release: prepare v0.4.0

CHANGELOG.md

@@ -6,6 +6,8 @@ semantic versioning once `0.1.0` ships; pre-alpha releases may break anything.
 
 ## [Unreleased]
 
+## [0.4.0] — 2026-05-05 — Content-addressed provider snapshots
+
 ### Added
 
 - Track 026 adds content-addressed provider snapshots:
@@ -69,7 +71,7 @@ Includes approved/prohibited phrasings for academic and licensing contexts.
   physics) → `elume` (cognitive substrate) → application layers.
 - v0.3.0 leaves `elume.linoss` in place (numerical behavior pinned by 1177
   passing tests). Code-level consolidation between `elume.linoss` and
-  `linoss-dynamics` is planned for v0.4.0.
+  `linoss-dynamics` is planned for a follow-up release.
 - Consumers can now `from linoss_dynamics import energy, delta_energy,
   convergence_window` directly to access diagnostic helpers not exposed by
   `elume.linoss`.

README.md

@@ -207,7 +207,7 @@ elume/
 Downstream projects pin a versioned PyPI release:
 
 ```bash
-pip install elume==0.1.0
+pip install elume==0.4.0
 ```
 
 For co-development against an unreleased branch, an editable install also works:

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "elume"
-version = "0.3.0"
+version = "0.4.0"
 description = "Agentic memory with mental-model architecture. Deterministic, replay-safe substrate for cognitive-architecture agents: LinOSS temporal encoding, Hopfield attractor memory, mental-model and metacognitive primitives, strategy evolution, pluggable MemEvolve benchmark adapter. Contribution: byte-identical replay, immutable lineage, provider abstraction, composable operators."
 readme = "README.md"
 requires-python = ">=3.11"

src/elume/__init__.py

@@ -1,12 +1,10 @@
 """Elume
 
-An open-source agentic memory engine for long-horizon adaptive learning.
+Deterministic, replay-safe substrate for cognitive-architecture agents.
 
-Elume unifies three mechanisms into a single self-evolving memory kernel:
-LinOSS oscillatory state-space models for encoding long-horizon temporal
-trajectories, attractor-basin dynamics for content-addressable memory,
-and MemEvolve-style evolutionary optimization for adapting memory
-retrieval and management strategies over time.
+Elume unifies long-horizon temporal encoding, attractor-basin memory,
+mental-model and metacognitive primitives, and MemEvolve-style evolutionary
+optimization behind clean provider boundaries.
 
 The contribution is integration, not invention. The underlying
 techniques are drawn from:
@@ -28,4 +26,4 @@
 See conductor/product.md for the full product specification.
 """
 
-__version__ = "0.2.0"
+__version__ = "0.4.0"

tests/unit/test_version.py

@@ -1,5 +1,8 @@
 """Version sanity check. Pinned so accidental version drift is caught."""
 
+import tomllib
+from pathlib import Path
+
 import elume
 
 
@@ -8,4 +11,5 @@ def test_version_is_declared() -> None:
 
 
 def test_version_matches_pyproject() -> None:
-    assert elume.__version__ == "0.2.0"
+    pyproject = tomllib.loads(Path("pyproject.toml").read_text())
+    assert elume.__version__ == pyproject["project"]["version"]