atmem 2.2.6b2__tar.gz → 2.2.6b3__tar.gz

{atmem-2.2.6b2/atmem.egg-info → atmem-2.2.6b3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: atmem
-Version: 2.2.6b2
+Version: 2.2.6b3
 Summary: Host-neutral Agent Black Box and reversible memory control plane.
 Author-email: aetna000 <aetna000@gmail.com>
 Maintainer-email: aetna000 <aetna000@gmail.com>
@@ -31,6 +31,13 @@ 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: mem0
+Requires-Dist: mem0ai<2,>=1; extra == "mem0"
+Provides-Extra: langgraph-provider
+Requires-Dist: langgraph<2,>=1.1.5; extra == "langgraph-provider"
+Provides-Extra: pydantic-provider
+Requires-Dist: pydantic-ai-slim<3,>=2.0; extra == "pydantic-provider"
+Requires-Dist: pydantic<3,>=2.0; extra == "pydantic-provider"
 Provides-Extra: atbot
 Provides-Extra: pydantic-ai
 Requires-Dist: pydantic-ai-slim<3,>=2.0; extra == "pydantic-ai"
@@ -45,7 +52,7 @@ Dynamic: license-file
 
 # AtMem
 
-[![Version 2.2.6b2](https://img.shields.io/badge/version-2.2.6b2-blue)](./docs/releases/v2.2.6b2.md)
+[![Version 2.2.6b3](https://img.shields.io/badge/version-2.2.6b3-blue)](./docs/releases/v2.2.6b3.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.**
@@ -60,7 +67,7 @@ authorizes, stores, scopes, injects, corrects, and deletes memory.
 ### 1. Install AtMem and choose memory intelligence
 
 ```bash
-python -m pip install --pre --upgrade atmem==2.2.6b2
+python -m pip install --pre --upgrade atmem==2.2.6b3
 atmem atbot setup
 atmem atbot doctor
 atmem dashboard
@@ -88,7 +95,7 @@ connection. Do not install the npm package yourself.
 Already using AtMem 2.1 with OpenClaw? Upgrade in place:
 
 ```bash
-python -m pip install --pre --upgrade atmem==2.2.6b2
+python -m pip install --pre --upgrade atmem==2.2.6b3
 atmem openclaw upgrade
 atmem control verify
 ```
@@ -111,7 +118,7 @@ selected by `python`, rather than an unrelated `pip` executable on `PATH`.
 #### Pydantic AI — native capability
 
 ```bash
-python -m pip install --pre 'atmem[pydantic-ai]==2.2.6b2'
+python -m pip install --pre 'atmem[pydantic-ai]==2.2.6b3'
 atmem control shadow --host generic --memory-db ~/.atmem/memories.db
 ```
 
@@ -135,7 +142,7 @@ agent = Agent("openai:gpt-5-mini", capabilities=[memory])
 #### LangChain/LangGraph — native middleware
 
 ```bash
-python -m pip install --pre 'atmem[langgraph]==2.2.6b2'
+python -m pip install --pre 'atmem[langgraph]==2.2.6b3'
 atmem control shadow --host generic --memory-db ~/.atmem/memories.db
 ```
 
@@ -192,7 +199,7 @@ If AtBot or its selected model is unavailable, AtMem continues with safe local
 capture and hybrid ranking. Memory authority and agent operation do not depend
 on a hosted model.
 
-> **Release status:** this repository describes **AtMem 2.2.6b2**. AtBot is a
+> **Release status:** this repository describes **AtMem 2.2.6b3**. AtBot is a
 > separately packaged, headless component installed and managed by AtMem; it is
 > not an independent agent or a second memory authority.
 
@@ -217,6 +224,27 @@ run native retrieval after an accepted delegated decision. Provider failure
 withholds context unless the operator explicitly registered native fallback.
 See the [delegated context-provider guide](docs/delegated-context-provider.md).
 
+AtMem also ships provider-side adapters, each as an independent optional extra:
+
+```bash
+# Mem0 chooses context; AtMem verifies and records delivery
+python -m pip install 'atmem[mem0]'
+atmem provider init mem0-local --kind mem0 --mode oss --port 8788
+
+# Or use an application-owned LangGraph / Pydantic AI factory
+python -m pip install 'atmem[langgraph-provider]'
+atmem provider init graph-context --kind langgraph \
+  --factory myapp.memory_graph:build_graph --port 8789
+
+python -m pip install 'atmem[pydantic-provider]'
+atmem provider init ai-context --kind pydantic-ai \
+  --factory myapp.memory_agent:build_agent --port 8790
+```
+
+`atmem provider init` creates private signing material and prints the exact
+registration command; it does not enable delegated authority. See the
+[context-provider adapter quick starts](docs/context-provider-adapters.md).
+
 It gives agent runtimes one governed memory source and one tamper-evident record
 of what the host observed: memory considered and injected, model boundaries,
 tool requests and completions, turn termination, and linked external outcome
@@ -232,7 +260,7 @@ and restores it exactly.
 ## Installation details
 
