atmem 2.3.2__tar.gz → 2.3.4__tar.gz

{atmem-2.3.2/atmem.egg-info → atmem-2.3.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: atmem
-Version: 2.3.2
+Version: 2.3.4
 Summary: Host-neutral Agent Black Box and reversible memory control plane.
 Author-email: Javad Taghia <taghia.javad@gmail.com>
 Maintainer-email: Aetna / AtMem <aetna000@gmail.com>
@@ -25,12 +25,16 @@ Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: atmem-atbot==0.1.0
+Requires-Dist: atflows==0.1.1
 Requires-Dist: cryptography<49,>=42
+Provides-Extra: atflows
 Provides-Extra: dev
 Requires-Dist: pytest>=8.0.0; extra == "dev"
 Requires-Dist: tomli>=2.0; python_version < "3.11" and extra == "dev"
 Provides-Extra: semantic
 Requires-Dist: sentence-transformers<6.0.0,>=5.0.0; extra == "semantic"
+Provides-Extra: semantic-accelerated
+Requires-Dist: numpy<3,>=1.26; extra == "semantic-accelerated"
 Provides-Extra: mem0
 Requires-Dist: mem0ai<3,>=2.0.20; extra == "mem0"
 Provides-Extra: langgraph-provider
@@ -75,12 +79,12 @@ Dynamic: license-file
 
 # AtMem
 
-[![Version 2.3.2](https://img.shields.io/badge/version-2.3.2-blue)](./docs/releases/v2.3.2.md)
+[![Version 2.3.4](https://img.shields.io/badge/version-2.3.4-blue)](./docs/releases/v2.3.4.md)
 [![CI](https://github.com/aetna000/atmem/actions/workflows/ci.yml/badge.svg)](https://github.com/aetna000/atmem/actions/workflows/ci.yml)
 
 **AtMem is a host-neutral Agent Black Box and reversible memory control plane.**
 
-> **Release status:** AtMem 2.3.2 is the stable encrypted Agent Black Box release.
+> **Release status:** AtMem 2.3.4 is the stable Agent Black Box and governed-memory release. It installs AtFlows 0.1.1 by default, with optional AtMem-owned login and read-only review leads.
 > It records exact host-observed text and supported multimodal boundaries by default,
 > while keeping every claim limited to what the connected agent actually supplied.
 
@@ -97,7 +101,7 @@ authorizes, stores, scopes, injects, corrects, and deletes memory.
 ### 1. Install AtMem and choose memory intelligence
 
 ```bash
-python -m pip install --upgrade atmem==2.3.2
+python -m pip install --upgrade atmem==2.3.4
 atmem atbot setup
 atmem atbot doctor
 atmem init
@@ -130,7 +134,7 @@ package yourself.
 Already using AtMem 2.1 with OpenClaw? Upgrade in place:
 
 ```bash
-python -m pip install --upgrade atmem==2.3.2
+python -m pip install --upgrade atmem==2.3.4
 atmem openclaw upgrade
 atmem control verify
 ```
@@ -153,7 +157,7 @@ selected by `python`, rather than an unrelated `pip` executable on `PATH`.
 #### Pydantic AI — native capability
 
 ```bash
-python -m pip install 'atmem[pydantic-ai]==2.3.2'
+python -m pip install 'atmem[pydantic-ai]==2.3.4'
 atmem control shadow --host generic --memory-db ~/.atmem/memories.db
 ```
 
@@ -178,7 +182,7 @@ agent = Agent("openai:gpt-5-mini", capabilities=[memory])
 #### LangChain/LangGraph — native middleware
 
 ```bash
-python -m pip install 'atmem[langgraph]==2.3.2'
+python -m pip install 'atmem[langgraph]==2.3.4'
 atmem control shadow --host generic --memory-db ~/.atmem/memories.db
 ```
 
@@ -242,7 +246,7 @@ open tasks. AtBot may propose a change, but AtMem revalidates and commits it.
 
 See the [Governed Task State guide](docs/governed-task-state.md) for lifecycle,
 correction, provenance, expiry, benchmark, and automation examples, and read
-the [2.3.2 release notes](docs/releases/v2.3.2.md) before upgrading.
+the [2.3.4 release notes](docs/releases/v2.3.4.md) before upgrading.
 
 ### Prove memory quality locally
 
@@ -250,12 +254,17 @@ the [2.3.2 release notes](docs/releases/v2.3.2.md) before upgrading.
 atmem benchmark run --output benchmark.json
 ```
 
