Represent the operational and canonical processing identity of a paper inside
victus-processing.
papers tracks the general processable state of a paper. It does not state
which artifacts exist for that paper.
paper_idpaper_id is the primary key for processing registry rows.paper_id must reference the same paper identity used by scientificpaper_id must not be derived from storage paths as its source of truth.papers is owned by victus-processing.
CREATE TABLE papers (
paper_id TEXT PRIMARY KEY,
status TEXT NOT NULL,
doi TEXT,
title TEXT,
source_fingerprint TEXT,
created_at TIMESTAMPTZ NOT NULL DEFAULT now(),
updated_at TIMESTAMPTZ NOT NULL DEFAULT now(),
deprecated_at TIMESTAMPTZ
);
| Field | Type | Description |
|---|---|---|
paper_id |
TEXT | Primary processing registry identifier for the paper. |
status |
TEXT | Current general processable state of the paper. |
doi |
TEXT / NULL | DOI when available. |
title |
TEXT / NULL | Paper title when available. |
source_fingerprint |
TEXT / NULL | Source-level fingerprint used for reconciliation or deduplication. |
created_at |
TIMESTAMPTZ | Registry row creation timestamp. |
updated_at |
TIMESTAMPTZ | Last registry row update timestamp. |
deprecated_at |
TIMESTAMPTZ / NULL | Timestamp when this paper registry row was deprecated. |
papers must:
papers must not store:
paper_id and status are required.paper_id must be unique.status must be one of the allowed states.doi, title, source_fingerprint, and deprecated_at may be null.deprecated_at must be set when status is deprecated.updated_at must change when the row is materially updated.registeredingestedprocessingprocessedpartially_processedfaileddeprecatedCreated when a paper is registered for processing.
Updated when the paper advances through processing state or metadata is
corrected.
Not deleted under normal operation.
Deprecated when the processing registry entry is superseded or should no longer
be used.
PaperPipelineRunsStageRunsProcessingArtifactsProcessingErrorspipeline_runs.paper_id -> papers.paper_idstage_runs.paper_id -> papers.paper_idartifacts.paper_id -> papers.paper_idprocessing_errors.paper_id -> papers.paper_idpapers is a registry table, not an artifact inventory. Artifact availability
must be determined from artifacts.
Documentation clarification or validation wording refinement.
Backward-compatible additions such as nullable columns or new non-breaking
status values.
Breaking schema changes, identity changes, status meaning changes, or field
removals.