The Multilingual Support (Danish) epic involves modifying Backend files to implement language preference and LLM prompt injection.

Backend system supports autonomous data extraction, natural language querying, document RAG, and findings visualization. Recent updates include integrating findings generation into API and agent workflows, with components ready for deployment and testing. The consolidation entity is to be connected and used within the analysis pipeline. The analysis pipeline is planned to use a consolidated view entity when conduction analyses. The analysis pipeline uses Arctic to generate SQL queries over the unified schema.

The backend service depends on the PostgreSQL 16 service as its metadata storage with connection configured via environment variable. The backend service depends on the Redis 7 service for job queueing with connection configured via environment variable. The backend service uses DS-STAR agents and their Python virtual environment from the mounted /home/ops/datalens directory for execution. The Backend implements the DS-STAR intelligence capability. The Backend implements the Text-to-SQL capability. The Backend implements the Document RAG capability. The DataLens Platform contains the Backend component. Coolify Application depends on the Backend for API functionality. Backend integrates with PostgreSQL for authentication and project metadata. Backend includes a production-ready Dockerfile for deployment. The Data Discovery feature includes the Discovery backend service. The Discovery backend service uses defined API endpoints for its operation. The Backend Service uses the defined API endpoints including the discovery endpoints. The Backend API integrates with the Discovery backend service via the defined routes. The Data Discovery feature includes the Discovery backend service. The Discovery backend service uses defined API endpoints for its operation. The Backend Service uses the defined API endpoints including the discovery endpoints. The Backend API integrates with the Discovery backend service via the defined routes. Integration requires registering and using the Backend service's API endpoints for discovery.

The Backend Service question_router.py uses the Backend Service consolidation.py to manage transient consolidated views as part of query execution.

The Backend Service question_router.py depends on the Backend Service query_enhancer.py to extract entities and enhance queries for consolidated SQL generation. The Backend Service query_enhancer.py uses Danish language keywords and entity recognition to extract columns and values from Danish questions.

The Arctic-Text2SQL-R1-7B model is used via the Backend Service question_router.py to generate SQL. The Qwen3 model is used within the Backend Service question_router.py for schema selection and question ranking tasks. The Backend Service question_router.py depends on the Backend Service schema_graph.py for schema graph construction and join discovery. The Backend Service question_router.py depends on the Backend Service query_enhancer.py to extract entities and enhance queries for consolidated SQL generation. The Backend Service question_router.py uses the Backend Service consolidation.py to manage transient consolidated views as part of query execution.

The Backend Service question_router.py depends on the Backend Service schema_graph.py for schema graph construction and join discovery. The Backend Service schema_graph.py uses the Danish Budget Tables to perform join key analysis and table clustering for consolidation.

Existing service for establishing table indexes; no detailed updates provided.

The Arctic-Text2SQL-R1-7B model is utilized by the Backend Service text_to_sql.py for SQL query generation. The Qwen3 model is used in the Backend Service text_to_sql.py for answer synthesis after SQL execution.

Backend tests validate the Discovery feature's API endpoints and server-side logic.

Docling is installed and running on the backend-qg4kk8ccggw844wscsossogs server container to support batch processing.

The backend API agent code is part of the IronClaw-powered Agent Mode implementation. The Full Findings Visualization Layer depends on integration with Agent.py to return findings on each query. agent.py logging is a specific logging to be added when creating AgentFinding records to confirm saving of findings. agent.py logging is to be implemented in backend/app/api/agent.py at the creation of AgentFinding records. agent.py logging is to be implemented in backend/app/api/agent.py at the creation of AgentFinding records. agent.py logging is a specific logging to be added when creating AgentFinding records to confirm saving of findings. agent.py logging is to be implemented in backend/app/api/agent.py at the creation of AgentFinding records. agent.py logging is to be implemented in backend/app/api/agent.py at the creation of AgentFinding records. The backend/app/api/agent.py file is part of the IronClaw-powered Agent Mode implementation. The Full Findings Visualization Layer is planned to be integrated with agent.py to return findings on each query. Logging will be added in agent.py when AgentFinding records are created to verify findings saving to the database. Logging will be added in agent.py when AgentFinding records are created to verify findings saving to the database.

The file backend/app/api/analysis.py is part of the API LAYER. The QuestionRouter class is used by the API LAYER including the backend/app/api/analysis.py endpoint for processing queries. The API Layer contains the backend/app/api/analysis.py file. The file backend/app/api/analysis.py contains the /analysis/ask endpoint. The file backend/app/api/analysis.py contains the /analysis/query endpoint. The file backend/app/api/analysis.py contains the /analysis/history endpoint. The analysis.py API layer uses QuestionRouter to route questions to appropriate services.

The backend app api discovery router is part of the Backend discovery service to provide API integration.

The generate-goal API endpoint is implemented in backend/app/api/projects.py.

Uses Docling GPU on elin for DOCX extraction with semantic chunking, heading hierarchy, and table embedding as JSON. Falls back to python-docx for simple files; refactored for Phase 2 GPU-first processing. The DocxExtractor is part of the Docling extraction system for DOCX documents using GPU extraction. The docx_extractor.py extractor interfaces with Docling on elin GPU via SSH for DOCX extraction with semantic chunking and rich metadata. GPU-first document extraction includes extracting DOCX files using backend/app/extractors/docx_extractor.py that calls Docling on elin GPU. The DOCX extractor uses Docling for extraction on the elin GPU as a mandatory tool to perform semantic chunking with embedded JSON tables and rich metadata. The RQ worker calls the DOCX extractor for extraction using Docling and fails hard if extraction fails, enforcing the no fallback policy. The DOCX extractor produces semantic chunks with rich metadata including hierarchy and provenance to support DS-STAR queries for document reasoning.

The PPTX extractor uses Docling on elin GPU via SSH to extract PPTX files preserving slide semantics and metadata, without fallback. It is based on python-pptx for slide and text extraction, with enhancements for semantic chunking and slide layout detection (title, bullets, blank). It extracts image metadata such as counts and types for added context, integrated into GPU-first extraction workflow.

The backend main application includes the IronClaw-powered Agent Mode feature. Discovery API router was registered in the backend app main.py to expose discovery endpoints. IronClaw agent back-end logic is integrated with the main app via API router registration. Discovery API router was registered in the backend app main.py to expose discovery endpoints. IronClaw agent back-end logic is integrated with the main app via API router registration.

Backend models initialized, no additional details provided.

The agent models are part of the IronClaw-powered Agent Mode subsystem.

The models directory contains backend/app/models/models.py and backend/app/models/database.py files for ORM and database connections.

The agent skills service contributes to the IronClaw-powered Agent Mode functionality. The backend/app/services/agent_skills.py module is part of the IronClaw-powered Agent Mode feature. The code in backend/app/services/agent_skills.py depends on backend/app/services/text_to_sql.py for SQL extraction functionality within agent workflows. backend/app/services/agent_skills.py depends on backend/app/services/findings_generator.py to generate findings after query execution in the agent workflow.

Agent warming service is included in the IronClaw-powered Agent Mode implementation. AgentWarmingService uses ExploreSchemaSkill to assemble warm context by computing schema profiles.

Contains the consolidation logic, implemented for intelligent table merging and creation of transient views. Integration uses backend/app/services/consolidation.py for handling consolidated views in the analysis pipeline.