@leemour/cli-messaging 0.145.0 → 0.146.0

package/dist/search/stem.d.ts

@@ -0,0 +1,35 @@
+/** The version of the vendored files in `./snowball/`; `bin/snowball-update` regenerates them. */
+export declare const SNOWBALL_VERSION = "3.1.1";
+export declare const CYRILLIC_STEMMERS: readonly ["russian", "none"];
+export declare const LATIN_STEMMERS: readonly ["spanish", "english", "none"];
+export type CyrillicStemmer = (typeof CYRILLIC_STEMMERS)[number];
+export type LatinStemmer = (typeof LATIN_STEMMERS)[number];
+export type Stemmers = {
+    cyrillic: CyrillicStemmer;
+    latin: LatinStemmer;
+};
+export declare const DEFAULT_STEMMERS: Stemmers;
+/** Clear-all rather than LRU: the 1M benchmark corpus holds 523k distinct words, and a Map stays cheap below this. */
+export declare const STEM_CACHE_LIMIT = 200000;
+/** Recorded beside an index built with these choices; a different value means the index must be rebuilt. */
+export declare const analyzerIdentity: (stemmers: Stemmers) => string;
+/** Checks choices that come from outside the type system — the store, a config value, an MCP argument. */
+export declare const parseStemmers: (value: unknown) => Stemmers;
+export type Stemmer = {
+    readonly stemmers: Stemmers;
+    readonly identity: string;
+    /** The indexed form of one token: stemmed by the script of its letters, then folded with `normalize()`. */
+    stemToken(token: string): string;
+    /** `stemToken` over every word of a text, split as `unicode61` splits it; tokens that fold to nothing are dropped. */
+    stemTokens(text: string): string[];
+};
+/**
+ * One stemmer per fill run or write drain: it owns its cache, so a cache never outlives the choices it was
+ * built with. A token is stemmed only when all its letters are of one script and it has no digit: mixed scripts
+ * (`Москвa` with a Latin `a`) are usually homoglyph spoofing or code, and `covid19` is a name, not a word form.
+ * Ukrainian and other Cyrillic languages go through the Russian stemmer — there is no language detection.
+ */
+export declare const createStemmer: (stemmers?: Stemmers, { cacheLimit }?: {
+    cacheLimit?: number;
+}) => Stemmer;
+//# sourceMappingURL=stem.d.ts.map

package/dist/search/stem.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"stem.d.ts","sourceRoot":"","sources":["../../src/search/stem.ts"],"names":[],"mappings":"AAMA,kGAAkG;AAClG,eAAO,MAAM,gBAAgB,UAAU,CAAA;AAEvC,eAAO,MAAM,iBAAiB,YAAI,SAAS,EAAE,MAAM,CAAU,CAAA;AAC7D,eAAO,MAAM,cAAc,YAAI,SAAS,EAAE,SAAS,EAAE,MAAM,CAAU,CAAA;AACrE,MAAM,MAAM,eAAe,GAAG,CAAC,OAAO,iBAAiB,CAAC,CAAC,MAAM,CAAC,CAAA;AAChE,MAAM,MAAM,YAAY,GAAG,CAAC,OAAO,cAAc,CAAC,CAAC,MAAM,CAAC,CAAA;AAC1D,MAAM,MAAM,QAAQ,GAAG;IAAE,QAAQ,EAAE,eAAe,CAAC;IAAC,KAAK,EAAE,YAAY,CAAA;CAAE,CAAA;AAEzE,eAAO,MAAM,gBAAgB,EAAE,QAAoD,CAAA;AAEnF,sHAAsH;AACtH,eAAO,MAAM,gBAAgB,SAAU,CAAA;AAEvC,4GAA4G;AAC5G,eAAO,MAAM,gBAAgB,aAAc,QAAQ,KAAG,MACgC,CAAA;AAEtF,0GAA0G;AAC1G,eAAO,MAAM,aAAa,UAAW,OAAO,KAAG,QAe9C,CAAA;AAeD,MAAM,MAAM,OAAO,GAAG;IACpB,QAAQ,CAAC,QAAQ,EAAE,QAAQ,CAAA;IAC3B,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAA;IACzB,2GAA2G;IAC3G,SAAS,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,CAAA;IAChC,sHAAsH;IACtH,UAAU,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,EAAE,CAAA;CACnC,CAAA;AAED;;;;;GAKG;AACH,eAAO,MAAM,aAAa,cACd,QAAQ,mBACiB;IAAE,UAAU,CAAC,EAAE,MAAM,CAAA;CAAE,KACzD,OA6BF,CAAA"}

