[feature]for each document, generate some propertites

omnibox_wizard/resources/prompts/property.md

@@ -0,0 +1,290 @@
+# Task Description
+
+Extract structured properties from the given text content, identifying key entities and concepts that appear in the document.
+
+# Guidelines
+
+- Extract specific entities mentioned in the text, not general categories
+- For each property type, only include items that are explicitly mentioned or clearly referenced
+- If no relevant entities are found for a property type, return an empty array
+- Your response must be in JSON format with the specified property keys
+- Extract entities in user's preference language when possible
+- Prioritize accuracy and specificity
+
+# Universal Properties (Always Extract)
+
+- **main_topic**: The primary subject or theme of the document (1-2 items max)
+- **key_concepts**: Important concepts, ideas, or themes discussed (3-5 items max)
+- **persons**: Specific people mentioned by name
+- **organizations**: Companies, government departments, schools, institutions mentioned
+- **locations**: Countries, cities, landmarks, geographic places mentioned
+- **products**: Specific product names or services mentioned (e.g. iPhone 16, ChatGPT-5)
+- **technologies**: Professional technical terms mentioned (e.g. Large Language Models, Quantum Computing, Blockchain)
+- **events**: Specific events mentioned (e.g. 2025 World AI Conference)
+
+# Document-Specific Properties
+
+## Meeting Notes (meeting_notes)
+- **meeting_title**: Meeting subject/title
+- **meeting_date**: Meeting date (ISO format)
+- **attendees**: List of attendees
+- **agenda_items**: Meeting agenda items
+- **key_decisions**: Important decisions made
+- **action_items**: Tasks with owner and due date [{"task": "...", "owner": "...", "due_date": "..."}]
+- **unresolved_issues**: Issues that need follow-up
+- **sentiment**: Meeting atmosphere ("Positive", "Neutral", "Controversial")
+
+## Technical Design (technical_design)
+- **project_name**: Project or system name
+- **tech_stack**: Key technologies and frameworks used
+- **architecture_type**: System architecture style
+- **api_endpoints**: API definitions [{"method": "...", "path": "...", "description": "..."}]
+- **database_schema**: Database tables [{"table": "...", "fields": [...]}]
+- **dependencies**: External services/components
+- **performance_targets**: Performance requirements
+- **risks_and_mitigations**: Risks and solutions [{"risk": "...", "mitigation": "..."}]
+
+## Market Analysis (market_analysis)
+- **report_type**: Specific report type
+- **target_market**: Target market or user group
+- **analyzed_companies**: Companies analyzed
+- **key_findings**: Core findings and conclusions
+- **data_sources**: Referenced data sources
+- **swot_analysis**: SWOT analysis {"strengths": [...], "weaknesses": [...], "opportunities": [...], "threats": [...]}
+- **market_size**: Market size and growth data
+- **future_trends**: Predicted market trends
+
+## Legal Contract (legal_contract)
+- **contract_type**: Type of contract
+- **party_a**: First party full name
+- **party_b**: Second party full name
+- **effective_date**: Contract effective date
+- **term**: Contract duration/term
+- **governing_law**: Applicable law and jurisdiction
+- **payment_terms**: Payment terms and amounts
+- **liability_clause**: Liability limitation summary
+- **confidentiality_period**: Confidentiality period
+
+## Resume (resume)
+- **candidate_name**: Candidate's name
+- **contact_info**: Contact information {"email": "...", "phone": "..."}
+- **education_history**: Education background [{"school": "...", "major": "...", "degree": "...", "period": "..."}]
+- **work_experience**: Work history [{"company": "...", "position": "...", "period": "...", "summary": "..."}]
+- **skills_set**: Skills list
+- **years_of_experience**: Total work experience in years
+- **current_location**: Current location
+- **certifications**: Professional certifications
+
+## User Interview (user_interview)
+- **interviewee_profile**: User profile {"user_id": "...", "role": "...", "industry": "..."}
+- **interview_date**: Interview date
+- **user_goals**: User goals and objectives
+- **pain_points**: User pain points and frustrations
+- **key_quotes**: Important user quotes
+- **feature_requests**: Requested features
