@aliou/pi-neuralwatt 0.7.4 → 0.8.0

package/{src/extensions → extensions}/provider/models/public-models.ts

@@ -18,6 +18,7 @@ export const NEURALWATT_MODELS: ProviderModelConfig[] = [
     contextWindow: 1048560,
     maxTokens: 65536,
     thinkingLevelMap: {
+      off: "none",
       minimal: null,
       low: null,
       medium: null,
@@ -64,6 +65,7 @@ export const NEURALWATT_MODELS: ProviderModelConfig[] = [
     contextWindow: 199984,
     maxTokens: 65536,
     thinkingLevelMap: {
+      off: "none",
       minimal: null,
       low: null,
       medium: null,
@@ -95,52 +97,6 @@ export const NEURALWATT_MODELS: ProviderModelConfig[] = [
       maxTokensField: "max_tokens",
     },
   },
-  // Kimi K2.5 - MoonshotAI
-  {
-    id: "moonshotai/Kimi-K2.5",
-    name: "Kimi K2.5",
-    reasoning: true,
-    input: ["text", "image"],
-    cost: {
-      input: 0.52,
-      output: 2.59,
-      cacheRead: 0.13,
-      cacheWrite: 0,
-    },
-    contextWindow: 262128,
-    maxTokens: 65536,
-    thinkingLevelMap: {
-      minimal: null,
-      low: null,
-      medium: "medium",
-      high: null,
-      xhigh: null,
-    },
-    compat: {
-      supportsDeveloperRole: false,
-      maxTokensField: "max_tokens",
-      requiresReasoningContentOnAssistantMessages: true,
-    },
-  },
-  // Kimi K2.5 Fast - MoonshotAI
-  {
-    id: "kimi-k2.5-fast",
-    name: "Kimi K2.5 Fast",
-    reasoning: false,
-    input: ["text", "image"],
-    cost: {
-      input: 0.52,
-      output: 2.59,
-      cacheRead: 0.13,
-      cacheWrite: 0,
-    },
-    contextWindow: 262128,
-    maxTokens: 65536,
-    compat: {
-      supportsDeveloperRole: false,
-      maxTokensField: "max_tokens",
-    },
-  },
   // Kimi K2.6 - MoonshotAI
   {
     id: "kimi-k2.6",
@@ -288,6 +244,136 @@ export const NEURALWATT_MODELS: ProviderModelConfig[] = [
       requiresReasoningContentOnAssistantMessages: true,
     },
   },
+  // GLM-5.2 Short Fast Flex - ZhipuAI (flex variant, reasoning disabled)
+  {
+    id: "glm-5.2-short-fast-flex",
+    name: "GLM-5.2 (short, fast, flex)",
+    reasoning: false,
+    input: ["text"],
+    cost: {
+      input: 1.45,
+      output: 4.5,
+      cacheRead: 0.3625,
+      cacheWrite: 0,
+    },
+    contextWindow: 199984,
+    maxTokens: 65536,
+    compat: {
+      supportsDeveloperRole: false,
+      maxTokensField: "max_tokens",
+    },
+  },
+  // GLM-5.2 Short Flex - ZhipuAI (flex variant)
+  {
+    id: "glm-5.2-short-flex",
+    name: "GLM-5.2 (short, flex)",
