Retire Chronicle target profiles

.github/chronicle-agents.yml

@@ -19,20 +19,6 @@ approved_agents:
     required_judges:
       - ledger-source-fidelity
       - ledger-boundary
-  - id: ledger-target-profile-author
-    purpose: Add source-backed target profiles and model measurement contracts without target values.
-    allowed_paths:
-      - policyengine_chronicle/target_profiles/**
-      - chronicle/targets/**
-      - tests/test_policyengine_chronicle_target_profiles.py
-      - tests/test_us_poverty_target_coverage.py
-    required_deterministic_checks:
-      - no embedded target values
-      - sum-only operation
-      - selector references raw Chronicle facts
-    required_judges:
-      - ledger-target-profile
-      - ledger-boundary
   - id: ledger-contract-maintainer
     purpose: Change Chronicle schemas, identity, provenance, or consumer contracts.
     allowed_paths:
@@ -54,8 +40,6 @@ approved_agents:
 required_judges:
   ledger-source-fidelity:
     verdict: PASS if every new fact is directly traceable to publisher bytes/cells and no source values are invented.
-  ledger-target-profile:
-    verdict: PASS if profiles contain selectors and measurement contracts only, with no target values or active calibration decisions.
   ledger-contract:
     verdict: PASS if schema or identity changes preserve source provenance and do not move Microcosm responsibilities into Chronicle.
   ledger-boundary:

.github/pull_request_template.md

@@ -2,17 +2,15 @@
 
 ## Chronicle Governance
 
-For any source package, target profile, consumer contract, schema, or source-data
+For any source package, consumer contract, schema, or source-data
 boundary change:
 
 - Approved Chronicle agent role:
   - `ledger-source-ingestor`
-  - `ledger-target-profile-author`
   - `ledger-contract-maintainer`
 - Deterministic checks run:
 - LLM judge verdicts:
   - `ledger-source-fidelity`:
-  - `ledger-target-profile`:
   - `ledger-contract`:
   - `ledger-boundary`:
 

.github/workflows/ci.yml

@@ -58,7 +58,6 @@ jobs:
           import policyengine_chronicle
           import policyengine_chronicle.normalization
           import policyengine_chronicle.sources
-          import policyengine_chronicle.target_profiles
           import policyengine_chronicle.targets
           from policyengine_chronicle.schema import (
               CONSUMER_FACT_SCHEMA_SHA256,

AGENTS.md

@@ -1,7 +1,8 @@
 # Chronicle Agent Rules
 
 Chronicle is a source-backed fact store. It may parse publisher artifacts, normalize
-representation, preserve provenance, and declare target-profile contracts.
+representation, and preserve provenance. Selection and measurement contracts live
+in consumers such as Microcosm.
 
 Every fact value must trace to a publisher. The boundary is who asserted the
 value, not level versus projection: a publisher's own projection (CBO baseline,
@@ -16,13 +17,13 @@ Do not put Microcosm work in Chronicle:
 - no imputation
 - no support-aware target activation
 - no solver-ready target construction
-- no target values in target profiles
+- no target profiles or model-measurement bindings
 - no PolicyEngine-computed values stored as facts
 
-Resolving profile targets at a period other than a fact's reference period
-requires the consumer's explicit `PeriodAlignmentDeclaration`; Chronicle records
-the declaration and returns the published level, never the aligned number.
+Chronicle records every fact's publisher reference period. Consumers own and
+enforce any declaration that aligns those facts to another period; Chronicle
+returns the published level and never an aligned number.
 
 Only approved Chronicle agent roles in `.github/chronicle-agents.yml` should add or
-modify source packages, target profiles, or contract schemas. Source-data PRs
+modify source packages or contract schemas. Source-data PRs
 need deterministic validation plus the listed Chronicle judge reviews before merge.

LANE_C2_REPORT.md

@@ -13,8 +13,8 @@ The cached PDF resolves three material ambiguities in the task shorthand:
 
 - The Annex columns put most requested External series in publisher periods
   2021–2023, not 2022–2024. Chronicle preserves the printed periods. A consumer
-  that needs another period must supply an explicit `PeriodAlignmentDeclaration`;
-  Chronicle returns the published level and never an aligned number.
+  that needs another period owns the contract for aligning it; Chronicle
+  returns the published level and never an aligned number.
 - The `poa=2,273k` and `psu=79k` observations are in the SILC column. Their
   External cells are `NaN`, so they are stored as `survey_aggregate` and are
   validation-only.
@@ -439,8 +439,8 @@ acceptance, and both required Chronicle judge reviews pass.
   `Add Belgium national accounts and JRC external facts`.
 - No push and no R2 upload were performed. The two manifests merely declare the
   required content-addressed R2 keys.
-- A consumer using a fact outside its publisher period must supply an explicit
-  `PeriodAlignmentDeclaration`; Chronicle returns the published value and does
-  not compute an aligned value.
+- A consumer using a fact outside its publisher period owns the alignment
+  contract; Chronicle returns the published value and does not compute an
+  aligned value.
 
