Document the v2 operator migration

This commit is contained in:
2026-09-07 04:55:13 +00:00
parent 2b97ef3722
commit a4f37ff427
19 changed files with 263 additions and 74 deletions

View File

@@ -44,8 +44,9 @@ the durable copied session input.
### Read-only effective pipeline inspection
`narratio config validate`, `narratio config show`, and `narratio config
sources` use the same `--config`, `--campaign`, `--campaign-file`, and
`narratio config validate`, `narratio config show`, `narratio config sources`,
and `narratio config artifacts` use the same `--config`, `--campaign`,
`--campaign-file`, and
`--profile` selection rules as pipeline commands, but do not select, discover,
or load a session. They resolve configured typed credential references as part
of the ordinary ConfigKit load but expose only their environment-variable names;
@@ -61,10 +62,12 @@ then parses the campaign-owned party and expands concrete artifacts and any
family publish rules before validation. `config validate` prints the resulting
root, profile, and effective digest. `config show` emits the normalized
effective pipeline YAML, with defaults and concrete expansion included but
ConfigKit directives and family declarations omitted. `config sources` prints a stable
ConfigKit directives and family declarations omitted. This review projection is
not promised to be a reloadable authoring document. `config sources` prints a stable
source projection instead of effective values: root/import/profile/default
ownership plus campaign/party and generated-family records. Canonical derived
players trace to the party. The [CLI reference](cli.md#config-validate-config-show-config-sources-and-config-diff)
players trace to the party. `config artifacts` lists the concrete artifact map,
effective dependencies, and defining sources. The [CLI reference](cli.md#config-validate-config-show-config-sources-config-artifacts-and-config-diff)
owns command syntax and output conventions.
`config diff` compares normalized field values rather than YAML text or source
@@ -78,14 +81,14 @@ difference.
### Migrating to the maintained bundle
Use the [production/testing bundle](../examples/production-testing/pipeline.yml)
as the complete copyable migration reference. Split stable pipeline settings
into explicit additive imports and place production/testing differences in the
fixed `profiles/production.yml` and `profiles/testing.yml` overlays. Convert
campaign
rosters to [canonical party input](integrations/party.md), then express
character work as families. Inspect the result
with `config validate`, `config show`, and `config sources`; use `config diff`
to review profiles before running a session. Unversioned party documents are
as the complete copyable migration reference. Keep production settings in the
base, split them into explicit additive imports, and place testing differences
in the fixed `profiles/testing.yml` overlay. Put ordinary artifacts and families
in the recommended one-definition files shown by that bundle. Convert campaign
rosters to [canonical party input](integrations/party.md), then express character
work as families. Inspect the result with `config validate`, `config show`,
`config sources`, and `config artifacts`; use `config diff default testing` to
review the overlay before running a session. Unversioned party documents are
rejected.
### Identity segments
@@ -116,6 +119,11 @@ remote state with an unsafe legacy identity must be migrated before use.
- Session files must be concrete; unresolved `{{ ... }}` placeholders fail load.
- Session and party inputs inspected before typed loading are limited to 8 MiB.
- Pipeline defaults are applied before validation.
- Defaults apply only to absent fields. Explicit empty strings, zero values,
`false`, and `null` retain ConfigKit's typed meaning and are then checked by
Narratio domain validation; they are not silently replaced by defaults.
Explicit null optional stage sections stay absent, while required sections
fail validation.
- Campaign and session identities must agree.
- Required stable files (`speakers_file`, `autocorrect_file`, `glossary_file`,
`party_file`) are campaign-owned; the first three may be overridden by the

View File

@@ -58,10 +58,12 @@ Configured definitions are always registered. Without an explicit selection,
the effective analyze set contains enabled definitions. With `--artifacts`, the
exact named configured definitions become the effective set for that invocation,
regardless of their `enabled` value. The effective-set resolver itself does not
expand dependencies; the analyze work planner closes those targets over their
configured prerequisite graph. Availability is separate from executability.
Configuration may normalize a family selection into its concrete generated
members before this resolver runs. The effective set retains optional family
expand dependencies. Configuration has already materialized the sorted union of
explicit ordering edges and configured-artifact input edges, and the analyze
work planner closes targets over that one concrete prerequisite graph.
Availability is separate from executability. Configuration may normalize a
family selection into its concrete generated members before this resolver runs.
The effective set retains optional family
and character origin metadata, but its keys, catalog sources, and runtime
lookups remain concrete configured-artifact identities.
Family publish policies are likewise expanded into ordinary configured-source