-This offline release gate runs 16 isolated extraction, contradiction, recall,
+This offline release gate runs 24 isolated extraction, contradiction, recall,
 withholding, injection, privacy, poisoning and fallback cases. Safety must be
 perfect; other quality metrics cannot fall below checked-in baselines. Optional
 local/hosted profiles, LongMemEval import and fair Mem0 OSS comparison are also
 available without adding Mem0 or model SDKs to the base install. See the
 [memory benchmark guide](docs/benchmarks.md) for commands and honest limits.
+The retrieval and archive improvements from [2.3.3](docs/releases/v2.3.3.md)
+remain in [2.3.4](docs/releases/v2.3.4.md), which adds the optional AtFlows
+observability and shared-login scope. Developer opt-in
+fusion, scoped graph nomination and matrix reuse are available for evaluation;
+broader performance targets remain research, not release claims.
 
 ### 3. Review, then activate
 
@@ -340,7 +349,7 @@ and restores it exactly.
 ## Installation details
 
 ```bash
-python -m pip install atmem==2.3.2
+python -m pip install atmem==2.3.4
 atmem --version
 ```
 
@@ -350,7 +359,7 @@ embedding model, while the semantic extra adds local sentence-transformer
 choices:
 
 ```bash
-python -m pip install 'atmem[semantic]==2.3.2'
+python -m pip install 'atmem[semantic]==2.3.4'
 ```
 
 For repository development, install both workspace packages:
@@ -538,7 +547,7 @@ atmem control restore
 Existing 2.1 installations upgrade without starting a new migration:
 
 ```bash
-python -m pip install --upgrade atmem==2.3.2
+python -m pip install --upgrade atmem==2.3.4
 atmem openclaw upgrade
 atmem control verify
 ```
@@ -709,6 +718,11 @@ system-of-record verification.
 
 ## Documentation map
 
+Official website documentation is maintained through the public
+[documentation inventory](docs/website/manifest.json). See
+[how to contribute and update release docs](docs/website-contributing.md).
+Website publication follows a separate owner-reviewed, main-only deployment.
+
 - [Current implementation status](docs/current-status.md)
 - [Generic runtime adapter](docs/generic-adapter.md)
 - [Delegated context-provider contract](docs/contracts/delegated-context-provider-v1.md)
@@ -766,8 +780,8 @@ npm test
 npm run smoke
 ```
 
-Current repository metadata is version **2.3.2**, with the matched OpenClaw
-bridge **2.3.2**. Release validation requires exact Python/bridge alignment;
+Current source metadata is the **2.3.4 stable target**, with matched OpenClaw
+bridge source **2.3.4**. Release validation requires exact Python/bridge alignment;
 AtBot retains its independent stable compatible version **0.1.0**.
 
 ## License

{atmem-2.3.2 → atmem-2.3.4}/README.md

@@ -1,11 +1,11 @@
 # AtMem
 
-[![Version 2.3.2](https://img.shields.io/badge/version-2.3.2-blue)](./docs/releases/v2.3.2.md)
+[![Version 2.3.4](https://img.shields.io/badge/version-2.3.4-blue)](./docs/releases/v2.3.4.md)
 [![CI](https://github.com/aetna000/atmem/actions/workflows/ci.yml/badge.svg)](https://github.com/aetna000/atmem/actions/workflows/ci.yml)
 
 **AtMem is a host-neutral Agent Black Box and reversible memory control plane.**
 
-> **Release status:** AtMem 2.3.2 is the stable encrypted Agent Black Box release.
+> **Release status:** AtMem 2.3.4 is the stable Agent Black Box and governed-memory release. It installs AtFlows 0.1.1 by default, with optional AtMem-owned login and read-only review leads.
 > It records exact host-observed text and supported multimodal boundaries by default,
 > while keeping every claim limited to what the connected agent actually supplied.
 
@@ -22,7 +22,7 @@ authorizes, stores, scopes, injects, corrects, and deletes memory.
 ### 1. Install AtMem and choose memory intelligence
 