+    reasoning: true,
+    input: ["text"],
+    cost: {
+      input: 1.45,
+      output: 4.5,
+      cacheRead: 0.3625,
+      cacheWrite: 0,
+    },
+    contextWindow: 199984,
+    maxTokens: 65536,
+    thinkingLevelMap: {
+      off: "none",
+      minimal: null,
+      low: null,
+      medium: null,
+      high: "high",
+      xhigh: "max",
+    },
+    compat: {
+      supportsDeveloperRole: false,
+      maxTokensField: "max_tokens",
+      requiresReasoningContentOnAssistantMessages: true,
+    },
+  },
+  // Kimi K2.6 Flex - MoonshotAI (flex variant)
+  {
+    id: "kimi-k2.6-flex",
+    name: "Kimi K2.6 (flex)",
+    reasoning: true,
+    input: ["text", "image"],
+    cost: {
+      input: 0.69,
+      output: 3.22,
+      cacheRead: 0.1725,
+      cacheWrite: 0,
+    },
+    contextWindow: 262128,
+    maxTokens: 65536,
+    thinkingLevelMap: {
+      minimal: null,
+      low: null,
+      medium: "medium",
+      high: null,
+      xhigh: null,
+    },
+    compat: {
+      supportsDeveloperRole: false,
+      maxTokensField: "max_tokens",
+      requiresReasoningContentOnAssistantMessages: true,
+    },
+  },
+  // Kimi K2.7 Code Flex - MoonshotAI (flex variant)
+  {
+    id: "kimi-k2.7-code-flex",
+    name: "Kimi K2.7 Code (flex)",
+    reasoning: true,
+    input: ["text", "image"],
+    cost: {
+      input: 0.95,
+      output: 4.0,
+      cacheRead: 0.2375,
+      cacheWrite: 0,
+    },
+    contextWindow: 262128,
+    maxTokens: 65536,
+    thinkingLevelMap: {
+      off: null,
+      minimal: null,
+      low: null,
+      medium: "medium",
+      high: null,
+      xhigh: null,
+    },
+    compat: {
+      supportsDeveloperRole: false,
+      maxTokensField: "max_tokens",
+      requiresReasoningContentOnAssistantMessages: true,
+    },
+  },
+  // GLM-5.2 Flex - ZhipuAI (flex variant)
+  {
+    id: "glm-5.2-flex",
+    name: "GLM-5.2 (flex)",
+    reasoning: true,
+    input: ["text"],
+    cost: {
+      input: 1.45,
+      output: 4.5,
+      cacheRead: 0.3625,
+      cacheWrite: 0,
+    },
+    contextWindow: 1048560,
+    maxTokens: 65536,
+    thinkingLevelMap: {
+      off: "none",
+      minimal: null,
+      low: null,
+      medium: null,
+      high: "high",
+      xhigh: "max",
+    },
+    compat: {
+      supportsDeveloperRole: false,
+      maxTokensField: "max_tokens",
+      requiresReasoningContentOnAssistantMessages: true,
+    },
+  },
   // Qwen3.6 35B Fast - Qwen
   {
     id: "qwen3.6-35b-fast",

