Invocation#

bash

spis discover <start-url> --catalog <slug> --limit <n>

Discover important same-origin pages behind a start URL and scaffold selected catalog records.

Inputs and options#

<start-url>, --catalog <slug> and --limit <n> are required. --limit is how many pages you want recorded from this site; Spis chooses no default.
Brama is asked through MODEL_ROUTER_URL, the router alias MODEL_ROUTER_MODEL and the bearer the vault role MODEL_ROUTER_TOKEN_ROLE names; each missing one is refused by name.

Output and state effects#

Fetches the start page, extracts every same-origin page link on it, asks Brama which pages to record and in which family, records at most --limit of them, fetches thumbnails, and invokes the maintained Rust catalog and record scaffolding path. It is networked and mutating.

Refusals#

These messages are emitted by the current Rust implementation:

discover: the following arguments are required: start_url, discover: the following arguments are required: --catalog or discover: the following arguments are required: --limit (how many pages to record from this site)
Missing/nonnumeric option values are refused as discover: argument FLAG: expected one argument or discover: argument FLAG: invalid int value VALUE. Missing required arguments, missing option values and invalid integers exit 2; fetch or ranking failures exit 1.
Unknown options are discover: unrecognized argument VALUE.
Selection failures are exact: discover: no same-origin links found and discover: nothing selected for this corpus.