System boundry and persistance workplan

This commit is contained in:
2026-05-05 10:37:01 +02:00
parent 9afc3800e9
commit aed08061f7
3 changed files with 334 additions and 0 deletions

View File

@@ -12,6 +12,7 @@ Start here:
- `docs/stack-decision.md`
- `docs/markitect-main-scope-assessment.md`
- `docs/markitect-tool-reuse-boundary.md`
- `docs/phase-memory-boundary.md`
- `docs/system-layer-extraction-inventory.md`
- `docs/system-layer-migration-backlog.md`
- `docs/service-api-boundary.md`

View File

@@ -0,0 +1,107 @@
# Phase Memory Boundary
Date: 2026-05-05
This note defines how `kontextual-engine` should relate to the sibling
`phase-memory` project while closing the current persistence gap.
## Source Scope
`phase-memory` exists to provide profile-driven memory infrastructure for
agentic systems. Its intent covers memory phases, memory graphs, event paths,
stores, compaction, retention, service profiles, activation planning, and
adapters for advanced memory backends.
`kontextual-engine` exists to provide an AI-first, headless knowledge and
content engine. Its intent covers persistent knowledge artifacts, collections,
relationships, ingestion, workflow state, query, service access, and agent-
operable knowledge processes.
The projects overlap around durable state, graph-shaped records, context
packages, provenance, and agent operation. They must stay separated by purpose.
## Ownership Boundary
`kontextual-engine` owns:
- Persistent knowledge collections, artifacts, relationships, and revisions.
- Content identity, artifact metadata, source provenance, and lifecycle state.
- Durable workflow run records and operation manifests.
- Query and retrieval over knowledge artifacts and relationships.
- Service and programmatic APIs for operating knowledge systems.
- Adapter surfaces that can call `markitect-tool`, `phase-memory`, storage
backends, and future workflow engines.
`phase-memory` owns:
- Ephemeral, fluid, stabilized, and rigid memory phase semantics.
- Reasoning graphs, conversational paths, activation memory, identity memory,
preference memory, and reviewed memory records.
- Memory profiles with latency, token, byte, retention, decay, compaction,
freshness, merge, conflict, and policy parameters.
- Memory-specific lifecycle enforcement, including retention, deletion,
stabilization, compaction, and summarization.
- Memory retrieval planning and activation-package compilation.
- Adapters to graph databases, vector stores, audit sinks, policy systems, and
memory-specific backend infrastructure.
## Interop Boundary
`kontextual-engine` may persist memory-related artifacts when they are treated
as knowledge assets, such as:
- reviewed decisions,
- source-backed facts,
- durable project context packages,
- workflow outputs,
- policy-visible provenance records,
- exported memory graph slices.
It should not implement memory-specific behavior such as:
- conversational path branching and merge semantics,
- memory decay and retention policy execution,
- compaction or summarization strategies,
- activation planning heuristics,
- preference or identity memory governance,
- graph/vector memory backend selection,
- memory service profile validation.
Those behaviors belong to `phase-memory`. `kontextual-engine` should integrate
with them through explicit adapters and stable exchange formats.
## Persistence Implication
The persistence gap in `kontextual-engine` should be closed as durable knowledge
runtime storage, not as a general memory runtime.
The first durable backend should persist:
- collections,
- artifacts,
- artifact revisions,
- relationships,
- workflow runs,
- run manifests,
- change records,
- queryable metadata,
- context-package references.
It should not persist phase-memory internals beyond opaque references or
exported artifacts unless an integration adapter explicitly maps them.
## Integration Shape
The preferred integration shape is:
1. Core `kontextual-engine` repository contracts remain independent.
2. A future `phase-memory` adapter can read/write durable knowledge artifacts
through those contracts.
3. Context packages can move between the systems as inspectable payloads.
4. Memory-specific policies stay in `phase-memory`; knowledge persistence
policy stays in `kontextual-engine`.
5. Cross-repo references should use stable identifiers, content digests, and
provenance metadata rather than implicit filesystem coupling.
This keeps `kontextual-engine` focused on persistent, operable knowledge while
leaving sophisticated agentic memory behavior to the project designed for it.