+- **overall_sentiment**: User's overall sentiment
+- **researcher**: Interview researcher/conductor
+
+## Project Proposal (project_proposal)
+- **project_title**: Official project name
+- **project_owner**: Project owner/sponsor
+- **objectives**: Main project objectives
+- **stakeholders**: Key stakeholders
+- **timeline_milestones**: Key milestones [{"milestone": "...", "date": "..."}]
+- **budget_estimate**: Estimated budget
+- **deliverables**: Project deliverables
+- **success_metrics**: Success metrics/KPIs
+
+## Press Release (press_release)
+- **announcement_subject**: Core announcement
+- **company_involved**: Companies involved
+- **key_persons**: Key people mentioned with titles
+- **release_date**: Release date
+- **financial_figures**: Key financial numbers
+- **purpose_of_funds**: (For funding news) Fund usage
+- **forward_looking_statements**: Future plans summary
+- **media_contact**: Media contact info {"name": "...", "email": "..."}
+
+## How-To Guide / SOP (how_to_guide)
+- **procedure_title**: Guide or process title
+- **target_audience**: Intended audience/readers
+- **prerequisites**: Required conditions or preparations
+- **steps**: Structured operation steps [{"step": 1, "action": "..."}]
+- **expected_outcome**: Expected result after completion
+- **troubleshooting**: Common issues and solutions [{"problem": "...", "solution": "..."}]
+- **tools_required**: Required software/hardware tools
+- **last_verified_date**: Last verified date (ISO format)
+
+## Research Paper / Academic Article (research_paper)
+- **paper_title**: Paper title
+- **authors**: Author list with affiliations
+- **publication_info**: Publication details {"journal": "...", "year": ...}
+- **problem_statement**: Research problem being addressed
+- **methodology**: Research method used
+- **key_findings**: Most important results/conclusions
+- **datasets_used**: Datasets used in experiments
+- **limitations**: Acknowledged limitations or future improvements
+
+## Financial Report (financial_report)
+- **report_period**: Reporting period (e.g., "Q3 2024", "FY 2024")
+- **company_name**: Reporting company
+- **revenue_figures**: Key revenue numbers
+- **profit_loss**: Profit/loss statements
+- **key_metrics**: Important financial KPIs
+- **segment_performance**: Performance by business segment
+- **risk_factors**: Identified financial risks
+- **outlook**: Future financial guidance
+- **auditor**: External auditor information
+
+## Marketing Campaign (marketing_campaign)
+- **campaign_name**: Campaign title/name
+- **campaign_period**: Campaign duration
+- **target_demographics**: Target audience segments
+- **marketing_channels**: Channels used (social media, email, etc.)
+- **budget_allocation**: Budget breakdown by channel
+- **key_messages**: Core marketing messages
+- **success_metrics**: Campaign KPIs and goals
+- **creative_assets**: Mentioned creative materials
+- **campaign_manager**: Person responsible for campaign
+
+## Product Specification (product_spec)
+- **product_name**: Product name/title
+- **version**: Product version number
+- **feature_list**: Key features and capabilities
+- **technical_requirements**: System/hardware requirements
+- **user_stories**: User scenarios or use cases
+- **acceptance_criteria**: Definition of done criteria
+- **dependencies**: Required integrations or components
+- **constraints**: Technical or business limitations
+- **priority_level**: Development priority
+
+## Customer Feedback (customer_feedback)
+- **feedback_type**: Type of feedback (Complaint, Suggestion, Praise, etc.)
+- **customer_info**: Customer details {"id": "...", "segment": "..."}
+- **product_service**: Product/service being discussed
+- **satisfaction_score**: Numerical ratings if mentioned
+- **specific_issues**: Detailed problems or concerns
+- **improvement_suggestions**: Customer suggestions
+- **follow_up_required**: Whether follow-up is needed
+- **feedback_channel**: How feedback was received
+
+## Training Material (training_material)
+- **course_title**: Training course or module name
+- **target_learners**: Intended audience/learners
+- **learning_objectives**: Educational goals
+- **course_duration**: Expected time to complete
+- **prerequisite_knowledge**: Required background knowledge
+- **assessment_methods**: How learning is evaluated
+- **certification_offered**: Available certifications