View File

@@ -53,13 +53,16 @@ the same typed loader and retain the existing post-load identity expectations,
stable-input precedence, and source-relative path semantics.
The application also has a separate read-only inspection resolver for `config
validate`, `config show`, and `config sources`. It uses the same production root/profile and
validate`, `config show`, `config sources`, and `config artifacts`. It uses the same production root/profile and
campaign selection functions, but never routes through session discovery,
remote-session download, secret loading, adapter composition, workspace
remote-session download, adapter composition, workspace
initialization, manifest access, or cleanup. A pipeline with retained artifact
family declarations must resolve its selected campaign before ordinary pipeline
validation, which expands its canonical-party members and generated publish
rules. A pipeline without those declarations may be validated by itself.
Secret-directory bootstrap remains invocation-scoped and happens before the
ordinary ConfigKit load because explicit typed references must resolve even for
inspection.
`MarshalEffectivePipeline` is the configuration-owned projection for `config
show`. It serializes one explicit typed safe projection with the existing YAML
@@ -77,6 +80,12 @@ Campaign records identify campaign-owned fields and party inputs; canonical
derived players point to the party source. The application command only joins these sorted
records with selection metadata and never reparses configuration files.
`EffectiveArtifactCatalog` projects the same finalized concrete artifact map for
`config artifacts`. It derives source identity from each map key, carries sorted
effective dependencies, aggregates ConfigKit definition contributors, and adds
family/character/party provenance for generated members. It does not inspect
files named after artifacts or infer another runtime graph.
`config diff` uses a thin paired loader that performs two independent ConfigKit
loads. When campaign
resolution is needed, the command loads one selected campaign and party and

View File

@@ -76,12 +76,15 @@ value from the last value persisted in the session manifest, which helps
diagnose profile switches without changing resume authority.
Before switching an operational profile, compare its effective meaning with the
current selection through `narratio config diff <left-profile> <right-profile>`.
production base through `narratio config diff default <profile>`.
The command is read-only and succeeds whether it finds differences or not. Its
sorted records describe defaulted, expanded concrete configuration—not source
file layout—so it can be used to review model, artifact, and publish changes
without creating a session or run. Select the same campaign explicitly when
profiles could resolve different campaign paths; see the [CLI reference](cli.md#config-validate-config-show-config-sources-and-config-diff) for syntax and record format.
without creating a session or run. Use `narratio config artifacts` to review the
concrete artifact identities, definitions, outputs, and dependencies. Select the
same campaign explicitly when profiles could resolve different campaign paths;
see the [CLI reference](cli.md#config-validate-config-show-config-sources-config-artifacts-and-config-diff)
for syntax and record format.
## Stage Execution and Continuation Behavior

View File

@@ -22,5 +22,7 @@ See the [release procedure](../release.md) for creating a candidate and tag.
When asynchronous publication succeeds, the corresponding Gitea release is the
canonical source for downloadable binaries and checksums.
- [v2.0.0](v2.0.0.md)
- [v1.7.0](v1.7.0.md)
- [v1.6.0](v1.6.0.md)
- [v1.5.0](v1.5.0.md)