package/dist/search/stem.js

@@ -0,0 +1,76 @@
+import { CliError } from "@leemour/cli-core";
+import { normalize } from "../store/normalize.js";
+import EnglishStemmer from "./snowball/english-stemmer.js";
+import RussianStemmer from "./snowball/russian-stemmer.js";
+import SpanishStemmer from "./snowball/spanish-stemmer.js";
+/** The version of the vendored files in `./snowball/`; `bin/snowball-update` regenerates them. */
+export const SNOWBALL_VERSION = "3.1.1";
+export const CYRILLIC_STEMMERS = ["russian", "none"];
+export const LATIN_STEMMERS = ["spanish", "english", "none"];
+export const DEFAULT_STEMMERS = { cyrillic: "russian", latin: "spanish" };
+/** Clear-all rather than LRU: the 1M benchmark corpus holds 523k distinct words, and a Map stays cheap below this. */
+export const STEM_CACHE_LIMIT = 200_000;
+/** Recorded beside an index built with these choices; a different value means the index must be rebuilt. */
+export const analyzerIdentity = (stemmers) => `snowball-${SNOWBALL_VERSION} cyrillic=${stemmers.cyrillic} latin=${stemmers.latin}`;
+/** Checks choices that come from outside the type system — the store, a config value, an MCP argument. */
+export const parseStemmers = (value) => {
+    const record = (typeof value === "object" && value !== null ? value : {});
+    const pick = (script, allowed, fallback) => {
+        const choice = record[script] ?? fallback;
+        if (allowed.includes(choice))
+            return choice;
+        throw new CliError("validation_error", `${script} words cannot be stemmed with ${JSON.stringify(choice)} — choose ${allowed.join(", ")}`, { reason: "invalid_stemmer", script, allowed: [...allowed] });
+    };
+    return {
+        cyrillic: pick("cyrillic", CYRILLIC_STEMMERS, DEFAULT_STEMMERS.cyrillic),
+        latin: pick("latin", LATIN_STEMMERS, DEFAULT_STEMMERS.latin),
+    };
+};
+/** What FTS5 `unicode61` keeps inside a token: letters, digits, private use, and combining marks. */
+const TOKEN = /[\p{L}\p{N}\p{Co}\p{M}]+/gu;
+const MARKS = /\p{M}+/gu;
+const CYRILLIC = /^\p{Script=Cyrillic}+$/u;
+const LATIN = /^\p{Script=Latin}+$/u;
+const snowball = {
+    russian: () => new RussianStemmer(),
+    spanish: () => new SpanishStemmer(),
+    english: () => new EnglishStemmer(),
+    none: () => null,
+};
+/**
+ * One stemmer per fill run or write drain: it owns its cache, so a cache never outlives the choices it was
+ * built with. A token is stemmed only when all its letters are of one script and it has no digit: mixed scripts
+ * (`Москвa` with a Latin `a`) are usually homoglyph spoofing or code, and `covid19` is a name, not a word form.
+ * Ukrainian and other Cyrillic languages go through the Russian stemmer — there is no language detection.
+ */
+export const createStemmer = (stemmers = DEFAULT_STEMMERS, { cacheLimit = STEM_CACHE_LIMIT } = {}) => {
+    const valid = parseStemmers(stemmers);
+    const cyrillic = snowball[valid.cyrillic]();
+    const latin = snowball[valid.latin]();
+    const cache = new Map();
+    const compute = (token) => {
+        // Stress marks survive NFC (кварти́ру) and would hide the suffix from Snowball; ё and й compose and stay.
+        const word = token.normalize("NFC").toLowerCase().replace(MARKS, "");
+        const engine = CYRILLIC.test(word) ? cyrillic : LATIN.test(word) ? latin : null;
+        const stem = engine?.stemWord(word) || word;
+        return normalize(stem);
+    };
+    const stemToken = (token) => {
+        const cached = cache.get(token);
+        if (cached !== undefined)
+            return cached;
+        const stem = compute(token);
+        if (cache.size >= cacheLimit)
+            cache.clear();
+        if (cacheLimit > 0)
+            cache.set(token, stem);
+        return stem;
+    };
+    return {
+        stemmers: valid,
+        identity: analyzerIdentity(valid),
+        stemToken,
+        stemTokens: (text) => (text.normalize("NFC").toLowerCase().match(TOKEN) ?? []).map(stemToken).filter(Boolean),
+    };
+};
+//# sourceMappingURL=stem.js.map