 ```bash
-python -m pip install --pre atmem==2.2.6b2
+python -m pip install --pre atmem==2.2.6b3
 atmem --version
 ```
 
@@ -240,7 +268,7 @@ AtMem requires Python 3.10 or newer. It always creates a dependency-free local
 vector sidecar; the semantic extra adds an optional local embedding upgrade:
 
 ```bash
-python -m pip install --pre 'atmem[semantic]==2.2.6b2'
+python -m pip install --pre 'atmem[semantic]==2.2.6b3'
 ```
 
 For repository development, install both workspace packages:
@@ -420,7 +448,7 @@ atmem control restore
 Existing 2.1 installations upgrade without starting a new migration:
 
 ```bash
-python -m pip install --pre --upgrade atmem==2.2.6b2
+python -m pip install --pre --upgrade atmem==2.2.6b3
 atmem openclaw upgrade
 atmem control verify
 ```
@@ -572,9 +600,9 @@ npm test
 npm run smoke
 ```
 
-Current repository metadata is version **2.2.6b2**. Python and npm release
-versions are intentionally kept equal because the OpenClaw installer pins the
-matching bridge.
+Current repository metadata is version **2.2.6b3**. The installer pins the
+compatible OpenClaw bridge independently; provider-only Python releases do not
+force an unnecessary npm publication.
 
 ## License
 

{atmem-2.2.6b2 → atmem-2.2.6b3}/README.md

@@ -1,6 +1,6 @@
 # AtMem
 
-[![Version 2.2.6b2](https://img.shields.io/badge/version-2.2.6b2-blue)](./docs/releases/v2.2.6b2.md)
+[![Version 2.2.6b3](https://img.shields.io/badge/version-2.2.6b3-blue)](./docs/releases/v2.2.6b3.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.**
@@ -15,7 +15,7 @@ authorizes, stores, scopes, injects, corrects, and deletes memory.
 ### 1. Install AtMem and choose memory intelligence
 
 ```bash
-python -m pip install --pre --upgrade atmem==2.2.6b2
+python -m pip install --pre --upgrade atmem==2.2.6b3
 atmem atbot setup
 atmem atbot doctor
 atmem dashboard
@@ -43,7 +43,7 @@ connection. Do not install the npm package yourself.
 Already using AtMem 2.1 with OpenClaw? Upgrade in place:
 
 ```bash
-python -m pip install --pre --upgrade atmem==2.2.6b2
+python -m pip install --pre --upgrade atmem==2.2.6b3
 atmem openclaw upgrade
 atmem control verify
 ```
@@ -66,7 +66,7 @@ selected by `python`, rather than an unrelated `pip` executable on `PATH`.
 #### Pydantic AI — native capability
 
 ```bash
-python -m pip install --pre 'atmem[pydantic-ai]==2.2.6b2'
+python -m pip install --pre 'atmem[pydantic-ai]==2.2.6b3'
 atmem control shadow --host generic --memory-db ~/.atmem/memories.db
 ```
 
@@ -90,7 +90,7 @@ agent = Agent("openai:gpt-5-mini", capabilities=[memory])
 #### LangChain/LangGraph — native middleware
 
