to 1.0""" if totalChapters == 0: return 1.0 # Progress from completed chapters (0 to chapterIndex-1) completedChaptersProgress = chapterIndex / totalChapters # Progress from current chapter (sectionIndex / totalSections) currentChapterProgress = (sectionIndex / totalSections) / totalChapters if totalSections > 0 else 0 return min(1.0, completedChaptersProgress + currentChapterProgress) # Process chapters sequentially with chapter-level progress chapterIndex = 0 for doc in chapterStructure.get("documents", []): for chapter in doc.get("chapters", []): chapterIndex += 1 chapterId = chapter.get("id", "unknown") chapterTitle = chapter.get("title", "Untitled Chapter") sections = chapter.get("sections", []) totalSections = len(sections) # Start chapter operation chapterOperationId = f"{fillOperationId}_chapter_{chapterId}" self.services.chat.progressLogStart( chapterOperationId, "Chapter Generation", f"Chapter {chapterIndex}/{totalChapters}", chapterTitle, parentOperationId=fillOperationId ) # Process sections within chapter in parallel sectionTasks = [] for sectionIndex, section in enumerate(sections): # Create task for parallel processing task = self._processSingleSection( section=section, sectionIndex=sectionIndex, totalSections=totalSections, chapterIndex=chapterIndex, totalChapters=totalChapters, chapterId=chapterId, chapterOperationId=chapterOperationId, fillOperationId=fillOperationId, contentParts=contentParts, userPrompt=userPrompt, all_sections_list=all_sections_list, language=language, calculateOverallProgress=calculateOverallProgress ) sectionTasks.append((sectionIndex, section, task)) # Execute all section tasks in parallel if sectionTasks: # Create list of tasks (without indices for gather) tasks = [task for _, _, task in sectionTasks] # Execute in parallel with error handling results = await asyncio.gather(*tasks, return_exceptions=True) # Process results in order and assign elements to sections for (originalIndex, originalSection, _), result in zip(sectionTasks, results): if isinstance(result, Exception): logger.error(f"Error processing section {originalSection.get('id')}: {str(result)}") # Set error element originalSection["elements"] = [{ "type": "error", "message": f"Error processing section: {str(result)}", "sectionId": originalSection.get("id") }] else: # Assign elements to section in correct order originalSection["elements"] = result # Finish chapter operation after all sections processed self.services.chat.progressLogFinish(chapterOperationId, True) # Update overall progress after chapter completion overallProgress = chapterIndex / totalChapters if totalChapters > 0 else 1.0 self.services.chat.progressLogUpdate( fillOperationId, overallProgress, f"Chapter {chapterIndex}/{totalChapters} completed: {chapterTitle}" ) return chapterStructure def _addContentPartsMetadata( self, structure: Dict[str, Any], contentParts: List[ContentPart] ) -> Dict[str, Any]: """ Fügt ContentParts-Metadaten zur Struktur hinzu, wenn contentPartIds vorhanden sind. Dies hilft der Validierung, den Kontext der ContentParts zu verstehen. """ # Erstelle Mapping von ContentPart-ID zu Metadaten contentPartsMap = {} for part in contentParts: contentPartsMap[part.id] = { "id": part.id, "format": part.metadata.get("contentFormat", "unknown"), "type": part.typeGroup, "mimeType": part.mimeType, "originalFileName": part.metadata.get("originalFileName"), "usageHint": part.metadata.get("usageHint"), "documentId": part.metadata.get("documentId"), "dataSize": len(str(part.data)) if part.data else 0 } # Füge Metadaten zu Sections hinzu, die contentPartIds haben for doc in structure.get("documents", []): # Prüfe ob Chapters vorhanden sind (neue Struktur) if "chapters" in doc: for chapter in doc.get("chapters", []): # Füge Metadaten zu Chapter-Level contentPartIds hinzu chapterContentPartIds = chapter.get("contentPartIds", []) if chapterContentPartIds: chapter["contentPartsMetadata"] = [] for partId in chapterContentPartIds: if partId in