package/dist/search/stem.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"stem.js","sourceRoot":"","sources":["../../src/search/stem.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,mBAAmB,CAAA;AAC5C,OAAO,EAAE,SAAS,EAAE,MAAM,uBAAuB,CAAA;AACjD,OAAO,cAAc,MAAM,+BAA+B,CAAA;AAC1D,OAAO,cAAc,MAAM,+BAA+B,CAAA;AAC1D,OAAO,cAAc,MAAM,+BAA+B,CAAA;AAE1D,kGAAkG;AAClG,MAAM,CAAC,MAAM,gBAAgB,GAAG,OAAO,CAAA;AAEvC,MAAM,CAAC,MAAM,iBAAiB,GAAG,CAAC,SAAS,EAAE,MAAM,CAAU,CAAA;AAC7D,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,SAAS,EAAE,SAAS,EAAE,MAAM,CAAU,CAAA;AAKrE,MAAM,CAAC,MAAM,gBAAgB,GAAa,EAAE,QAAQ,EAAE,SAAS,EAAE,KAAK,EAAE,SAAS,EAAE,CAAA;AAEnF,sHAAsH;AACtH,MAAM,CAAC,MAAM,gBAAgB,GAAG,OAAO,CAAA;AAEvC,4GAA4G;AAC5G,MAAM,CAAC,MAAM,gBAAgB,GAAG,CAAC,QAAkB,EAAU,EAAE,CAC7D,YAAY,gBAAgB,aAAa,QAAQ,CAAC,QAAQ,UAAU,QAAQ,CAAC,KAAK,EAAE,CAAA;AAEtF,0GAA0G;AAC1G,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,KAAc,EAAY,EAAE;IACxD,MAAM,MAAM,GAAG,CAAC,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,IAAI,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAA4B,CAAA;IACpG,MAAM,IAAI,GAAG,CAAmB,MAAc,EAAE,OAAqB,EAAE,QAAW,EAAK,EAAE;QACvF,MAAM,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC,IAAI,QAAQ,CAAA;QACzC,IAAI,OAAO,CAAC,QAAQ,CAAC,MAAW,CAAC;YAAE,OAAO,MAAW,CAAA;QACrD,MAAM,IAAI,QAAQ,CAChB,kBAAkB,EAClB,GAAG,MAAM,iCAAiC,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,aAAa,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,EACjG,EAAE,MAAM,EAAE,iBAAiB,EAAE,MAAM,EAAE,OAAO,EAAE,CAAC,GAAG,OAAO,CAAC,EAAE,CAC7D,CAAA;IACH,CAAC,CAAA;IACD,OAAO;QACL,QAAQ,EAAE,IAAI,CAAC,UAAU,EAAE,iBAAiB,EAAE,gBAAgB,CAAC,QAAQ,CAAC;QACxE,KAAK,EAAE,IAAI,CAAC,OAAO,EAAE,cAAc,EAAE,gBAAgB,CAAC,KAAK,CAAC;KAC7D,CAAA;AACH,CAAC,CAAA;AAED,qGAAqG;AACrG,MAAM,KAAK,GAAG,4BAA4B,CAAA;AAC1C,MAAM,KAAK,GAAG,UAAU,CAAA;AACxB,MAAM,QAAQ,GAAG,yBAAyB,CAAA;AAC1C,MAAM,KAAK,GAAG,sBAAsB,CAAA;AAEpC,MAAM,QAAQ,GAAG;IACf,OAAO,EAAE,GAAG,EAAE,CAAC,IAAI,cAAc,EAAE;IACnC,OAAO,EAAE,GAAG,EAAE,CAAC,IAAI,cAAc,EAAE;IACnC,OAAO,EAAE,GAAG,EAAE,CAAC,IAAI,cAAc,EAAE;IACnC,IAAI,EAAE,GAAG,EAAE,CAAC,IAAI;CACjB,CAAA;AAWD;;;;;GAKG;AACH,MAAM,CAAC,MAAM,aAAa,GAAG,CAC3B,QAAQ,GAAa,gBAAgB,EACrC,EAAE,UAAU,GAAG,gBAAgB,EAAE,GAA4B,EAAE,EACtD,EAAE;IACX,MAAM,KAAK,GAAG,aAAa,CAAC,QAAQ,CAAC,CAAA;IACrC,MAAM,QAAQ,GAAG,QAAQ,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAA;IAC3C,MAAM,KAAK,GAAG,QAAQ,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAA;IACrC,MAAM,KAAK,GAAG,IAAI,GAAG,EAAkB,CAAA;IAEvC,MAAM,OAAO,GAAG,CAAC,KAAa,EAAU,EAAE;QACxC,0GAA0G;QAC1G,MAAM,IAAI,GAAG,KAAK,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAA;QACpE,MAAM,MAAM,GAAG,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAAA;QAC/E,MAAM,IAAI,GAAG,MAAM,EAAE,QAAQ,CAAC,IAAI,CAAC,IAAI,IAAI,CAAA;QAC3C,OAAO,SAAS,CAAC,IAAI,CAAC,CAAA;IACxB,CAAC,CAAA;IAED,MAAM,SAAS,GAAG,CAAC,KAAa,EAAU,EAAE;QAC1C,MAAM,MAAM,GAAG,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,CAAA;QAC/B,IAAI,MAAM,KAAK,SAAS;YAAE,OAAO,MAAM,CAAA;QACvC,MAAM,IAAI,GAAG,OAAO,CAAC,KAAK,CAAC,CAAA;QAC3B,IAAI,KAAK,CAAC,IAAI,IAAI,UAAU;YAAE,KAAK,CAAC,KAAK,EAAE,CAAA;QAC3C,IAAI,UAAU,GAAG,CAAC;YAAE,KAAK,CAAC,GAAG,CAAC,KAAK,EAAE,IAAI,CAAC,CAAA;QAC1C,OAAO,IAAI,CAAA;IACb,CAAC,CAAA;IAED,OAAO;QACL,QAAQ,EAAE,KAAK;QACf,QAAQ,EAAE,gBAAgB,CAAC,KAAK,CAAC;QACjC,SAAS;QACT,UAAU,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC;KAC9G,CAAA;AACH,CAAC,CAAA"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@leemour/cli-messaging",
-  "version": "0.145.0",
+  "version": "0.146.0",
   "description": "The messenger-neutral half of a messaging CLI: domain model, message locators, rendering, name resolution, the SQLite seam, the command skeleton and shared read commands, the send guard, run records and the message store",
   "license": "MIT",
   "author": "Viacheslav Ptsarev",
@@ -65,7 +65,7 @@
     "./parity.json": "./parity.json"
   },
   "scripts": {
-    "build": "tsc --build && node scripts/bundle-drizzle.ts",
+    "build": "tsc --build && node scripts/bundle-drizzle.ts && node scripts/copy-snowball.ts",
     "check:dist": "node scripts/check-dist.ts",
     "typecheck": "tsc --build --force --emitDeclarationOnly && tsc -p tsconfig.test.json && tsc -p tsconfig.scripts.json",
     "lint": "biome check .",