 ```bash
-python -m pip install --pre 'atmem[langgraph]==2.2.6b2'
+python -m pip install --pre 'atmem[langgraph]==2.2.6b3'
 atmem control shadow --host generic --memory-db ~/.atmem/memories.db
 ```
 
@@ -147,7 +147,7 @@ If AtBot or its selected model is unavailable, AtMem continues with safe local
 capture and hybrid ranking. Memory authority and agent operation do not depend
 on a hosted model.
 
-> **Release status:** this repository describes **AtMem 2.2.6b2**. AtBot is a
+> **Release status:** this repository describes **AtMem 2.2.6b3**. AtBot is a
 > separately packaged, headless component installed and managed by AtMem; it is
 > not an independent agent or a second memory authority.
 
@@ -172,6 +172,27 @@ run native retrieval after an accepted delegated decision. Provider failure
 withholds context unless the operator explicitly registered native fallback.
 See the [delegated context-provider guide](docs/delegated-context-provider.md).
 
+AtMem also ships provider-side adapters, each as an independent optional extra:
+
+```bash
+# Mem0 chooses context; AtMem verifies and records delivery
+python -m pip install 'atmem[mem0]'
+atmem provider init mem0-local --kind mem0 --mode oss --port 8788
+
+# Or use an application-owned LangGraph / Pydantic AI factory
+python -m pip install 'atmem[langgraph-provider]'
+atmem provider init graph-context --kind langgraph \
+  --factory myapp.memory_graph:build_graph --port 8789
+
+python -m pip install 'atmem[pydantic-provider]'
+atmem provider init ai-context --kind pydantic-ai \
+  --factory myapp.memory_agent:build_agent --port 8790
+```
+
+`atmem provider init` creates private signing material and prints the exact
+registration command; it does not enable delegated authority. See the
+[context-provider adapter quick starts](docs/context-provider-adapters.md).
+
 It gives agent runtimes one governed memory source and one tamper-evident record
 of what the host observed: memory considered and injected, model boundaries,
 tool requests and completions, turn termination, and linked external outcome
@@ -187,7 +208,7 @@ and restores it exactly.
 ## Installation details
 
 ```bash
-python -m pip install --pre atmem==2.2.6b2
+python -m pip install --pre atmem==2.2.6b3
 atmem --version
 ```
 
@@ -195,7 +216,7 @@ AtMem requires Python 3.10 or newer. It always creates a dependency-free local
 vector sidecar; the semantic extra adds an optional local embedding upgrade:
 
 ```bash
-python -m pip install --pre 'atmem[semantic]==2.2.6b2'
+python -m pip install --pre 'atmem[semantic]==2.2.6b3'
 ```
 
 For repository development, install both workspace packages:
@@ -375,7 +396,7 @@ atmem control restore
 Existing 2.1 installations upgrade without starting a new migration:
 
 ```bash
-python -m pip install --pre --upgrade atmem==2.2.6b2
+python -m pip install --pre --upgrade atmem==2.2.6b3
 atmem openclaw upgrade
 atmem control verify
 ```
@@ -527,9 +548,9 @@ npm test
 npm run smoke
 ```
 
-Current repository metadata is version **2.2.6b2**. Python and npm release
-versions are intentionally kept equal because the OpenClaw installer pins the
-matching bridge.
+Current repository metadata is version **2.2.6b3**. The installer pins the
+compatible OpenClaw bridge independently; provider-only Python releases do not
+force an unnecessary npm publication.
 
 ## License
 

{atmem-2.2.6b2 → atmem-2.2.6b3}/atmem/cli.py

@@ -256,6 +256,59 @@ wants AtMem to resume native context preparation when the provider fails.""",
         command_parser = delegated_commands.add_parser(name, help=help_text)
         command_parser.add_argument("--json", action="store_true")
 