contentPartsMap: chapter["contentPartsMetadata"].append(contentPartsMap[partId]) # Füge Metadaten zu Sections hinzu for section in chapter.get("sections", []): contentPartIds = section.get("contentPartIds", []) if contentPartIds: section["contentPartsMetadata"] = [] for partId in contentPartIds: if partId in contentPartsMap: section["contentPartsMetadata"].append(contentPartsMap[partId]) return structure def _flattenChaptersToSections( self, chapterStructure: Dict[str, Any] ) -> Dict[str, Any]: """ Flattening: Konvertiert Chapters zu finaler Section-Struktur. Jedes Chapter wird zu einer Heading-Section (Level 1) + dessen Sections. IMPORTANT: Chapters are the main structure elements (heading level 1). All section headings with level < 2 are adjusted to level 2. """ result = { "metadata": chapterStructure.get("metadata", {}), "documents": [] } for doc in chapterStructure.get("documents", []): flattened_doc = { "id": doc.get("id"), "title": doc.get("title"), "filename": doc.get("filename"), "sections": [] } for chapter in doc.get("chapters", []): # 1. Vordefinierte Heading-Section für Chapter-Title (ALWAYS Level 1) heading_section = { "id": f"{chapter['id']}_heading", "content_type": "heading", "elements": [{ "type": "heading", "content": { "text": chapter.get("title", ""), "level": 1 # Chapters are always level 1 } }] } flattened_doc["sections"].append(heading_section) # 2. Generierte Sections - adjust heading levels for section in chapter.get("sections", []): adjusted_section = self._adjustSectionHeadingLevels(section) flattened_doc["sections"].append(adjusted_section) result["documents"].append(flattened_doc) return result def _adjustSectionHeadingLevels(self, section: Dict[str, Any]) -> Dict[str, Any]: """ Adjust heading levels in sections: sections with type heading and level < 2 are changed to level 2. Only chapter headings have level 1. """ adjusted_section = copy.deepcopy(section) # Check if this is a heading section if adjusted_section.get("content_type") == "heading": elements = adjusted_section.get("elements", []) for element in elements: if isinstance(element, dict) and element.get("type") == "heading": content = element.get("content", {}) if isinstance(content, dict): level = content.get("level", 1) # If level < 2, change to level 2 (only chapters have level 1) if level < 2: content["level"] = 2 return adjusted_section def _buildChapterSectionsStructurePrompt( self, chapterId: str, chapterLevel: int, chapterTitle: str, generationHint: str, contentPartIds: List[str], contentPartInstructions: Dict[str, Any], contentParts: List[ContentPart], userPrompt: str, language: str = "en" ) -> str: """Baue Prompt für Chapter-Sections-Struktur-Generierung.""" # Baue ContentParts-Index (nur IDs, keine Previews!) contentPartsIndex = "" for partId in contentPartIds: part = self._findContentPartById(partId, contentParts) if not part: continue contentFormat = part.metadata.get("contentFormat", "unknown") instruction = contentPartInstructions.get(partId, {}).get("instruction", "Use content as needed") contentPartsIndex += f"\n- ContentPart ID: {partId}\n" contentPartsIndex += f" Format: {contentFormat}\n" contentPartsIndex += f" Type: {part.typeGroup}\n" contentPartsIndex += f" Instruction: {instruction}\n" if not contentPartsIndex: contentPartsIndex = "\n(No content parts specified for this chapter)" prompt = f"""TASK: Generate Chapter Sections Structure LANGUAGE: Generate all content in {language.upper()} language. All text, titles, headings, paragraphs, and content must be written in {language.upper()}. CHAPTER: {chapterTitle} (Level {chapterLevel}, ID: {chapterId}) GENERATION HINT: {generationHint} NOTE: Chapter already has a heading section. Do NOT generate a heading for the chapter title. IMPORTANT - SECTION INDEPENDENCE: - Each section is independent and self-contained - One section does NOT have information about another section - Each section must provide its own context and be understandable alone AVAILABLE