SQLAgent realizes the Natural Language Querying capability converting questions into SQL queries. RAG Agent realizes the Natural Language Querying capability by semantic search and answer generation over documents.

The Platform Backend implements a natural language query endpoint to generate and execute SQL queries from user questions.

Provides users a way to ask questions in natural language and receive structured answers. Processes queries to generate SQL or semantic search, returning results with confidence scores and source attribution, aiming to improve data system accessibility and interaction.

SQLCoder-7B deployed on elin GPU, running at 2-3 seconds inference speed, enabling 3-5x faster DataLens queries than previous models. It generates valid DuckDB SQL and is integrated with the backend for improved query performance.

NDJSON response streaming endpoint implemented for real-time query processing, helping to mitigate timeout issues during long-running DataLens analyses.

Infrastructure includes Network as a monitored specification. Network between theo and elin relies on SSH tunnel managed by theo stakeholder. Network between theo and elin relies on SSH tunnel managed by the elin server. The Docker deployment includes network configurations for service communication. The Docker deployment configures networking for container communication Network connection stability, such as SSH tunnel between theo and elin and OpenClaw node status, is part of resource management.

Planned UI component displaying table relationships, join confidence, and user-customizable consolidation steps.

new project page is implemented as a SvelteKit page. projects > new page is built using the SvelteKit framework. projects > new page is part of the DataLens system. The projects > new page is part of the DataLens project.

Post-deployment activities include creating a user, uploading data, verifying system functionality, and initial testing to ensure successful system operation and readiness.

The Frontend is implemented using Node.js runtime environment.

Embedding service uses the nomic-embed-text 768-dimensional model via Ollama for GPU accelerated vector embedding. The nomic-embed-text component is part of Ollama embeddings used for GPU batch embedding processing of document chunks. The Nomic-embed-text embedding model is used by the async embedding queue for batch GPU embedding processing. Semantic Search capability uses nomic-embed-text component for creating table embeddings to improve table ranking in schema selection. TableEmbeddingIndex uses nomic-embed-text for embedding table names and descriptions. RAG Agent employs the nomic-embed-text embedding model for creating document embeddings.

TableEmbeddingIndex uses the nomic-embed-text model to compute semantic embeddings for table names and schemas. QdrantService uses the nomic-embed-text model via Ollama API to embed texts into 768D vectors.

DataLens uses notebooks as part of its data analysis work. DataLens agent uses notebooks as part of its workflow for data analysis.

Data from NOVANA files integrated into the SVGV project; on-demand extraction and schema standardization implemented.

DataLens uses npm to run local tests as part of its deployment workflow. DataLens deployment workflow uses npm commands for testing frontend before pushing code.

Script for running and managing E2E discovery tests with real SVGV data, covering core functionality, UX, and performance validations.

NSQL-6B identified as a potential alternative for SQL generation to improve inference speed and SQL quality in future upgrades, complementing current SQLCoder-7B deployment.

DataLens checks GPU usage via nvidia-smi before any GPU work.

DataLens uses Ollama (ollama/qwen3-coder-next) on elin for local LLM inference, including embeddings and query synthesis. Integrated with TextToSQLService and vector search. Requires Ollama to run on elin (176.9.90.154) with port 11434 accessible. Deployed on GPU for high-performance AI tasks. Critical for Text-to-SQL and document embedding pipelines, and part of the Arctic-Text2SQL setup. IronClaw Gateway uses Ollama with the Qwen3-coder-next model as fallback for local inference when Anthropic Claude not available.

Runs local LLM models like qwen3-coder-next on elin's GPU for inference, integral for autonomous cataloging and text-to-SQL. Ollama and Arctic LLM inference engines are used by the Data Discovery system for intelligent table discovery and query processing. Qdrant Service uses Ollama Embedding Service to create vector embeddings.

The ollama user operates the Ollama inference service hosting the Qwen3-coder-next model used as a fallback LLM for query generation.

DataLens integrates with the Ollama Arctic Model for Text-to-SQL queries in production.

The Qdrant vector search service uses Ollama embeddings for generating vector representations of data. RAG Agent uses Ollama embeddings via nomic-embed-text for document retrieval The Docling extraction system utilizes Ollama embeddings (nomic-embed-text) to generate vector embeddings for semantic search and reasoning. The nomic-embed-text component is part of Ollama embeddings used for GPU batch embedding processing of document chunks. Qdrant vectors store the vector embeddings generated by Ollama embeddings from Docling extracted chunks for semantic search. Ollama embeddings run on the RTX 4000 SFF Ada 20GB GPU for batch processing of text chunks.

The Ollama GPU qwen3-coder-next 80B model is used by the async embedding queue to generate GPU embeddings for text chunks.

The ollama user operates the Ollama inference service hosting the Qwen3-coder-next model used as a fallback LLM for query generation.