+- **instructional_design**: Teaching methodology used
+- **course_materials**: Required textbooks, software, etc.
+
+## Incident Report (incident_report)
+- **incident_title**: Incident summary
+- **incident_date**: When incident occurred
+- **severity_impact**: Impact level and affected users
+- **root_cause**: Underlying cause of incident
+- **timeline**: Key events timeline [{"time": "...", "event": "..."}]
+- **resolution_steps**: Actions taken to resolve
+- **lessons_learned**: Key takeaways
+- **preventive_measures**: Steps to prevent recurrence
+- **incident_commander**: Person who led response
+
+## Survey Results (survey_results)
+- **survey_title**: Survey name/topic
+- **survey_period**: Data collection timeframe
+- **sample_size**: Number of respondents
+- **response_rate**: Participation rate percentage
+- **key_demographics**: Respondent characteristics
+- **major_findings**: Most important results
+- **statistical_significance**: Confidence levels mentioned
+- **methodology**: Data collection approach
+- **recommendations**: Suggested actions based on results
+
+## Policy Document (policy_document)
+- **policy_title**: Official policy name
+- **effective_date**: When policy takes effect
+- **policy_owner**: Department/person responsible
+- **scope_applicability**: Who/what policy applies to
+- **key_requirements**: Main policy requirements
+- **compliance_procedures**: How compliance is ensured
+- **violations_consequences**: Penalties for non-compliance
+- **review_schedule**: Policy review frequency
+- **related_policies**: Connected or referenced policies
+
+## Email Communication (email_communication)
+- **email_subject**: Email subject line
+- **sender**: Email sender information
+- **recipients**: Primary recipients list
+- **cc_bcc**: CC/BCC recipients if mentioned
+- **email_type**: Type (announcement, request, update, etc.)
+- **key_requests**: Specific asks or requests made
+- **deadlines**: Mentioned deadlines or dates
+- **attachments**: Referenced file attachments
+- **follow_up_needed**: Whether response/action is required
+
+## Job Description (job_description)
+- **job_title**: Position title
+- **department**: Department or team
+- **employment_type**: Full-time, part-time, contract, etc.
+- **location_type**: Remote, on-site, hybrid
+- **experience_required**: Years of experience needed
+- **key_responsibilities**: Main job duties
+- **required_qualifications**: Must-have skills/education
+- **preferred_qualifications**: Nice-to-have skills
+- **compensation_range**: Salary range if mentioned
+- **benefits_package**: Benefits offered
+
+## Event Planning (event_planning)
+- **event_name**: Event title
+- **event_date**: Scheduled date and time
+- **event_location**: Venue information
+- **event_type**: Conference, workshop, social, etc.
+- **expected_attendance**: Number of attendees
+- **agenda_schedule**: Event schedule/program
+- **speakers_presenters**: Featured speakers
+- **budget_costs**: Event budget breakdown
+- **logistics_requirements**: Equipment, catering, etc.
+- **registration_process**: How people sign up
+
+# Output Format
+
+## Base Structure (Always Include)
+```json
+{
+  "main_topic": ["topic1", "topic2"],
+  "key_concepts": ["concept1", "concept2", "concept3"],
+  "persons": ["person1", "person2"],
+  "organizations": ["org1", "org2"],
+  "locations": ["location1", "location2"],
+  "products": ["product1", "product2"],
+  "technologies": ["tech1", "tech2"],
+  "events": ["event1", "event2"]
+}
+```
+
+## Complete Output (Include document-specific properties when applicable)
+```json
+{
+  // Universal properties (always include)
+  "main_topic": ["topic1"],
+  "key_concepts": ["concept1", "concept2"],
+  "persons": ["person1"],
+  "organizations": ["org1"],
+  "locations": ["location1"],
+  "products": ["product1"],
+  "technologies": ["tech1"],
+  "events": ["event1"],
+
+  // Document-specific properties (include based on document type)
+  "meeting_title": "...",        // for meeting_notes
+  "project_name": "...",         // for technical_design
+  "candidate_name": "...",       // for resume
+  "contract_type": "...",        // for legal_contract
+  // ... other document-specific properties as needed
+}
+
+# Meta info
+
+- Current time: {{ now }}
+- User's preference language: {{ lang }}