CONTENT PARTS: {contentPartsIndex} CONTENT TYPES: table, bullet_list, heading, paragraph, code_block, image useAiCall RULES: - useAiCall: true ONLY if ContentPart Format is "extracted" AND transformation needed - useAiCall: false if Format is "object" or "reference" (direct insertion) - useAiCall: false if Format is "extracted" AND simple "include full text" instruction - useAiCall: true if NO ContentPartIds provided (content must be generated from scratch); Sections without ContentParts MUST have a clear, detailed generationHint explaining what content to generate RETURN JSON: {{ "sections": [ {{ "id": "section_1", "content_type": "paragraph", "contentPartIds": ["extracted_part_1"], "generationHint": "Include full text", "useAiCall": false, "elements": [] }} ] }} EXAMPLES (all content types): - paragraph: {{"id": "s1", "content_type": "paragraph", "contentPartIds": ["extracted_1"], "generationHint": "Include full text", "useAiCall": false, "elements": []}} - bullet_list: {{"id": "s2", "content_type": "bullet_list", "contentPartIds": ["extracted_1"], "generationHint": "Create bullet list", "useAiCall": true, "elements": []}} - table: {{"id": "s3", "content_type": "table", "contentPartIds": ["extracted_1", "extracted_2"], "generationHint": "Create table", "useAiCall": true, "elements": []}} - heading: {{"id": "s4", "content_type": "heading", "contentPartIds": ["extracted_1"], "generationHint": "Extract heading", "useAiCall": true, "elements": []}} - code_block: {{"id": "s5", "content_type": "code_block", "contentPartIds": ["extracted_1"], "generationHint": "Format code", "useAiCall": true, "elements": []}} - image: {{"id": "s6", "content_type": "image", "contentPartIds": ["obj_1"], "generationHint": "Display image", "useAiCall": false, "elements": []}} - reference: {{"id": "s7", "content_type": "paragraph", "contentPartIds": ["ref_1"], "generationHint": "Reference", "useAiCall": false, "elements": []}} - NO CONTENT PARTS (generate from scratch): {{"id": "s8", "content_type": "paragraph", "contentPartIds": [], "generationHint": "Write a detailed professional paragraph explaining [specific topic or purpose]. Include [key points to cover]. Address [important aspects]. Conclude with [summary or recommendations].", "useAiCall": true, "elements": []}} CRITICAL: Return ONLY valid JSON. Do not include any explanatory text outside the JSON. """ return prompt def _getContentStructureExample(self, contentType: str) -> str: """Get the JSON structure example for a specific content type.""" structures = { "table": '{{"headers": ["Column1", "Column2"], "rows": [["Value1", "Value2"], ["Value3", "Value4"]]}}', "bullet_list": '{{"items": ["Item 1", "Item 2", "Item 3"]}}', "heading": '{{"text": "Section Title", "level": 2}}', "paragraph": '{{"text": "This is paragraph text."}}', "code_block": '{{"code": "function example() {{ return true; }}", "language": "javascript"}}', "image": '{{"base64Data": "", "altText": "Description", "caption": "Optional caption"}}' } return structures.get(contentType, '{{"text": ""}}') def _buildSectionGenerationPrompt( self, section: Dict[str, Any], contentParts: List[Optional[ContentPart]], userPrompt: str, generationHint: str, allSections: Optional[List[Dict[str, Any]]] = None, sectionIndex: Optional[int] = None, isAggregation: bool = False, language: str = "en" ) -> str: """Baue Prompt für Section-Generierung mit vollständigem Kontext.""" # Filtere None-Werte validParts = [p for p in contentParts if p is not None] # Section-Metadaten sectionId = section.get("id", "unknown") contentType = section.get("content_type", "paragraph") # Baue ContentParts-Beschreibung contentPartsText = "" if isAggregation: # Aggregation: Zeige nur Metadaten, nicht Previews contentPartsText += f"\n## CONTENT PARTS (Aggregation)\n" contentPartsText += f"- Anzahl: {len(validParts)} ContentParts\n" contentPartsText += f"- Alle ContentParts werden als Parameter übergeben (nicht im Prompt!)