 LANE C2 DONE

LANE_C5_REPORT.md

@@ -355,7 +355,7 @@ a within-source pair ending in 2025 in this lane.
 Any choice of proxy family, ratio calculation, aging, cross-source mapping,
 period alignment, target activation, or solver construction belongs to the
 consumer. When a target period differs from a fact reference period, the
-consumer must provide its explicit `PeriodAlignmentDeclaration`; Chronicle
-returns the published levels, never the aligned number.
+consumer-owned contract must declare the alignment; Chronicle returns the
+published levels, never the aligned number.
 
 LANE C5 DONE

README.md

@@ -10,15 +10,13 @@ values as structured, queryable facts.
 
 Chronicle may normalize structure: parse files, type values, declare units and
 scales, assign geography and period identifiers, preserve lineage back to
-source artifacts, and publish target profiles that identify source-backed facts
-and measurement contracts. Chronicle does not reconcile inconsistent sources,
-impute missing data, store raw survey microdata, or execute simulator-specific
-calibration.
+source artifacts, and publish source-backed facts. Chronicle does not own
+selection or measurement contracts, reconcile inconsistent sources, impute
+missing data, store raw survey microdata, or execute simulator-specific calibration.
 
-Microcosm consumes Chronicle facts and target profiles, selects the subset its
-current support universe can target, applies minimal period alignment when
-declared, and runs calibration. Thesis can consume the same facts and
-measurement contracts as official observations.
+Microcosm consumes Chronicle facts, owns the contracts that select and bind
+them, applies declared period alignment, and runs calibration. Thesis can
+consume the same facts as official observations.
 
 ## Purpose
 
@@ -35,11 +33,8 @@ This repository provides:
   their `survey_instrument`.
 - **Normalization**: Low-assumption representation changes such as unit/scale
   conversion and source-published total/share arithmetic.
-- **Target profiles**: Source-backed target contracts and model-measurement
-  bindings that Microcosm, Thesis, and future rule engines can consume.
 - **Consumer artifacts**: Versioned, reproducible bundles of consumer-contract
-  fact rows plus profiles (`chronicle build-consumer-artifact`) with a resolution
-  API that enforces the period contract.
+  fact rows plus manifest hashes (`chronicle build-consumer-artifact`).
 - **Jurisdiction loaders**: Source-specific ETL that emits the shared Chronicle
   schema.
 
@@ -50,27 +45,24 @@ They are source-backed claims with provenance.
 
 The load-bearing rule:
 
-> Chronicle may re-express a published value and declare target contracts, but may
-> not reconcile, impute, or transform published values in ways that change their
-> meaning.
+> Chronicle may re-express a published value, but may not select it for a
+> consumer or transform it in ways that change its meaning.
 
 The store is facts-only, and the line is who asserted the value. Everything a
 publisher asserted — including the publisher's own projections — is a fact.
 Everything PolicyEngine computes (aged, uprated, forecast, or reconciled
 levels) is a downstream build artifact and never enters the store; Microcosm
 owns aging as a named, versioned model over Chronicle growth-factor facts. A
-fact's `period` is the period its value refers to, and resolving a target at
-any other period hard-fails without an explicit consumer
-`PeriodAlignmentDeclaration` — the guard against silent un-aged calibration
-(see [`docs/adr-chronicle-facts-only.md`](docs/adr-chronicle-facts-only.md)).
+fact's `period` is the period its value refers to. Consumers must enforce any
+contract that aligns it to another period (see
+[`docs/adr-chronicle-facts-only.md`](docs/adr-chronicle-facts-only.md)).
 
 | Layer | Owns | Examples |
 |-------|------|----------|
 | Chronicle Sources | Source artifacts and provenance | URLs, checksums, source files, parsed tables/cells |
 | Chronicle Facts | Structured source claims | SOI cells, ACS estimates, CPI values, CBO-published projections |
 | Chronicle Normalization | Representation changes | Unit scales, typed values, geography/date identifiers |
-| Chronicle Target Profiles | Source-backed calibration contracts | SOI EITC totals, CBO baselines, source-published growth factors, measurement bindings |
-| Microcosm Targets | Build-ready active subset | Support-aware activation, solver inputs, diagnostics |
+| Microcosm Target Contracts | Selection, measurement bindings, and active subset | Period alignment, support-aware activation, solver inputs, diagnostics |
 
