1 lightning.finite_horizon.studies.inventory

lightning.finite_horizon.studies.inventory

Project archived scorer receipts into an immutable, reproducible inventory.

The TOML selection describes which historical observations to expose, rather than recording mutable workflow state. All scientific metadata comes from byte-verified archive files and explicit JSON pointers. A report-only run is not a fit, and a planned cell is not an executed attempt. This projection does not re-admit a fitted model or establish independent scene observations.

1.1 Classes

Name Description
InventoryField One metadata value read from authenticated historical evidence.
InventoryRowKind Observation roles; only actual fits and fitted controls count as attempts.
InventoryStatus Execution disposition, independent of scientific acceptance gates.
ScorerInventoryRecord One selected fit, control group, intention, or zero-update probe.
ScorerStudyInventory Receipt-derived snapshot with independently authenticated export provenance.

1.1.1 InventoryField

lightning.finite_horizon.studies.inventory.InventoryField(
    name,
    logical_path,
    historical_path,
    source_sha256,
    pointer,
    value_json,
)

One metadata value read from authenticated historical evidence.

1.1.1.1 Attributes

Name Description
name Scientific field name selected for the inventory projection.
logical_path Archive logical identity, independent of current filesystem location.
historical_path Original source path retained for historical readback.
source_sha256 Exact digest of the file containing the selected JSON value.
pointer RFC 6901 pointer within a JSON document or its JSONL row array.
value_json Immutable JSON serialization of the admitted value; null stays unknown.

1.1.1.2 Methods

Name Description
value Return a fresh projection; mutation cannot change admitted metadata.
to_jsonable Emit exact source references, optionally including the selected value.
1.1.1.2.1 value
lightning.finite_horizon.studies.inventory.InventoryField.value()

Return a fresh projection; mutation cannot change admitted metadata.

1.1.1.2.2 to_jsonable
lightning.finite_horizon.studies.inventory.InventoryField.to_jsonable(
    include_value=False,
)

Emit exact source references, optionally including the selected value.

1.1.2 InventoryRowKind

lightning.finite_horizon.studies.inventory.InventoryRowKind()

Observation roles; only actual fits and fitted controls count as attempts.

1.1.3 InventoryStatus

lightning.finite_horizon.studies.inventory.InventoryStatus()

Execution disposition, independent of scientific acceptance gates.

1.1.4 ScorerInventoryRecord

lightning.finite_horizon.studies.inventory.ScorerInventoryRecord(
    row_id,
    kind,
    status,
    evidence_role,
    fields,
    unknowns,
)

One selected fit, control group, intention, or zero-update probe.

1.1.4.1 Attributes

Name Description
row_id Unique stable selection identity; never used as a runtime attempt ledger.
kind Separates actual fits from grouped evidence and unexecuted intentions.
status Receipt-backed execution disposition or explicit frozen intention status.
evidence_role Direct receipt versus historical wrapper assertion qualification.
fields Immutable source-bound metadata, including available identity/budget fields.
unknowns Missing proof remains explicit rather than inferred from similar studies.

1.1.4.2 Methods

Name Description
field Select a documented field or fail on unavailable metadata.
to_jsonable Emit references without duplicating large source populations by default.
1.1.4.2.1 field
lightning.finite_horizon.studies.inventory.ScorerInventoryRecord.field(name)

Select a documented field or fail on unavailable metadata.

1.1.4.2.2 to_jsonable
lightning.finite_horizon.studies.inventory.ScorerInventoryRecord.to_jsonable(
    include_values=False,
)

Emit references without duplicating large source populations by default.

1.1.5 ScorerStudyInventory

lightning.finite_horizon.studies.inventory.ScorerStudyInventory(
    archive_manifest_sha256,
    selection_sha256,
    records,
)

Receipt-derived snapshot with independently authenticated export provenance.

1.1.5.1 Attributes

Name Description
archive_manifest_sha256 Externally pinned archive identity admitted by the transport owner.
selection_sha256 Exact human-authored selection bytes used for this projection.
records Stable ordered observations; repeated references to a fit are rejected.

1.1.5.2 Methods

Name Description
to_jsonable Export a portable inventory without making runtime or scientific claims.
1.1.5.2.1 to_jsonable
lightning.finite_horizon.studies.inventory.ScorerStudyInventory.to_jsonable(
    include_values=False,
)

Export a portable inventory without making runtime or scientific claims.

1.2 Functions

Name Description
read_scorer_inventory Admit an explicit TOML selection against an already verified study archive.

1.2.1 read_scorer_inventory

lightning.finite_horizon.studies.inventory.read_scorer_inventory(
    archive,
    selection_path,
)

Admit an explicit TOML selection against an already verified study archive.

Each field declares logical_path, sha256, and pointer. JSONL sources are treated as an ordered array of parsed rows. Optional per-row equal assertions compare two source references, rejecting identity or population mismatches before any output is returned. A source reference may declare schema to require that exact top-level JSON schema version.

Actual fits and fitted controls require a source-bound fit_identity or exact evidence receipt reference; normalized run IDs must be unique even when referenced through different report wrappers. Completed fits require positive realized_updates. Missing metadata must have a matching explicit unknown. Frozen intentions retain source-bound configuration but can never carry an executed fit ID.

Back to top