+    provider_parser = subparsers.add_parser(
+        "provider",
+        help="Run an optional Mem0, LangGraph, or Pydantic AI context authority",
+        description=(
+            "Create a signed local context-provider service. Initialization and startup "
+            "do not change AtMem authority; registration and enablement remain explicit."
+        ),
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+        epilog="""Examples:
+  python -m pip install 'atmem[mem0]'
+  atmem provider init memory-provider --kind mem0 --mode oss --port 8788
+  atmem provider start memory-provider
+  atmem provider doctor memory-provider
+
+  python -m pip install 'atmem[langgraph-provider]'
+  atmem provider init graph-provider --kind langgraph \
+    --factory myapp.context:build_graph --port 8789
+
+  python -m pip install 'atmem[pydantic-provider]'
+  atmem provider init ai-provider --kind pydantic-ai \
+    --factory myapp.context:build_agent --egress hosted --port 8790
+
+The init result prints the separate `atmem delegated register` command. Run its
+matching `atmem delegated enable` command only after reviewing the exact scopes.""",
+    )
+    provider_commands = provider_parser.add_subparsers(dest="provider_command")
+    provider_init = provider_commands.add_parser("init", help="Create private keys and secret-free configuration")
+    provider_init.add_argument("instance")
+    provider_init.add_argument("--kind", required=True, choices=("mem0", "langgraph", "pydantic-ai"))
+    provider_init.add_argument("--port", type=int, default=8788)
+    provider_init.add_argument("--factory", default=None, help="Operator factory in module:attribute form")
+    provider_init.add_argument("--mode", choices=("oss", "platform"), default=None, help="Mem0 client mode")
+    provider_init.add_argument("--provider-id", default=None)
+    provider_init.add_argument("--provider-version", default="1.0")
+    provider_init.add_argument("--egress", choices=("local", "hosted"), default="local")
+    provider_init.add_argument("--json", action="store_true")
+    for name, help_text in (
+        ("serve", "Run one provider in the foreground"),
+        ("start", "Start one private background provider process"),
+        ("stop", "Stop only the PID owned by this provider instance"),
+        ("doctor", "Check dependencies, files, service health, and next trust step"),
+    ):
+        command_parser = provider_commands.add_parser(name, help=help_text)
+        command_parser.add_argument("instance")
+        command_parser.add_argument("--json", action="store_true")
+    provider_status = provider_commands.add_parser("status", help="Show redacted provider state")
+    provider_status.add_argument("instance", nargs="?")
+    provider_status.add_argument("--json", action="store_true")
+    provider_remove = provider_commands.add_parser("remove", help="Remove a stopped provider; AtMem evidence is retained")
+    provider_remove.add_argument("instance")
+    provider_remove.add_argument("--yes", action="store_true")
+    provider_remove.add_argument("--json", action="store_true")
+
     benchmark_parser = subparsers.add_parser(
         "benchmark",
         help="Run reproducible memory-quality gates and compare external results",
@@ -885,6 +938,13 @@ External evaluation:
         _run_delegated(args)
         return
 
+    if args.command == "provider":
+        if args.provider_command is None:
+            provider_parser.print_help()
+            return
+        _run_provider(args)
+        return
+
     if args.command == "benchmark":
         if args.benchmark_command is None:
             benchmark_parser.print_help()
@@ -2040,6 +2100,67 @@ def _run_delegated(args: argparse.Namespace) -> None:
         print(f"Removed {result['registration_id']}")
     else:
         print(json.dumps(result, indent=2, sort_keys=True, default=str))
+
+
+def _run_provider(args: argparse.Namespace) -> None:
+    from atmem.provider_adapters import lifecycle
+
+    command = args.provider_command
+    try:
+        if command == "init":
+            result = lifecycle.initialize(
+                instance=args.instance, kind=args.kind, port=args.port,
+                factory=args.factory, mode=args.mode, provider_id=args.provider_id,
+                provider_version=args.provider_version, egress=args.egress,
+            )
+        elif command == "serve":
+            from atmem.provider_adapters.server import serve
+
+            _, config = lifecycle.load_config(args.instance)
+            serve(lifecycle.build_runtime(args.instance), config["host"], config["port"])
+            return
+        elif command == "start":
+            result = lifecycle.start(args.instance)
+        elif command == "stop":
+            result = lifecycle.stop(args.instance)
+        elif command == "doctor":
+            result = lifecycle.doctor(args.instance)
+        elif command == "status":
+            if args.instance:
+                result = lifecycle.status(args.instance)
+            else:
+                root = lifecycle.provider_root()
+                result = {
+                    "format": "atmem-provider-status-list-v1",
+                    "providers": [
+                        lifecycle.status(path.name)
+                        for path in sorted(root.iterdir())
+                        if path.is_dir() and not path.is_symlink()
+                    ] if root.is_dir() else [],
+                }
+        elif command == "remove":
+            if not args.yes:
+                raise ValueError("provider removal requires --yes")
+            result = lifecycle.remove(args.instance)
+        else:  # pragma: no cover
+            raise ValueError(f"unknown provider command: {command}")
+    except (ImportError, OSError, RuntimeError, ValueError) as exc:
+        print(f"error: {exc}", file=sys.stderr)
+        raise SystemExit(2) from exc
+    if getattr(args, "json", False):
+        print(json.dumps(result, indent=2, sort_keys=True, default=str))
+        return
+    if command == "init":
+        print(f"Created provider {result['instance']} (authority remains native AtMem).")
+        print("Start it:")
+        print(f"  atmem provider start {result['instance']}")
+        print("Register exact trust scopes (still disabled):")
+        print(f"  {result['registration_command']}")
+        print(f"Then review and enable: atmem delegated enable {result['provider_id']}:{result['instance']}")
+    else:
+        print(json.dumps(result, indent=2, sort_keys=True, default=str))
+
+
 def _interactive_atbot_setup(manager: Any) -> dict[str, Any]:
     from atmem.control.atbot_service import PROVIDER_PROFILES
 

