Gregius Data – WPCLI

Overview

Gregius Data ships a full WP-CLI command surface under wp gg-data — eight command families covering sync, vectors, RAG answers, benchmarks, evaluation, listing, and logs. Every subcommand accepts a --format flag (table, json, or csv), producing deterministic output for scripts and CI/CD pipelines.

Automation and scripting are a site operator’s most reliable tools. Without a command-line interface, routine tasks — syncing content after a migration, rebuilding vectors after a model change, auditing logs — require dashboard clicks or REST calls that compound into brittle workflows. The wp gg-data namespace turns these into repeatable, versionable one-liners.

By the end of this page, you will know every command family, its subcommands, flags, defaults, and output formats — enough to script your entire Gregius Data operational surface.

Prerequisites

• WP-CLI installed and configured for the target WordPress site
• Gregius Data plugin installed and activated
• At least one data connection configured and active (for sync, vectors, answer, and listing commands)
• AI models registered (for vector generation and RAG answer commands)

---

How to: Get Started

Every subcommand responds to --help with its full flag contract. Start here before running anything:

wp gg-data --help
wp gg-data sync --help
wp gg-data vectors --help
wp gg-data answer --help
wp gg-data benchmark --help
wp gg-data evaluation --help
wp gg-data list-connections --help
wp gg-data list-models --help
wp gg-data logs --help

All subcommands produce output in one of three formats controlled by --format:

Format	Use case
table	Interactive terminal use (default)
json	Script consumption, API bridges, jq pipelines
csv	Spreadsheet import, reporting pipelines

---

Sync Commands

Sync commands move WordPress content into PostgreSQL. Three subcommands cover posts, terms, and a combined full sync.

Subcommands:

Subcommand	Scope
wp gg-data sync posts	Sync posts by post type
wp gg-data sync terms	Sync all taxonomies, term taxonomies, and term relationships
wp gg-data sync all	Run posts + terms in sequence

Flags:

Flag	Default	Description
--connection	gregius-data	Target data connection name
--post-type	all	Specific post type (posts only)
--batch-size	100	Posts per batch (max 1000)
--dry-run	off	Report what would sync without writing
--format	table	table, json, or csv

Examples:

# Full sync with custom batch size
wp gg-data sync all --connection=gregius-data --batch-size=200

# Sync only pages, dry-run first
wp gg-data sync posts --post-type=page --dry-run

# Sync all posts as CSV for reporting
wp gg-data sync posts --format=csv

---

Vectors Commands

Vector commands generate and rebuild semantic embeddings from your synced content. Two subcommands handle initial generation and full rebuilds.

Subcommands:

Subcommand	Scope
wp gg-data vectors generate	Generate vectors for content that needs them
wp gg-data vectors rebuild	Rebuild all vectors for the connection

Flags:

Flag	Default	Description
--connection	gregius-data	Target data connection name
--embedding-model	tfidf-300	Embedding model ID to use
--post-type	all	Specific post type to process
--batch-size	50	Items per batch (max 200)
--force	off	Force regeneration even when vectors exist
--format	table	table, json, or csv

Examples:

# Generate vectors with a specific model
wp gg-data vectors generate --embedding-model=hashingtf-murmur3-1024 --batch-size=100

# Full rebuild
wp gg-data vectors rebuild --connection=gregius-data

---

Answer Command

Ask a question through the RAG pipeline directly from the command line. Useful for testing retrieval quality and debugging prompt behavior.

Command: wp gg-data answer <query>

Flags:

Flag	Default	Description
--connection	gregius-data	Target data connection
--embedding-model	tfidf-300	Embedding model for search
--agentic-model	(none)	Model for agentic routing
--rerank-model	(none)	Model for result reranking
--answer-model	gpt-4o-mini	Model for answer generation
--prompt-id	(none)	Specific prompt ID (numeric)
--prompt	(none)	Prompt by slug or exact title
--no-track	off	Suppress interaction recording
--format	table	table or json

Examples:

# Basic RAG query
wp gg-data answer "What is dementia care?"

# With model overrides and JSON output
wp gg-data answer "What is dementia care?" --embedding-model=text-embedding-3-small --answer-model=gpt-4o --format=json

# Using a specific system prompt by ID
wp gg-data answer "Summarize our privacy policy" --prompt-id=42

---

Benchmark Commands

Benchmark commands validate and run the RAG quality benchmark. Designed for QA reviewers who need repeatable response validation.

