Explore

The explore command parses a file and lists the concepts, datasources, and imports in its environment. It answers "what can I query from this file?" without making you read the raw source.

This matters most on fact models, which import a handful of dimensions and can easily expose several hundred concepts. The default view collapses concepts by namespace, so a 378-concept import becomes roughly 25 group lines.

Usage

trilogy explore <path> [options]

Arguments

Argument	Description
path	Path to a Trilogy file to parse.

Options

Option	Description
--show [all|concepts|datasources|imports|groups]	Which section to print. Default groups.
--purpose TEXT	Filter concepts by purpose: key, property, metric, constant, rowset. Repeatable.
--regex TEXT	Case-insensitive regex filter over concept addresses. Repeatable, OR semantics.
--include-hidden	Include concepts normally hidden from public view.
--include-builtins	Include internal/builtin concepts.
--expand-imports	Render imported concepts in full instead of collapsing to a name-only listing.
--expand-roles	Render every role-played conformed dimension in full.

Full Examples

# What does this model expose?
trilogy explore raw/orders.preql

# The full flat concept table
trilogy explore raw/orders.preql --show concepts

# Everything, including datasources and imports
trilogy explore raw/orders.preql --show all

# Just the metrics
trilogy explore models/sales.preql --purpose metric

# Anything date-related
trilogy explore models/sales.preql --regex 'date\.(year|week_seq)'

# A bare word matches as a substring
trilogy explore models/sales.preql --regex customer

Sample output:

Available Concepts (1 namespaces, 4 concepts)

# (root)
order_id : enum<1, 2, 3>
    KEY
amount : float
    PROP @order_id
customer_id : enum<10, 11>
    PROP @order_id
order_date : date
    PROP @order_id

Collapsing Rules

Two defaults keep the output small, and both can be turned off:

Imports collapse. A fact file typically pulls in five to ten dimensions, each contributing dozens of concepts. Imported concepts are listed by name only so local declarations stay readable. --expand-imports renders them in full. Passing --regex also bypasses the collapse - when you are filtering, you want matching imports in full.

Role-played dimensions collapse. Namespaces parsed from the same source file with an identical shape - the eight date roles a fact plays, say - collapse to one canonical schema plus same_as references and a conformed map. This is most of the token savings on fact files. --expand-roles gives the literal per-role dump.

Notes

• --regex uses Python re.search, so full Python regex syntax works
• A malformed pattern exits non-zero with the underlying re.error message
• Pairs well with trilogy ingest: ingest a table, then explore it