{atmem-2.2.6b2 → atmem-2.2.6b3}/atmem/mcp/server.py

@@ -26,7 +26,7 @@ try:
 
     SERVER_VERSION = _pkg_version("atmem")
 except Exception:  # not installed (e.g. run from a checkout)
-    SERVER_VERSION = "2.2.6b2"
+    SERVER_VERSION = "2.2.6b3"
 
 _SUBJECT_PROPERTY = {
     "subject_id": {

atmem-2.2.6b3/atmem/provider_adapters/__init__.py

@@ -0,0 +1,12 @@
+"""Optional context-provider adapters for AtMem's delegated v1 contract."""
+
+from .models import ContextItem, ProviderProposal, ProviderRequest, ProviderRuntimeIdentity
+from .runtime import ProviderRuntime
+
+__all__ = [
+    "ContextItem",
+    "ProviderProposal",
+    "ProviderRequest",
+    "ProviderRuntime",
+    "ProviderRuntimeIdentity",
+]

atmem-2.2.6b3/atmem/provider_adapters/langgraph.py

@@ -0,0 +1,43 @@
+"""LangGraph-backed delegated context provider."""
+
+from __future__ import annotations
+
+import inspect
+from typing import Any
+
+from .models import ProviderProposal, ProviderRequest
+
+
+class LangGraphContextProvider:
+    def __init__(self, graph: Any):
+        self.graph = graph
+        self.attribution = {"adapter": "langgraph"}
+
+    async def decide(self, request: ProviderRequest) -> ProviderProposal:
+        graph_input = {"delegated_request": {
+            "query": request.query,
+            "query_sha256": request.query_sha256,
+            "binding": request.binding.to_dict(),
+            "max_context_bytes": request.max_context_bytes,
+            "deadline": request.deadline,
+        }}
+        config = {"configurable": {"thread_id": f"{request.binding.session_id}:{request.binding.turn_id}"}}
+        if hasattr(self.graph, "ainvoke"):
+            output = await self.graph.ainvoke(graph_input, config=dict(config))
+        elif hasattr(self.graph, "invoke"):
+            output = self.graph.invoke(graph_input, config=dict(config))
+            if inspect.isawaitable(output):
+                output = await output
+        elif callable(self.graph):
+            output = self.graph(graph_input, dict(config))
+            if inspect.isawaitable(output):
+                output = await output
+        else:
+            raise ValueError("LangGraph provider requires an invokable graph or callable")
+        interrupts = getattr(output, "interrupts", ())
+        if interrupts:
+            raise ValueError("LangGraph interrupted before a context decision")
+        value = getattr(output, "value", output)
+        if not isinstance(value, dict) or set(value) != {"context_decision"}:
+            raise ValueError("LangGraph output must contain only context_decision")
+        return ProviderProposal.from_dict(value["context_decision"])