 The storage split is documented in
 [`docs/storage-architecture.md`](docs/storage-architecture.md): `ledger-raw`
@@ -126,8 +118,8 @@ chronicle/
 └── docs/                    # Architecture and source documentation
 ```
 
-New code should prefer `policyengine_chronicle` for source-backed fact and target
-profile consumers. Existing in-repo implementation code may continue using
+New code should prefer `policyengine_chronicle` for source-backed fact
+consumers. Existing in-repo implementation code may continue using
 legacy implementation modules while the namespace migration is completed.
 Solver execution and calibrated dataset construction belong in Microcosm.
 
@@ -304,6 +296,11 @@ uv run chronicle build-bundle --suite uk --out /tmp/chronicle-uk --replace
 uv run chronicle build-consumer-artifact --facts /tmp/chronicle-uk --out /tmp/chronicle-uk-artifact --replace
 ```
 
+The command writes a `policyengine_ledger.consumer_artifact.v2` artifact containing
+only `consumer_facts.jsonl` and `manifest.json`. Version 2 is incompatible with the
+retired v1 profile-bearing contract: loaders reject v1 manifests so downstreams must
+adopt the facts-only surface explicitly.
+
 `--year` is inert for `--suite uk` because the UK packages are year-pinned.
 The US off-year bundle behavior is unchanged and out of scope here.
 
@@ -493,10 +490,8 @@ LEDGER_EXPLORER_DATA_DIRS=/tmp/chronicle-build-a,/tmp/chronicle-build-b npm run
 
 ```python
 from policyengine_chronicle.targets import DataSource, Target, TargetType, query_targets
-from policyengine_chronicle.target_profiles import load_target_profile
 
 target_rows = query_targets(jurisdiction="us", year=2024)
-profile = load_target_profile("us_fiscal")
 ```
 
 ## Target Input Schema
@@ -508,8 +503,8 @@ Target inputs use a three-table schema:
 - **stratum_constraints**: Rules defining each stratum.
 - **targets**: Source-published aggregate values linked to strata.
 
-These are inputs to Chronicle target profiles. Microcosm owns the active
-support-aware subset and calibrated solver execution.
+These are source-backed inputs. Microcosm owns the contracts that select them,
+the active support-aware subset, and calibrated solver execution.
 
 ## Chronicle Facts And Microcosm Targets
 
@@ -521,8 +516,8 @@ publishes the total/share relationship.
 