\n" contentPartsText += f"- Jeder Part kann sehr groß sein → Chunking automatisch\n" contentPartsText += f"- WICHTIG: Aggregiere ALLE Parts zu einem Element (z.B. eine Tabelle)\n\n" contentPartsText += f"ContentPart IDs:\n" for part in validParts: contentFormat = part.metadata.get("contentFormat", "unknown") contentPartsText += f" - {part.id} (Format: {contentFormat}, Type: {part.typeGroup}" if part.metadata.get("originalFileName"): contentPartsText += f", Source: {part.metadata.get('originalFileName')}" contentPartsText += ")\n" else: # Einzelverarbeitung: Zeige Previews for part in validParts: contentFormat = part.metadata.get("contentFormat", "unknown") contentPartsText += f"\n- ContentPart {part.id}:\n" contentPartsText += f" Format: {contentFormat}\n" contentPartsText += f" Type: {part.typeGroup}\n" if part.metadata.get("originalFileName"): contentPartsText += f" Source file: {part.metadata.get('originalFileName')}\n" if contentFormat == "extracted": # Zeige Preview von extrahiertem Text (länger für besseren Kontext) previewLength = 1000 if part.data: preview = part.data[:previewLength] + "..." if len(part.data) > previewLength else part.data contentPartsText += f" Content preview:\n```\n{preview}\n```\n" else: contentPartsText += f" Content: (empty)\n" elif contentFormat == "reference": contentPartsText += f" Reference: {part.metadata.get('documentReference')}\n" if part.metadata.get("usageHint"): contentPartsText += f" Usage hint: {part.metadata.get('usageHint')}\n" elif contentFormat == "object": dataLength = len(part.data) if part.data else 0 contentPartsText += f" Object type: {part.typeGroup}\n" contentPartsText += f" MIME type: {part.mimeType}\n" contentPartsText += f" Data size: {dataLength} chars (base64 encoded)\n" if part.metadata.get("usageHint"): contentPartsText += f" Usage hint: {part.metadata.get('usageHint')}\n" # Baue Section-Kontext (vorherige und nachfolgende Sections) contextText = "" if allSections and sectionIndex is not None: prevSections = [] nextSections = [] if sectionIndex > 0: for i in range(max(0, sectionIndex - 2), sectionIndex): prevSection = allSections[i] prevSections.append({ "id": prevSection.get("id"), "content_type": prevSection.get("content_type"), "generation_hint": prevSection.get("generation_hint", "")[:100] }) if sectionIndex < len(allSections) - 1: for i in range(sectionIndex + 1, min(len(allSections), sectionIndex + 3)): nextSection = allSections[i] nextSections.append({ "id": nextSection.get("id"), "content_type": nextSection.get("content_type"), "generation_hint": nextSection.get("generation_hint", "")[:100] }) if prevSections or nextSections: contextText = "\n## DOCUMENT CONTEXT\n" if prevSections: contextText += "\nPrevious sections:\n" for prev in prevSections: contextText += f"- {prev['id']} ({prev['content_type']}): {prev['generation_hint']}\n" if nextSections: contextText += "\nFollowing sections:\n" for next in nextSections: contextText += f"- {next['id']} ({next['content_type']}): {next['generation_hint']}\n" contentStructureExample = self._getContentStructureExample(contentType) # Special handling for image content type with IMAGE_GENERATE isImageGeneration = contentType == "image" and len(validParts) == 0 if isAggregation: prompt = f"""# TASK: Generate Section Content (Aggregation) LANGUAGE: Generate all content in {language.upper()} language. All text, titles, headings, paragraphs, and content must be written in {language.upper()}. ## SECTION METADATA - Section ID: {sectionId} - Content Type: {contentType} - Generation Hint: {generationHint} ## AVAILABLE CONTENT FOR THIS SECTION {contentPartsText if contentPartsText else "(No content parts specified for this section)"} ## INSTRUCTIONS 1. Generate content for section "{sectionId}" based on the generation hint above 2. **AGGREGATION**: Combine ALL provided ContentParts into ONE element (e.g., one table with all data) 3. For table content_type: Create a single table with headers and rows from all ContentParts 4. For bullet_list content_type: Create a single list with items from all ContentParts 5. Format appropriately based on content_type ({contentType}) 6. Ensure the generated content is self-contained and understandable independently 7. Return ONLY a JSON object with an "elements" array 8. Each element should match the content_type: {contentType} 9. CRITICAL - NO HTML/STYLING: Do NOT include HTML tags, CSS styles, or any formatting markup in text content. Return plain text only. Formatting is handled automatically by the renderer. 10. For paragraphs: Return plain text only, no HTML tags like