 ```bash
-python -m pip install --upgrade atmem==2.3.2
+python -m pip install --upgrade atmem==2.3.4
 atmem atbot setup
 atmem atbot doctor
 atmem init
@@ -55,7 +55,7 @@ package yourself.
 Already using AtMem 2.1 with OpenClaw? Upgrade in place:
 
 ```bash
-python -m pip install --upgrade atmem==2.3.2
+python -m pip install --upgrade atmem==2.3.4
 atmem openclaw upgrade
 atmem control verify
 ```
@@ -78,7 +78,7 @@ selected by `python`, rather than an unrelated `pip` executable on `PATH`.
 #### Pydantic AI — native capability
 
 ```bash
-python -m pip install 'atmem[pydantic-ai]==2.3.2'
+python -m pip install 'atmem[pydantic-ai]==2.3.4'
 atmem control shadow --host generic --memory-db ~/.atmem/memories.db
 ```
 
@@ -103,7 +103,7 @@ agent = Agent("openai:gpt-5-mini", capabilities=[memory])
 #### LangChain/LangGraph — native middleware
 
 ```bash
-python -m pip install 'atmem[langgraph]==2.3.2'
+python -m pip install 'atmem[langgraph]==2.3.4'
 atmem control shadow --host generic --memory-db ~/.atmem/memories.db
 ```
 
@@ -167,7 +167,7 @@ open tasks. AtBot may propose a change, but AtMem revalidates and commits it.
 
 See the [Governed Task State guide](docs/governed-task-state.md) for lifecycle,
 correction, provenance, expiry, benchmark, and automation examples, and read
-the [2.3.2 release notes](docs/releases/v2.3.2.md) before upgrading.
+the [2.3.4 release notes](docs/releases/v2.3.4.md) before upgrading.
 
 ### Prove memory quality locally
 
@@ -175,12 +175,17 @@ the [2.3.2 release notes](docs/releases/v2.3.2.md) before upgrading.
 atmem benchmark run --output benchmark.json
 ```
 
-This offline release gate runs 16 isolated extraction, contradiction, recall,
+This offline release gate runs 24 isolated extraction, contradiction, recall,
 withholding, injection, privacy, poisoning and fallback cases. Safety must be
 perfect; other quality metrics cannot fall below checked-in baselines. Optional
 local/hosted profiles, LongMemEval import and fair Mem0 OSS comparison are also
 available without adding Mem0 or model SDKs to the base install. See the
 [memory benchmark guide](docs/benchmarks.md) for commands and honest limits.
+The retrieval and archive improvements from [2.3.3](docs/releases/v2.3.3.md)
+remain in [2.3.4](docs/releases/v2.3.4.md), which adds the optional AtFlows
+observability and shared-login scope. Developer opt-in
+fusion, scoped graph nomination and matrix reuse are available for evaluation;
+broader performance targets remain research, not release claims.
 
 ### 3. Review, then activate
 
@@ -265,7 +270,7 @@ and restores it exactly.
 ## Installation details
 
 ```bash
-python -m pip install atmem==2.3.2
+python -m pip install atmem==2.3.4
 atmem --version
 ```
 
@@ -275,7 +280,7 @@ embedding model, while the semantic extra adds local sentence-transformer
 choices:
 
 ```bash
-python -m pip install 'atmem[semantic]==2.3.2'
+python -m pip install 'atmem[semantic]==2.3.4'
 ```
 
 For repository development, install both workspace packages:
@@ -463,7 +468,7 @@ atmem control restore
 Existing 2.1 installations upgrade without starting a new migration:
 