 Inflation, cross-source reconciliation, and support-aware activation belong in
 Microcosm unless the source itself publishes the adjusted or projected series.
-Target profiles in Chronicle may declare the source-backed rows and measurement
-bindings Microcosm is allowed to activate.
+Microcosm contracts declare which source-backed rows and measurement bindings a
+build may activate.
 
 ```python
 from policyengine_chronicle.facts import SourceFact
@@ -558,10 +553,10 @@ normalized_fact = convert_units(fact, 1000, "count")
 ## Boundaries
 
 - **Chronicle** owns government-statistics release artifacts, provenance, source
-  facts, aggregate facts, target profiles, and measurement contracts.
-- **Microcosm** owns support-aware target activation, minimal period alignment,
-  raw microdata access, simulation interfaces, entity modeling, weights,
-  diagnostics, and calibration execution.
+  facts, and aggregate facts.
+- **Microcosm** owns selection and measurement contracts, support-aware target
+  activation, period alignment, raw microdata access, simulation interfaces,
+  entity modeling, weights, diagnostics, and calibration execution.
 - **Jurisdiction source packages** such as `ledger-us` and `chronicle-uk` own
   source-specific parsers and specs that emit shared Chronicle records.
 - **Jurisdiction simulation packages** own simulation-specific variable

chronicle/__init__.py

@@ -1,7 +1,7 @@
 """Chronicle source-data foundation.
 
 Chronicle owns government-statistics releases: source artifacts, source-backed
-facts, constraints, provenance, and target profiles. Raw microdata storage,
+facts, constraints, and provenance. Selection contracts, raw microdata storage,
 source reconciliation, aging, imputation, target activation, and calibration
 belong in downstream systems such as Microcosm.
 """

chronicle/core.py

@@ -82,21 +82,6 @@
 }
 ALLOWED_ASSERTIONS = {"observation", "source_projection"}
 DEFAULT_ASSERTION = "observation"
-# How profile resolution treats the assertion axis. observed_only is the safe
-# default: projections are invisible and a projection-only family fails loudly.
-# prefer_observed applies per series (one geography/entity/dimension tuple):
-# a series with any observed fact resolves only from observations, and a
-# series with none may fall back to projections — no single series ever
-# mixes bases across periods, and a projection-only series is never starved
-# by a neighbouring series' observation. allow_source_projection treats both
-# equally (for forecast families such as the OBR EFO lines), with the
-# observation winning an exact-period tie within a series.
-ASSERTION_POLICIES = {
-    "observed_only",
-    "prefer_observed",
-    "allow_source_projection",
-}
-DEFAULT_ASSERTION_POLICY = "observed_only"
 ALLOWED_PROVENANCE_CLASSES = {
     "administrative",
     "census",

chronicle/harness.py

@@ -754,30 +754,15 @@ def main(argv: list[str] | None = None) -> int:
 
     consumer_artifact_parser = subparsers.add_parser(
         "build-consumer-artifact",
-        help=(
-            "Build a versioned consumer artifact from consumer-contract facts "
-            "and Chronicle target profiles"
-        ),
+        help="Build a versioned facts-only consumer artifact",
+        description="Build a versioned facts-only consumer artifact.",
     )
     consumer_artifact_parser.add_argument(
         "--facts",
         type=Path,
         required=True,
         help="Path to a consumer_facts.jsonl file or a bundle directory",
     )
-    consumer_artifact_parser.add_argument(
-        "--profile",
-        action="append",
-        default=[],
-        help="Packaged target profile id (may be repeated)",
-    )
-    consumer_artifact_parser.add_argument(
-        "--profile-path",
-        action="append",
-        type=Path,
-        default=[],
-        help="Path to a target profile JSON file (may be repeated)",
-    )
     consumer_artifact_parser.add_argument(
         "--out",
         type=Path,
@@ -1285,8 +1270,6 @@ def main(argv: list[str] | None = None) -> int:
         artifact_report = build_consumer_artifact(
             args.out,
             facts_path=args.facts,
-            profile_ids=args.profile,
-            profile_paths=args.profile_path,
             replace=args.replace,
         )
         print(json.dumps(artifact_report.to_dict(), indent=2, sort_keys=True))

docs/adr-chronicle-facts-only.md

@@ -20,22 +20,19 @@ the value**, not level versus projection:
    aging), implemented as named, versioned models that consume growth-factor
    facts from Chronicle and emit their own lineage.
 
-Instead of projection objects, Chronicle contributes three guarantees:
+Instead of projection objects, Chronicle contributes two guarantees:
 
 - **Reference-period semantics.** `PeriodDimension` identifies the period a
   value refers to; `PeriodCoverage` records non-identity provenance (start
   and end dates, basis, the publisher's period label, accounting basis) for
   cases like BE-SILC incomes that reference the year before the survey
   label.
-- **Period-contract enforcement.** Resolving a profile target at a period
-  other than the fact's reference period raises `PeriodContractError`
-  unless the consumer passes an explicit `PeriodAlignmentDeclaration`
-  (model id, version, parameters — never values). Chronicle records the
-  declaration in resolved rows and returns the published level untouched.
-- **Basis-aware diagnostics.** Resolved rows carry `basis` (`fact` or
-  `declared_alignment`), `fact_period`, `requested_period`, and the
-  declaration, so downstream diagnostics distinguish "missed a published
-  fact" from "missed an aged level."
+- **Facts-only consumer artifacts.** Chronicle publishes schema-validated fact
+  rows with manifest hashes. Consumers own the selection, measurement,
+  period-alignment, and model-binding contracts that interpret those rows. The
+  facts-only artifact is `policyengine_ledger.consumer_artifact.v2`; the version
+  bump makes the removal of v1's embedded profiles and resolution surface an
+  explicit incompatible transition.
 
 ## Why not facts plus projections in one schema
 
@@ -67,9 +64,13 @@ Instead of projection objects, Chronicle contributes three guarantees:
   values other than `observation` and `source_projection` fail validation
   with an error explaining that PolicyEngine-computed values are not facts.
 - Consumer-contract rows always carry `assertion` explicitly, and the
-  consumer artifact (`chronicle build-consumer-artifact`) embeds profiles,
-  fact rows, coverage diagnostics, and manifest hashes so Microcosm can
-  build a target registry without database access or copied values
-  (issue #61).
+  consumer artifact (`chronicle build-consumer-artifact`) contains only fact
+  rows and the manifest hashes needed to verify them. Microcosm packages its
+  own selection contracts and builds its target registry without Chronicle
+  profiles (issues #166 and #172).
+- The retired `policyengine_ledger.target_profile.v1` and
+  `policyengine_ledger.resolved_target.v1` schema IDs have no v2 successor in
+  issue #143. The Chronicle-side Belgian profile plan in issue #70 is
+  superseded; Belgian contracts also live consumer-side.
 - Geography vintage translation (microcosm#205) follows the same pattern: a
   declared consumer-side transform over facts, never an edit to them.