@simbimbo/memory-ocmemog 0.1.7 → 0.1.8
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +9 -0
- package/README.md +3 -0
- package/docs/architecture/local-runtime-2026-03-19.md +33 -0
- package/ocmemog/sidecar/app.py +1 -1
- package/package.json +1 -1
package/CHANGELOG.md
CHANGED
|
@@ -1,5 +1,14 @@
|
|
|
1
1
|
# Changelog
|
|
2
2
|
|
|
3
|
+
## 0.1.8 — 2026-03-19
|
|
4
|
+
|
|
5
|
+
Documentation and release follow-through after the llama.cpp migration and repo grooming pass.
|
|
6
|
+
|
|
7
|
+
### Highlights
|
|
8
|
+
- documented the stable local runtime architecture (gateway/sidecar/text/embed split)
|
|
9
|
+
- published the repo in a llama.cpp-first state with fixed ports and cleaned installers/scripts
|
|
10
|
+
- kept compatibility hooks only where still useful instead of leaving Ollama as the implied primary path
|
|
11
|
+
|
|
3
12
|
## 0.1.7 — 2026-03-19
|
|
4
13
|
|
|
5
14
|
llama.cpp-first cleanup after the 0.1.6 runtime cutover.
|
package/README.md
CHANGED
|
@@ -14,6 +14,9 @@ Architecture at a glance:
|
|
|
14
14
|
- **FastAPI sidecar (`ocmemog/sidecar/`)** exposes memory and continuity APIs
|
|
15
15
|
- **SQLite-backed runtime (`brain/runtime/memory/`)** powers storage, hydration, checkpoints, salience ranking, and pondering
|
|
16
16
|
|
|
17
|
+
Current local runtime architecture note:
|
|
18
|
+
- `docs/architecture/local-runtime-2026-03-19.md`
|
|
19
|
+
|
|
17
20
|
## Repo layout
|
|
18
21
|
|
|
19
22
|
- `openclaw.plugin.json`, `index.ts`, `package.json`: OpenClaw plugin package and manifest.
|
|
@@ -0,0 +1,33 @@
|
|
|
1
|
+
# Local Runtime Architecture — 2026-03-19
|
|
2
|
+
|
|
3
|
+
This repo is now documented and operated with a **llama.cpp-first** local runtime architecture.
|
|
4
|
+
|
|
5
|
+
## Stable loopback-only service split
|
|
6
|
+
|
|
7
|
+
- OpenClaw gateway/dashboard: `127.0.0.1:17890`
|
|
8
|
+
- ocmemog sidecar/dashboard: `127.0.0.1:17891`
|
|
9
|
+
- llama.cpp text inference: `127.0.0.1:18080`
|
|
10
|
+
- llama.cpp embeddings: `127.0.0.1:18081`
|
|
11
|
+
|
|
12
|
+
## Active local models
|
|
13
|
+
|
|
14
|
+
- Text: `Qwen2.5-7B-Instruct-Q4_K_M.gguf`
|
|
15
|
+
- Embeddings: `nomic-embed-text-v1.5.Q4_K_M.gguf`
|
|
16
|
+
|
|
17
|
+
## Configuration direction
|
|
18
|
+
|
|
19
|
+
Primary local envs:
|
|
20
|
+
|
|
21
|
+
- `OCMEMOG_LOCAL_LLM_BASE_URL=http://127.0.0.1:18080/v1`
|
|
22
|
+
- `OCMEMOG_LOCAL_LLM_MODEL=qwen2.5-7b-instruct`
|
|
23
|
+
- `OCMEMOG_LOCAL_EMBED_BASE_URL=http://127.0.0.1:18081/v1`
|
|
24
|
+
- `OCMEMOG_LOCAL_EMBED_MODEL=nomic-embed-text-v1.5`
|
|
25
|
+
|
|
26
|
+
Legacy Ollama knobs may remain in code for compatibility/rollback, but they are **not the primary runtime path**.
|
|
27
|
+
|
|
28
|
+
## Operational notes
|
|
29
|
+
|
|
30
|
+
- The sidecar should remain loopback-only by default.
|
|
31
|
+
- The old plain dashboard lives at `http://127.0.0.1:17891/dashboard`.
|
|
32
|
+
- Memory search and pondering should target the sidecar, not the OpenClaw gateway port.
|
|
33
|
+
- Avoid reusing `17890` for the sidecar; that previously caused a routing collision with the OpenClaw dashboard/gateway.
|
package/ocmemog/sidecar/app.py
CHANGED
|
@@ -19,7 +19,7 @@ from ocmemog.sidecar.transcript_watcher import watch_forever
|
|
|
19
19
|
|
|
20
20
|
DEFAULT_CATEGORIES = ("knowledge", "reflections", "directives", "tasks", "runbooks", "lessons")
|
|
21
21
|
|
|
22
|
-
app = FastAPI(title="ocmemog sidecar", version="0.1.
|
|
22
|
+
app = FastAPI(title="ocmemog sidecar", version="0.1.8")
|
|
23
23
|
|
|
24
24
|
API_TOKEN = os.environ.get("OCMEMOG_API_TOKEN")
|
|
25
25
|
|
package/package.json
CHANGED