@lmzhen/dsh-memory-files 0.3.82 → 0.4.0

package/README.md

@@ -1,28 +1,28 @@
-# @deepseek-ai/dsh-memory-files
+# @lmzhen/dsh-memory-files
 
-Local layered memory provider
+Local layered memory provider: registers the `files` provider into `ctx.memory` over the IO seam.
 
-## Model Experience
+## Model surface
 
-### Indirect model surface
+- **Model-visible:** nothing of its own: `tool-memory` owns the tool and the injection.
+- **Prompt prefix / KV cache:** unchanged by this package: family-level rules single-sourced in `packages/README.md` §"Model-visible prompt prefix and the KV cache".
+- **Mount it?** yes — the `memory-files` provider row, in `evolution-host`/`evolution-all`/one-click `evolution-preset`.
 
-#### What the model sees
+## Configuration
 
-`@deepseek-ai/dsh-memory-files` registers no direct prompt or tool schema itself. Model-visible effects are owned by the packages that consume this service.
+- `providerName` (default `files`): the registered name.
+- `memoryCharLimit` / `userCharLimit` (`2200` / `1375`) the budget the STORE enforces per target.
+- `root`: directory override; `''` means `$DSH_HOME/memories`.
+- `addDatePrefix` / `maxConsolidationFailures` (`false` / `3`) date prefixes on entries; consolidation failures one turn tolerates.
+- `threatExemptLabels` (default `[]`): benign threat labels for the MEMORY store (per-site rule in `evolution-threat`).
 
-#### Token effect
+## Known limitations
 
-Zero direct token effect from this package; consumers add any model-visible tokens.
-
-#### KV Cache effect
-
-Independent of request-prefix construction. This package does not alter the assembled prompt or tool list.
-
-## Known Limitations and Deferred Work
-
-- **The memory budget is configured in two places.** `memoryCharLimit`/`userCharLimit` (this package) is what the STORE enforces; `evolution-policy`'s `memoryChars`/`userChars` is what the review pipeline PLANS against. An UNSET limit here follows the mounted policy (row order permitting — the policy must already be mounted when this row assembles), an explicit value wins, and an explicit value that contradicts the policy warns once at load. `/evolution doctor` re-checks the pair at run time and reports a `memory budget:` finding, which also catches the row-order case. "Unset" is read as "equal to the schema default": the loader fills defaults into the row config, so a row that pins the default value explicitly is indistinguishable from one that omits it.
+- **The memory budget is configured in two places.** `memoryCharLimit`/`userCharLimit` (this package) is what the STORE enforces; `evolution-policy`'s `memoryChars`/`userChars` is what the review pipeline PLANS against. An unset limit follows the mounted policy, an explicit value wins, a contradiction warns once at load; `/evolution doctor` re-checks the pair at run time.
+- Nothing selects this provider implicitly: with `memory.provider` empty the FIRST registered provider serves every read/write: set the pin or row order decides.
 
+**Runtime invariant:** No companion is published. The platform auto-assembles nothing and the family mounts no `<pkg>/invariant` cordis row, so a companion here would never execute (v37 S2.1 / I-3).
 
-- `providerName` (default `files`) is the name this provider registers under. Nothing selects it implicitly: `memory.provider` pins the registry to one name (V27 G6.3), and with the pin empty the FIRST registered provider serves every read/write — so with two providers mounted, set the pin or the choice is row order.
+## Notes and history
 
-**Runtime invariant:** No companion is published. The platform auto-assembles nothing and the family mounts no `<pkg>/invariant` cordis row, so a companion here would never execute (v37 S2.1 / I-3).
+- `memory.provider` pins the registry to one name (V27 G6.3), and with the pin empty the FIRST registered provider serves every read/write: so with two providers mounted, set the pin or the choice is row order.

package/lib/index.js

@@ -1,4 +1,5 @@
 import { createHash } from "node:crypto";
+import { resolve } from "node:path";
 import z from "@deepseek-ai/schemastery";
 import { DEFAULT_CONSOLIDATION_FAILURES, DEFAULT_MEMORY_CHAR_LIMIT, DEFAULT_USER_CHAR_LIMIT, MemoryStore, clampedNumber, evolutionIoAdapter, makeSerialQueue } from "@lmzhen/dsh-evolution-core";
 //#region lib/types/index.js
@@ -45,7 +46,8 @@ function apply(ctx, rawConfig = {}) {
 	if (explicitLimit("userCharLimit") && budget.user !== void 0 && config.userCharLimit !== budget.user) ctx.logger.warn(`memory-files: userCharLimit=${config.userCharLimit} contradicts evolution-policy userChars=${budget.user} — same divergence as memoryCharLimit`);
 	const io = evolutionIoAdapter(() => ctx.evolutionIo.provider());
 	const serializedWrite = makeSerialQueue();
-	const resolvedRoot = (config.root || "").trim();
+	const trimmedRoot = (config.root || "").trim();
+	const resolvedRoot = trimmedRoot === "" ? "" : resolve(trimmedRoot);
 	const store = new MemoryStore({
 		memoryCharLimit: config.memoryCharLimit,
 		userCharLimit: config.userCharLimit,

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@lmzhen/dsh-memory-files",
   "description": "Local layered memory provider (community build)",
-  "version": "0.3.82",
+  "version": "0.4.0",
   "publishConfig": {
     "access": "public"
   },
@@ -27,16 +27,16 @@
   "license": "MIT",
   "dependencies": {
     "@deepseek-ai/schemastery": "^3.18.1",
-    "@lmzhen/dsh-evolution-core": "^0.3.82"
+    "@lmzhen/dsh-evolution-core": "^0.4.0"
   },
   "peerDependencies": {
     "@deepseek-ai/cordis": "^4.0.1",
-    "@lmzhen/dsh-evolution-io": "^0.3.82",
-    "@lmzhen/dsh-memory": "^0.3.82"
+    "@lmzhen/dsh-evolution-io": "^0.4.0",
+    "@lmzhen/dsh-memory": "^0.4.0"
   },
   "devDependencies": {
-    "@lmzhen/dsh-evolution-core": "^0.3.82",
-    "@lmzhen/dsh-evolution-io": "^0.3.82",
-    "@lmzhen/dsh-memory": "^0.3.82"
+    "@lmzhen/dsh-evolution-core": "^0.4.0",
+    "@lmzhen/dsh-evolution-io": "^0.4.0",
+    "@lmzhen/dsh-memory": "^0.4.0"
   }
 }