Every ticket as the root of its provenance graph — goal, patch, approval, runs and (as the ticket-class build lands) mandate, manifestations, guarantees, journey, artifacts and children. Rendered through the <<<TicketGraphList>>> tag.
Research: design the similarity search — code chunking, embeddings (nomic-embed-text), ANN, ticket join
goal
Design the pipeline: (1) chunk the forge code corpus (function/class-level via codebean's lower_* AST importer #183, not naive line windows); (2) embed with nomic-embed-text via litellm (#184); (3) ANN index over embeddings + omni-git's blob/occurrence index for exact/structural matches (hybrid semantic+lexical); (4) JOIN each hit to its ASSOCIATED TICKETS (via memgraph edges / ticket_works refs / commit->ticket links). Input forms: a raw snippet, or path:line-range. OUTPUT per hit: path, line-range, repository, associated tickets, similarity score. Spec the storage + query interface.
patch
none
approval
unapproved
runs
no runs recorded
mandate (clauses)
not yet recorded — lands with the ticket-class build
manifestations
not yet recorded — lands with the ticket-class build
guarantees
not yet recorded — lands with the ticket-class build
journey (blunders & successes)
not yet recorded — lands with the ticket-class build
artifacts (forge)
not yet recorded — lands with the ticket-class build