Advanced worked examples

These guides assume basic Python, familiarity with database queries, and ordinary terminal use. Each starts with a small supported example and explains the GraphForge concepts it needs. They are optional additions to basic graph use, not a required capability checklist.

For agent-assisted research without writing code, start with Your first research project. That Basic path connects observations, passages, interpretation, and a bounded conclusion with guided or agent-assisted execution.

Advanced guide	What you build
Knowledge graph construction	Source-linked mentions, explicit identity rules, candidate lookup
Network analysis	Counts, paths, and native graph algorithms with scoped interpretation
LLM-powered workflows	Proposed extraction, source inspection, and separate human review
Decision workflows	Validate external answers and apply an explicit caller policy
Agent grounding	A tool/capability graph for an application
Agent tool recall	Candidate tools and declared input requirements

Use the matching installation for these v0.6.0 docs. Examples state whether they are memory-only; follow save and reopen before relying on persistence and portable projects for transfer.

Research notes discuss interpretation and evaluation. Architecture and engineering contracts are separate expert references, not prerequisites for these examples.