What a snapshot claims about itself

Every urpx:RatePlanVersion and urpx:RatePlanModifierVersion carries a urpx:CoverageClaim: a verdict on what can be answered from that snapshot. It is the standard's own answer to "what am I holding?", and urpx.coverage reads it.

(require '[urpx.coverage :as cov])

(cov/of-version version)          ; => a report map, keys under :urpx.coverage/
(cov/of-version version doc)      ; doc follows a bare @id reference and may state the zone
(cov/of-version version doc zone) ; zone reads dateTimes written without an offset
(cov/claim version)               ; => the raw CoverageClaim node, or ::absent

(cov/of-model model)            ; => every part's report, plus counts
(cov/descriptions model-report) ; => the Evaluation prose, in part order

(cov/coverage-interval report zone)        ; => {:tick/beginning .. :tick/end ..}
(cov/verdict-applies? report zone interval) ; => does the verdict cover that window?

The verdict is not this library's judgement. Per the ontology, claims "are produced by an implementation or publishing workflow", so a claim says what the publisher asserts. Nothing here re-derives one or second-guesses one.

urpx.coverage requires urpx.index and nothing else of the library: not urpx.price, not urpx.scenario-cost-model. Nothing in src/ requires it.

Three things a caller should know

Absent is not unknown. A version with no claim reports :stated? false and :completeness nil, never "urpx:unknown". urpx:unknown is itself a verdict, meaning no evaluation has run, so a document asserting it has said something a document omitting the claim has not.

selfContained true is not "complete", and can be vacuous. It is derived as "false iff any PriceDefinitionReference is retrievedElsewhere", so on a snapshot with no such reference the existential ranges over an empty set. It means only "nothing here is fetched from elsewhere".

A verdict states no window. The CoverageClaim note says producers should attach a TemporalScope with scopeKind urpx:coverage saying what the verdict covers, and no shape requires one. So a consumer publishing a horizon cannot assume a verdict covers it. :scope-stated? and :coverage-window are separate for the same reason: scopeKind is the only required field on a scope, so one can name its role and bound nothing.

Conformance, as distinct from validity

Every version snapshot must carry urpx:hasCoverageClaim; the schema requires it. urpx.validate also reports one rule the standard states that the schemas do not enforce, :self-contained-disagrees: urpx:selfContained may appear on the claim and the version and "MUST agree". urpx.coverage reads such a document without refusing, because the same note settles which wins; the violated MUST is the validator's business. It prints as NONCONFORMANT and counts toward the exit code.

What it refuses

:reason	When
:urpx.coverage/unknown-completeness-result	the claim names a member outside the six. Reporting nil would read as "not complete", a guess about a verdict the document did state
:urpx.coverage/claim-ref-unresolved	the claim is a bare @id that cannot be followed. A reference going nowhere is not the same as no claim
:urpx.coverage/self-contained-contradicts-references	the claim asserts selfContained true while a PriceDefinitionReference is retrievedElsewhere, contradicting the property's own derivation with no stated precedence
:urpx.coverage/ambiguous-coverage-scope	more than one scope names urpx:coverage. A verdict covers one window and nothing says which is meant
:urpx.coverage/coverage-scope-absent	coverage-interval or verdict-applies? asked about a claim with no urpx:coverage scope, which no shape requires
:urpx.coverage/coverage-scope-unbounded	a coverage scope naming its role and no bounds, which scopeKind being the only required field permits. It asserts a role, not a window
:urpx.coverage/zone-required	a coverage window was asked for with no ZoneId. URPX states the bounds without one and any default would be this library's invention

Claim-level and version-level selfContained disagreeing does not refuse. It is reported with :self-contained-disagrees? true and the claim winning, because the ontology settles precedence ("when both are present they MUST agree, and the claim is authoritative"); it is a violated MUST, which is a validation concern rather than a reading one.