 ```bash
-python -m pip install --upgrade atmem==2.3.2
+python -m pip install --upgrade atmem==2.3.4
 atmem openclaw upgrade
 atmem control verify
 ```
@@ -634,6 +639,11 @@ system-of-record verification.
 
 ## Documentation map
 
+Official website documentation is maintained through the public
+[documentation inventory](docs/website/manifest.json). See
+[how to contribute and update release docs](docs/website-contributing.md).
+Website publication follows a separate owner-reviewed, main-only deployment.
+
 - [Current implementation status](docs/current-status.md)
 - [Generic runtime adapter](docs/generic-adapter.md)
 - [Delegated context-provider contract](docs/contracts/delegated-context-provider-v1.md)
@@ -691,8 +701,8 @@ npm test
 npm run smoke
 ```
 
-Current repository metadata is version **2.3.2**, with the matched OpenClaw
-bridge **2.3.2**. Release validation requires exact Python/bridge alignment;
+Current source metadata is the **2.3.4 stable target**, with matched OpenClaw
+bridge source **2.3.4**. Release validation requires exact Python/bridge alignment;
 AtBot retains its independent stable compatible version **0.1.0**.
 
 ## License

atmem-2.3.4/atmem/benchmark/data/mem0-head-to-head-2.3.3b1-v1.json

@@ -0,0 +1,48 @@
+{
+  "format": "atmem-mem0-head-to-head-manifest-v1",
+  "frozen_at": "2026-09-16",
+  "dataset": {
+    "name": "LongMemEval-S cleaned",
+    "upstream_commit": "98d7416c24c778c2fee6e6f3006e7a073259d48f",
+    "source_sha256": "d6f21ea9d60a0d56f34a05b609c79c88a451d2ae03597821ea3d5a9678c3a442"
+  },
+  "selection": {
+    "rule": "sha256-question-id-stratified-v1",
+    "calibration_case_ids": [
+      "07741c45", "b6019101", "720133ac", "gpt4_7fce9456",
+      "c4f10528", "1d4da289", "54026fce", "06878be2",
+      "58bf7951", "c8c3f81d", "gpt4_d31cdae3", "gpt4_4edbafa2"
+    ],
+    "heldout_case_ids": [
+      "6071bd76", "a2f3aa27", "09ba9854", "00ca467f",
+      "70b3e69b", "ceb54acb", "195a1a1b", "6b7dfb22",
+      "ad7109d1", "36580ce8", "bcbe585f", "gpt4_6ed717ea"
+    ],
+    "heldout_selection_sha256": "sha256:4bd9ee57ced759f74e09283da571dd0b1283c1560aa26afabde50a2dd8e050a5"
+  },
+  "raw_retrieval": {
+    "ingestion": "raw-infer-disabled",
+    "chunk_chars": 1600,
+    "embedding_provider": "ollama",
+    "embedding_model": "nomic-embed-text:latest",
+    "embedding_model_id": "0a109f422b47",
+    "retrieved_chunk_limit": 100,
+    "session_cutoff": 5,
+    "quality_metrics": ["session_recall_any_at_5", "session_recall_all_at_5", "session_mrr_at_5"],
+    "latency_boundary": "query embedding through returned search candidates; excludes admission/indexing"
+  },
+  "native_preparation": {
+    "quality_metrics": ["useful_context_precision", "answerable_recall", "no_answer_correctness", "incorrect_injection_rate"],
+    "latency_boundary": "native control_prepare entry through authorized prepared context return",
+    "final_delivery_is_separately_confirmed": true
+  },
+  "competition": {
+    "mem0_oss_version": "2.0.20",
+    "mem0_nlp_and_bm25_status": "explicitly report base and hybrid-enabled profiles separately",
+    "speed_ratio_target": 2.0,
+    "speed_ratio_stretch": 10.0,
+    "speed_metric": "matched warm p95 on the same declared boundary",
+    "quality_condition": "no lower held-out answer/source quality and zero authority violations",
+    "claim_policy": "no cross-profile ratio or beta publication if a required comparable 2x gate fails"
+  }
+}

{atmem-2.3.2 → atmem-2.3.4}/atmem/benchmark/external.py

@@ -25,6 +25,8 @@ _METRIC_DIRECTIONS = {
     "fallback_completion": "higher",
     "latency_p50_ms": "lower",
     "latency_p95_ms": "lower",
+    "warm_latency_p50_ms": "lower",
+    "warm_latency_p95_ms": "lower",
     "token_usage": "lower",
     "model_cost_usd": "lower",
 }