, ,

, or style attributes 11. For headings: Return plain text only, no HTML tags or styling 12. For images: Do NOT include base64 data in JSON - images are handled separately ## OUTPUT FORMAT Return a JSON object with this structure: {{ "elements": [ {{ "type": "{contentType}", "content": {contentStructureExample} }} ] }} CRITICAL: - "content" MUST always be an object (never a string) - For text content: Return plain text only, NO HTML tags, NO CSS styles, NO formatting markup - Return ONLY valid JSON. Do not include any explanatory text outside the JSON. ## CONTEXT (for reference only) {contextText if contextText else ""} ``` {userPrompt} ``` """ else: prompt = f"""# TASK: Generate Section Content LANGUAGE: Generate all content in {language.upper()} language. All text, titles, headings, paragraphs, and content must be written in {language.upper()}. ## SECTION METADATA - Section ID: {sectionId} - Content Type: {contentType} - Generation Hint: {generationHint} ## AVAILABLE CONTENT FOR THIS SECTION {contentPartsText if contentPartsText else "(No content parts specified for this section)"} ## INSTRUCTIONS 1. Generate content for section "{sectionId}" based on the generation hint above 2. Use the available content parts to populate this section 3. For extracted text: Format appropriately based on content_type ({contentType}) 4. Ensure the generated content is self-contained and understandable independently 5. Return ONLY a JSON object with an "elements" array 6. Each element should match the content_type: {contentType} 7. CRITICAL - NO HTML/STYLING: Do NOT include HTML tags, CSS styles, or any formatting markup in text content. Return plain text only. Formatting is handled automatically by the renderer. 8. For paragraphs: Return plain text only, no HTML tags like

, ,

, or style attributes 9. For headings: Return plain text only, no HTML tags or styling 10. For images: If you need to reference an image, describe it in altText. Do NOT include base64 data - images are handled separately ## OUTPUT FORMAT Return a JSON object with this structure: {{ "elements": [ {{ "type": "{contentType}", "content": {contentStructureExample} }} ] }} CRITICAL: - "content" MUST always be an object (never a string) - For text content: Return plain text only, NO HTML tags, NO CSS styles, NO formatting markup - Return ONLY valid JSON. Do not include any explanatory text outside the JSON ## CONTEXT (for reference only) {contextText if contextText else ""} ``` {userPrompt} ``` """ return prompt def _findContentPartById(self, partId: str, contentParts: List[ContentPart]) -> Optional[ContentPart]: """Finde ContentPart nach ID.""" for part in contentParts: if part.id == partId: return part return None def _needsAggregation( self, contentType: str, contentPartCount: int ) -> bool: """ Bestimmt ob mehrere ContentParts aggregiert werden müssen. Aggregation nötig wenn: - content_type erfordert Aggregation (table, bullet_list) - UND mehrere ContentParts vorhanden sind (> 1) Args: contentType: Section content_type contentPartCount: Anzahl der ContentParts in dieser Section Returns: True wenn Aggregation nötig, False sonst """ aggregationTypes = ["table", "bullet_list"] if contentType in aggregationTypes and contentPartCount > 1: return True # Optional: Auch für paragraph wenn mehrere Parts vorhanden # (z.B. Vergleich mehrerer Dokumente) # Standard: Keine Aggregation für paragraph return False