package/parity.json

@@ -197,6 +197,9 @@
     "--html": {
       "meaning": "the text is HTML: <b>, <i>, <a href>, <code>"
     },
+    "--http": {
+      "meaning": "serve MCP over HTTP on 127.0.0.1 behind the owner's tunnel, with a one-owner login; every write asks first"
+    },
     "--idle": {
       "value": "<duration>",
       "meaning": "stop after this long with nobody using it — 15m, 1h"
@@ -373,10 +376,18 @@
       "value": "<file>",
       "meaning": "an image file — a profile photo in `account update`, a photo to send in `messages send`"
     },
+    "--port": {
+      "value": "<port>",
+      "meaning": "the local port for --http"
+    },
     "--provider": {
       "value": "<provider>",
       "meaning": "embed through a service with your key instead of on this machine: openai"
     },
+    "--public-url": {
+      "value": "<url>",
+      "meaning": "the tunnel's https address the browser apps use"
+    },
     "--qr-file": {
       "value": "<png>",
       "meaning": "write the QR code to this PNG instead of drawing it, for an agent to pass on"
@@ -407,6 +418,9 @@
     "--retract": {
       "meaning": "take your vote back, where the poll allows it"
     },
+    "--revoke": {
+      "meaning": "forget every login given to a browser app over --http"
+    },
     "--revote": {
       "meaning": "people may change their vote"
     },
@@ -1556,6 +1570,34 @@
         },
         "--allow-send": "all",
         "--confirm-send": "all",
+        "--http": {
+          "in": [],
+          "planned": {
+            "max": "CLI-58",
+            "tg": "tg's next cli-messaging bump"
+          }
+        },
+        "--port": {
+          "in": [],
+          "planned": {
+            "max": "CLI-58",
+            "tg": "tg's next cli-messaging bump"
+          }
+        },
+        "--public-url": {
+          "in": [],
+          "planned": {
+            "max": "CLI-58",
+            "tg": "tg's next cli-messaging bump"
+          }
+        },
+        "--revoke": {
+          "in": [],
+          "planned": {
+            "max": "CLI-58",
+            "tg": "tg's next cli-messaging bump"
+          }
+        },
         "--yes": {
           "in": [],
           "planned": {