View File

@@ -0,0 +1,226 @@
---
id: KONT-WP-0004
type: workplan
title: "Durable Persistence Foundation"
domain: markitect
repo: kontextual-engine
status: planned
owner: codex
topic_slug: markitect
created: "2026-05-05"
updated: "2026-05-05"
state_hub_workstream_id: pending
---
# KONT-WP-0004: Durable Persistence Foundation
## Purpose
Close the persistence gap identified after `KONT-WP-0003` by turning the current
in-memory repository contract into a durable, local-first storage foundation for
knowledge artifacts, collections, relationships, workflow state, and context
references.
This workplan deliberately does not implement `phase-memory` behavior. It uses
`docs/phase-memory-boundary.md` as the boundary: `kontextual-engine` persists
durable knowledge runtime state; `phase-memory` owns agentic memory phases,
profiles, compaction, retention, and activation planning.
## Persistence Scope
In scope:
- Durable storage for collections, artifacts, artifact revisions, and
relationships.
- Durable storage for workflow runs and run manifests.
- Explicit update and delete behavior for artifacts and relationships.
- Change records that make artifact evolution inspectable.
- Query support for identifiers, names, digests, metadata, content text, and
relationships.
- Local-first SQLite backend with a repository interface that can later be
backed by service, PostgreSQL, graph, or object-storage adapters.
- Tests proving data survives repository re-instantiation.
Out of scope:
- Memory-phase lifecycle behavior from `phase-memory`.
- Vector search, embedding storage, and memory activation planning.
- Markdown parsing or markdown transformations; use `markitect-tool` adapters.
- LLM provider execution; use future `llm-connect` adapters.
- Remote multi-tenant deployment concerns beyond schema choices that do not
block later migration.
## P4.1 - Finalize persistence boundary and ADR
```task
id: KONT-WP-0004-T001
status: todo
priority: high
```
Promote the storage decision from deferred to explicit: local-first SQLite for
the first durable backend, wrapped by repository contracts. Decide whether the
implementation uses direct `sqlite3` or SQLAlchemy for this slice.
Output: update `docs/stack-decision.md` or add an ADR under `docs/`.
Acceptance:
- The backend choice is explicit and justified.
- The decision references `docs/phase-memory-boundary.md`.
- Future service-backed storage remains possible.
## P4.2 - Complete repository contract semantics
```task
id: KONT-WP-0004-T002
status: todo
priority: high
```
Extend `KnowledgeRepository` from create/list/get into a durable lifecycle
contract. Define update, delete, revision, and transaction semantics without
binding callers to a specific backend.
Output: `src/kontextual_engine/storage.py` and focused tests.
Acceptance:
- Artifact update produces explicit revision/change semantics.
- Artifact delete behavior is defined for relationships and query results.
- Duplicate-name and referential-integrity behavior remains deterministic.
- Existing in-memory tests continue to pass.
## P4.3 - Design durable schema and migrations
```task
id: KONT-WP-0004-T003
status: todo
priority: high
```
Create the first durable schema for collections, artifacts, revisions,
relationships, workflow runs, run manifests, and change records.
Output: schema/migration files under `src/kontextual_engine/storage/` or an
equivalent package-owned location.
Acceptance:
- Schema stores content digest, artifact type, size, metadata, timestamps, and
provenance.
- Relationships enforce valid source and target artifacts.
- JSON metadata is preserved roundtrip.
- Migrations can initialize an empty local database deterministically.
## P4.4 - Implement SQLite repository backend
```task
id: KONT-WP-0004-T004
status: todo
priority: high
```
Implement `SQLiteKnowledgeRepository` behind the same repository contract used
by the in-memory backend.
Output: durable repository implementation and tests.
Acceptance:
- Collections, artifacts, relationships, and metadata survive closing and
reopening the repository.
- Query behavior matches the in-memory repository for supported filters.
- Tests cover duplicate artifact names, missing relationship endpoints, and
deterministic ordering.
- No markdown or memory-runtime logic is introduced.
## P4.5 - Persist artifact evolution
```task
id: KONT-WP-0004-T005
status: todo
priority: high
```
Add artifact revision and change-record support so persistent knowledge can be
versioned and audited over time.
Output: model additions, repository methods, and tests.
Acceptance:
- Updating an artifact records old and new digests.
- Revision history can be retrieved by artifact id.
- Deletion is traceable through a change record.
- Change records are backend-neutral at the programmatic API boundary.
## P4.6 - Persist workflow run state
```task
id: KONT-WP-0004-T006
status: todo
priority: medium
```
Persist `OperationRun`, `WorkflowStep`, `InputBundle`, and `RunManifest`
records so orchestration can resume and inspect prior execution.
Output: repository methods and persistence tests for workflow records.
Acceptance:
- Run status transitions survive repository re-instantiation.
- Run manifests roundtrip with inputs, outputs, diagnostics, and timestamps.
- Artifact outputs can be linked to producing runs.
## P4.7 - Add context and phase-memory reference hooks
```task
id: KONT-WP-0004-T007
status: todo
priority: medium
```
Add lightweight persistence for context-package references and external memory
references without implementing memory lifecycle behavior.
Output: context reference model and tests.
Acceptance:
- Context packages can refer to artifacts, relationships, runs, and external
memory records.
- External memory references are opaque and provenance-tagged.
- No retention, decay, compaction, activation planning, or preference-memory
behavior is added to this repo.
## P4.8 - Add import/export and smoke verification
```task
id: KONT-WP-0004-T008
status: todo
priority: medium
```
Provide a deterministic import/export path for repository snapshots so early
users and agents can inspect or migrate local state.
Output: programmatic snapshot helpers and tests.
Acceptance:
- A repository snapshot can be exported and imported into a fresh backend.
- Imported data preserves ids, digests, metadata, relationships, revisions, and
run links.
- Snapshot format does not become a replacement for the service API.
## Definition Of Done
- `python3 -m pytest` passes.
- Existing in-memory behavior remains compatible unless explicitly revised.
- SQLite-backed tests prove durable behavior across repository
re-instantiation.
- Persistence docs explain what is durable now and what remains deferred.
- `docs/phase-memory-boundary.md` remains the boundary for memory-specific
behavior.