omnibox_wizard/wizard/grimoire/common_ai.py

@@ -21,6 +21,8 @@ def __init__(self, config: GrimoireOpenAIConfig):
             self.title_system_prompt_template: str = f.read()
         with project_root.open("omnibox_wizard/resources/prompts/tag.md") as f:
             self.tag_system_prompt_template: str = f.read()
+        with project_root.open("omnibox_wizard/resources/prompts/property.md") as f:
+            self.property_system_prompt_template: str = f.read()
 
     @tracer.start_as_current_span("CommonAI._invoke")
     async def _invoke(
@@ -70,3 +72,9 @@ async def tags(self, text: str, *, lang: str | None = None, trace_info: TraceInf
         Create tags according to the given text
         """
         return (await self._invoke(text, self.tag_system_prompt_template, "mini", lang, trace_info))["tags"]
+
+    async def properties(self, text: str, *, lang: str | None = None, trace_info: TraceInfo | None = None) -> dict:
+        """
+        Extract properties according to the given text
+        """
+        return await self._invoke(text, self.property_system_prompt_template, "mini", lang, trace_info)

omnibox_wizard/worker/functions/property_generator.py

@@ -0,0 +1,33 @@
+from omnibox_wizard.common.trace_info import TraceInfo
+from omnibox_wizard.wizard.grimoire.common_ai import CommonAI
+from omnibox_wizard.worker.config import WorkerConfig
+from omnibox_wizard.worker.entity import Task
+from omnibox_wizard.worker.functions.base_function import BaseFunction
+
+
+class PropertyGenerator(BaseFunction):
+    def __init__(self, config: WorkerConfig):
+        self.config = config
+        self.common_ai = CommonAI(config.grimoire.openai)
+
+    async def run(self, task: Task, trace_info: TraceInfo) -> dict:
+        input_dict = task.input
+        text = input_dict.get("text", "")
+        lang = input_dict.get("lang", None)
+
+        if not text:
+            raise ValueError("Text input is required for property generation")
+
+        trace_info = trace_info.bind(text_length=len(text))
+        trace_info.info({"message": "Starting property generation"})
+
+        try:
+            properties = await self.common_ai.properties(text, trace_info=trace_info, lang=lang)
+
+            result_dict = {"properties": properties}
+            trace_info.info({"extracted_properties": properties, "message": "Property generation completed"})
+            return result_dict
+
+        except Exception as e:
+            trace_info.error({"error": str(e), "message": "Failed to generate properties"})
+            raise

omnibox_wizard/worker/worker.py

@@ -20,6 +20,7 @@
 from omnibox_wizard.worker.functions.tag_extractor import TagExtractor
 from omnibox_wizard.worker.functions.title_generator import TitleGenerator
 from omnibox_wizard.worker.functions.video_note_generator import VideoNoteGenerator
+from omnibox_wizard.worker.functions.property_generator import PropertyGenerator
 from omnibox_wizard.worker.health_tracker import HealthTracker
 from omnibox_wizard.worker.task_manager import TaskManager
 
@@ -44,6 +45,7 @@ def __init__(self, config: WorkerConfig, worker_id: int, health_tracker: HealthT
             "extract_tags": TagExtractor(config),
             "generate_title": TitleGenerator(config),
             "generate_video_note": VideoNoteGenerator(config),
+            "generate_properties": PropertyGenerator(config),
         }
 
         functions_enabled = set(self.worker_dict.keys())