92
docs/releases/v2.0.0.md Normal file
View File

@@ -0,0 +1,92 @@
# Narratio v2.0.0
Narratio v2.0.0 replaces its configuration infrastructure with ConfigKit,
simplifies the supported release platforms, and makes artifact definitions and
their effective dependency graph directly inspectable.
## Summary
Pipeline roots now use ConfigKit v0.1.1 for strict YAML loading, explicit
imports, fixed-path named profiles, provenance, defaults, and typed credential
references. The recommended production layout keeps production values in the
base, defines each ordinary artifact and artifact family in its own imported
file, and keeps testing changes in `profiles/testing.yml`.
Configured-artifact inputs now create their dependency edges automatically.
`narratio config artifacts` reports the finalized concrete catalog, including
generated party-family members, without loading a session or initializing
runtime adapters.
## Compatibility
This is a breaking configuration and platform release. There is no runtime
fallback for the retired Narratio composition/profile catalog, `pipeline.secrets`
or `secrets.env_dir`, credential `_env` fields, or the previous YAML
presence/default behavior. Defaults apply only to absent values; explicit empty,
zero, false, and null values retain their typed meaning and may fail Narratio
domain validation.
The maintained binaries are Linux amd64, Linux arm64, and macOS arm64. Windows
and macOS amd64 builds are no longer produced or supported. Existing v1 release
notes remain historical descriptions of their releases.
Inspection commands still avoid session, manifest, workspace, adapter, and
remote-service activity, but they are not credential-free loaders: every
explicit typed credential reference must resolve. Pipelines containing artifact
families also require a campaign so the canonical party can be expanded.
`config show` is a secret-safe review projection, not a promised reloadable
authoring file.
## Upgrade
Migrate each operational bundle before running v2:
1. Replace old composition syntax with root-owned
[`configkit.imports`](../config.md#pipeline-imports-and-profiles). Imports are
additive base sources; they do not override one another.
2. Fold the old production overlay into the base pipeline. Move testing changes
to `profiles/testing.yml`; compare the two independent resolutions with
`narratio config diff default testing`.
3. Follow the maintained
[`examples/production-testing`](../../examples/production-testing/pipeline.yml)
layout: one explicit artifact entry per `artifacts/<key>.yml` and one family
entry per `artifact-families/<key>.yml`. The map key, not the filename, is the
stable artifact identity.
4. Move credential files to an untracked `secrets/` directory beside
`pipeline.yml`, or pass `--secrets-dir`. Rename S3 fields to `access_key_id`
and `secret_access_key`, and rename the Audita field to `llm_api_key`.
5. For static S3 credentials, configure both S3 references (commonly
`OBJECT_STORAGE_KEY_ID` and `OBJECT_STORAGE_KEY`) and make them resolvable.
To use instance roles, workload identity, shared SDK files, or another AWS SDK
discovery provider, omit both references. A partial pair is invalid.
6. Remove a `depends_on` entry when it only duplicates a
`narratio.artifact.<key>` input. Keep explicit entries that express ordering
without supplying the upstream output as an input.
7. Review ConfigKit's stricter explicit-value/null semantics and correct values
that previously acquired defaults after being explicitly empty. Do not use
`config show` output as a source migration mechanism.
8. Run `config validate`, `config show`, `config sources`, and `config artifacts`
with the same profile, campaign, and secrets selection intended for runtime.
Resolve the campaign-owned party whenever families are present.
9. Update deployment and download automation to one of the three maintained
platform assets before removing the v1 binary.
The [configuration reference](../config.md), [CLI reference](../cli.md), and
[maintained examples](../../examples/README.md) are the canonical detailed
guides.
## Changes
- Adopted ConfigKit v0.1.1 across pipeline, campaign, session, and party loading
while keeping Narratio domain validation and post-campaign expansion explicit.
- Added ConfigKit imports and fixed `profiles/<name>.yml` pipeline overlays,
unified safe effective show/diff/digest projection, and source provenance.
- Added invocation-scoped ConfigKit secrets-directory bootstrap and typed S3 and
Audita credential references without exposing resolved values.
- Inferred, normalized, and validated configured-artifact input dependencies as
part of the one concrete graph consumed by analyze planning and resume logic.
- Added `config artifacts` with stable concrete identities, definition sources,
effective dependencies, and generated family/party provenance.
- Added the maintained production-base/testing-overlay artifact authoring layout.
- Reduced release assets and active platform support to Linux amd64/arm64 and
macOS arm64.

View File

@@ -426,7 +426,13 @@ Document the command in its canonical CLI owner and link to config layout exampl
## Stage 10 — Finish the operator layout and v2 migration documentation
Status: Pending
Status: Complete
Completion: Reworked the maintained bundle around an unprofiled production base,
one-entry artifact and family imports, and a testing overlay; updated the example
harness and inspection workflows; consolidated the v2 configuration guidance; and
added the v2.0.0 operator migration checklist. Focused documentation/example tests,
the offline inspection walkthrough, and the full normal/race suites pass.
Read: roadmap “Artifact Definition Ergonomics” and “Compatibility and Migration”;
`examples/README.md`, existing example bundles and their loader test,