skills.rest
AarushiShah
AarushiShahCOMMUNITY
Explore all skills by this author →
GitHub

databricks-unstructured-pdf-generation

Generate synthetic PDF documents with evaluation metadata for RAG testing on Databricks.

Updated Mar 24, 2026
Search Tags:#rag#pdf generation#synthetic data#unstructured data#databricks#unity catalog#evaluation dataset
One-click install
npx skills add https://github.com/AarushiShah/coding-agents-databricks-apps --skill databricks-unstructured-pdf-generation-aarushishah
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: databricks-unstructured-pdf-generation
Source: https://github.com/AarushiShah/coding-agents-databricks-apps/tree/main/.claude/skills/databricks-unstructured-pdf-generation
Command: npx skills add https://github.com/AarushiShah/coding-agents-databricks-apps --skill databricks-unstructured-pdf-generation-aarushishah

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Building and evaluating RAG systems requires realistic document corpora, but sourcing domain-specific PDFs with ground-truth questions is slow and often blocked by data privacy constraints. This Skill generates synthetic PDF documents with accompanying question-and-guideline JSON files, uploaded directly to Unity Catalog Volumes. ## Core Features & Use Cases - LLM-Based PDF Generation: Create batches of realistic documents (HR policies, technical docs, financial reports, training materials) from a natural-language description via the generate_pdf_documents MCP tool. - RAG Evaluation Metadata: Each PDF ships with a JSON file containing a title, category, test question, and evaluation guideline for measuring retrieval quality. - Unity Catalog Integration: Output lands in /Volumes/<catalog>/<schema>/<volume>/<folder>/ with configurable volume, folder, document size, and overwrite behavior. - Use Case: Generate 20 technical documentation PDFs for a SaaS analytics platform, ingest them into a vector search index, then use the bundled questions to evaluate your RAG pipeline's answer accuracy. ## Quick Start Ask the agent to generate 10 medium-sized HR policy PDFs into the ai_dev_kit catalog and hr_demo schema using the generate_pdf_documents tool.

Frequently Asked Questions about databricks-unstructured-pdf-generation

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate synthetic PDF documents for RAG testing?▼

Call the generate_pdf_documents MCP tool with a catalog, schema, detailed content description, and document count. The tool creates PDFs plus JSON files containing test questions and evaluation guidelines, saved to a Unity Catalog Volume.

How to create evaluation datasets for retrieval-augmented generation?▼

Each generated PDF includes a JSON file with a question field and a guideline field describing the expected answer. Query your RAG system with the questions and score responses against the guidelines to measure retrieval quality.

Can I use Azure OpenAI instead of Databricks models for PDF generation?▼

Yes, set LLM_PROVIDER=AZURE with AZURE_OPENAI_ENDPOINT, AZURE_OPENAI_API_KEY, and AZURE_OPENAI_DEPLOYMENT environment variables. The default provider is Databricks Foundation Models configured via DATABRICKS_MODEL.

Why does PDF generation fail with no LLM endpoint configured?▼

The tool requires an LLM to write document content, so it fails when neither DATABRICKS_MODEL nor AZURE_OPENAI_DEPLOYMENT is set. Configure one of these environment variables before running generation.

What document sizes are supported for generated PDFs?▼

Three sizes are available via the doc_size parameter: SMALL at roughly one page, MEDIUM at four to six pages, and LARGE at ten or more pages. MEDIUM is the default and suits most RAG testing scenarios.

Related Skills

Explore More Data & Analytics Skills→
FMurrayFMurray

unstructured-pdf-generation

Generates synthetic PDF documents with RAG evaluation JSON files and uploads them to Unity Catalog Volumes.

Community
Intermediate
HermeticOrmusHermeticOrmus

RAG Patterns

Provide expert patterns for building RAG systems with pgvector, BM25, and RAGAS evaluation.

Community
Advanced
JingyiWngJingyiWng

databricks-unstructured-pdf-generation

Generate synthetic PDFs with question-answer JSON metadata into Unity Catalog volumes.

Community
Intermediate
math-incmath-inc
1.2k

segment-anything-model

Generate object masks from prompts or automatically across images.

Official
Advanced