Subcommands:

Subcommand	Purpose
wp gg-data benchmark run	Execute the full prompt matrix
wp gg-data benchmark list_prompts	List configured benchmark prompts
wp gg-data benchmark validate_config	Validate the benchmark configuration file
wp gg-data benchmark scorecard	Regenerate scorecard from existing artifacts

Flags:

Flag	Default	Description
--config	uploads-based path	Path to benchmark config JSON
--only	(none)	Run a single prompt by ID (e.g. P1)
--scorecard-scope	all	all or selected
--output-dir	uploads-based path	Artifact output directory

Examples:

# Validate config before running
wp gg-data benchmark validate_config

# Targeted single-prompt run
wp gg-data benchmark run --only=P1 --scorecard-scope=selected

# Regenerate scorecard without re-running
wp gg-data benchmark scorecard --from-dir=/path/to/artifacts

---

Evaluation Commands

Evaluation commands produce framework-ready evaluation datasets from pre-collected RAG prompts. Designed for QA engineers running downstream quality measurement.

Subcommands:

Subcommand	Purpose
wp gg-data evaluation run	Execute the full prompt corpus
wp gg-data evaluation list_prompts	List configured evaluation prompts
wp gg-data evaluation validate_config	Validate the evaluation configuration
wp gg-data evaluation adapters	List supported framework adapters

Flags:

Flag	Default	Description
--config	uploads-based path	Path to evaluation config JSON
--only	(none)	Run a single prompt by ID
--framework	config value	Override framework (e.g. ragas)
--output-dir	uploads-based path	Artifact output directory

Examples:

# List supported framework adapters
wp gg-data evaluation adapters

# Validate config, then run
wp gg-data evaluation validate_config
wp gg-data evaluation run --framework=ragas

# Single prompt for targeted validation
wp gg-data evaluation run --only=P1 --framework=ragas

---

Listing Commands

Listing commands discover the configured data connections and registered AI models. Use them before running sync, vectors, or answer commands to confirm valid names.

Commands:

Command	Purpose
wp gg-data list-connections	List configured data connections
wp gg-data list-models	List registered AI models

Flags:

Flag	Default	Description
--format	table	table, json, or csv
--with-embedding-models	off	Include embedding model context (connections)
--with-model-details	off	Include full model details (connections; implies --with-embedding-models)
--connection	gregius-data	Connection scope (models)
--type	(all)	Filter by model type: embeddings, llm, or rerank (models)

Examples:

# Discover connection names
wp gg-data list-connections --format=json

# List embedding models only
wp gg-data list-models --type=embeddings --format=table

---

Logs Commands

Logs commands browse, export, purge, and summarize plugin operational logs. Designed for routine maintenance and diagnostics.

Subcommands:

Subcommand	Purpose
wp gg-data logs list	View log entries with filters
wp gg-data logs export	Export filtered logs to CSV or JSON
wp gg-data logs purge	Delete logs older than N days
wp gg-data logs stats	Summary statistics

Flags:

Flag	Default	Description
--level	(all)	Filter by severity: debug, info, warning, error, critical
--component	(all)	Filter by source: rag, search, sync, vectors, connection, model, cron, system
--connection	(all)	Filter by data connection name
--limit	50	Max entries returned (list)
--days	30	Age threshold for purge (must be ≥ 1)
--yes	off	Skip confirmation prompt (purge)
--format	table (list, stats), csv (export)	table, json, or csv

Examples:

# View recent error logs
wp gg-data logs list --level=error --limit=20

# Export sync component logs as CSV
wp gg-data logs export --component=sync --format=csv

# Purge logs older than 90 days
wp gg-data logs purge --days=90 --yes

# Quick stats overview
wp gg-data logs stats --format=json

---

Permissions

Command	Required capability	Notes
sync	manage_options	Writes to PostgreSQL
vectors	manage_options	Generates and stores embeddings
answer	read	Reads via Abilities API
benchmark	manage_options	Executes RAG pipeline
evaluation	manage_options	Executes RAG pipeline
list-connections	manage_options	Reads connection configuration
list-models	manage_options	Reads model registry
logs	manage_options	Reads and manages operational logs

---