package/{src/extensions → extensions}/provider/quota-store.ts

@@ -1,8 +1,8 @@
 import type { AuthStorage } from "@earendil-works/pi-coding-agent";
-import { getNeuralwattApiKey } from "../../lib/env";
-import { fetchQuotas } from "../../lib/neuralwatt-api";
-import type { NeuralwattQuotas } from "../../types/quota-api";
-import { parseQuotaHeaders } from "../../types/quota-events";
+import { parseQuotaHeaders } from "../../src/events";
+import { fetchQuotas } from "../../src/lib/neuralwatt-api";
+import type { NeuralwattQuotas } from "../../src/types/quota-api";
+import { getNeuralwattApiKey } from "../_shared/auth";
 
 export function buildQuotasFromHeaders(
   headers: Record<string, string>,

package/{src/extensions → extensions}/provider/sse-quotas.ts

@@ -1,4 +1,4 @@
-import type { NeuralwattQuotas } from "../../types/quota-api";
+import type { NeuralwattQuotas } from "../../src/types/quota-api";
 
 const JOULES_PER_KWH = 3_600_000;
 

package/{src/extensions → extensions}/quota-warnings/index.ts

@@ -2,29 +2,28 @@ import type {
   ExtensionAPI,
   ExtensionContext,
 } from "@earendil-works/pi-coding-agent";
+import { configLoader } from "../../src/config";
 import {
-  configLoader,
   NEURALWATT_CONFIG_UPDATED_EVENT,
   NEURALWATT_EXTENSIONS_REGISTER_EVENT,
   NEURALWATT_EXTENSIONS_REQUEST_EVENT,
-  type NeuralwattConfigUpdatedPayload,
-} from "../../config";
-import {
   NEURALWATT_QUOTAS_UPDATED_EVENT,
+  type NeuralwattConfigUpdatedPayload,
   type NeuralwattQuotasUpdatedPayload,
-} from "../../types/quota-events";
+} from "../../src/events";
 import { checkQuotas, clearAlertState } from "./notifier";
 
 export default async function (pi: ExtensionAPI) {
   await configLoader.load();
 
-  let enabled = configLoader.getConfig().quotaWarnings;
+  let enabled = configLoader.getConfig().quotaWarnings.enabled;
   let currentProvider: string | undefined;
   let currentContext: ExtensionContext | undefined;
 
   // Listen for config changes at runtime
   pi.events.on(NEURALWATT_CONFIG_UPDATED_EVENT, (data: unknown) => {
-    enabled = (data as NeuralwattConfigUpdatedPayload).config.quotaWarnings;
+    enabled = (data as NeuralwattConfigUpdatedPayload).config.quotaWarnings
+      .enabled;
 
     if (!enabled) {
       clearAlertState();

package/{src/extensions → extensions}/quota-warnings/notifier.ts

@@ -1,6 +1,6 @@
 import type { ExtensionContext } from "@earendil-works/pi-coding-agent";
-import type { NeuralwattQuotas } from "../../types/quota-api";
-import { formatKwh, formatUsd } from "../../utils/quota-format";
+import type { NeuralwattQuotas } from "../../src/types/quota-api";
+import { formatKwh, formatUsd } from "../../src/utils/quota-format";
 
 export type WarningSeverity = "warning" | "critical";
 

package/{src/extensions → extensions}/sub-bar-integration/index.ts

@@ -4,24 +4,22 @@ import type {
   ExtensionContext,
   Theme,
 } from "@earendil-works/pi-coding-agent";
+import { configLoader } from "../../src/config";
 import {
-  configLoader,
   NEURALWATT_CONFIG_UPDATED_EVENT,
   NEURALWATT_EXTENSIONS_REGISTER_EVENT,
   NEURALWATT_EXTENSIONS_REQUEST_EVENT,
-  type NeuralwattConfigUpdatedPayload,
-} from "../../config";
-import type { NeuralwattQuotas } from "../../types/quota-api";
-import {
   NEURALWATT_QUOTAS_REQUEST_EVENT,
   NEURALWATT_QUOTAS_UPDATED_EVENT,
+  type NeuralwattConfigUpdatedPayload,
   type NeuralwattQuotasUpdatedPayload,
-} from "../../types/quota-events";
+} from "../../src/events";
+import type { NeuralwattQuotas } from "../../src/types/quota-api";
 import {
   percentCreditsRemaining,
   percentEnergyRemaining,
-} from "../../utils/quota-bar";
-import { formatKwh, formatUsd } from "../../utils/quota-format";
+} from "../../src/utils/quota-bar";
+import { formatKwh, formatUsd } from "../../src/utils/quota-format";
 import { toUsageSnapshot } from "./snapshot";
 
 function formatStatus(quotas: NeuralwattQuotas, theme: Theme): string {
@@ -57,7 +55,7 @@ function formatStatus(quotas: NeuralwattQuotas, theme: Theme): string {
 export default async function (pi: ExtensionAPI) {
   await configLoader.load();
 
-  let enabled = configLoader.getConfig().subBarIntegration;
+  let enabled = configLoader.getConfig().subBarIntegration.enabled;
   let subCoreReady = false;
   let currentProvider: string | undefined;
   let currentAuthStorage: AuthStorage | undefined;
@@ -65,7 +63,8 @@ export default async function (pi: ExtensionAPI) {
 
   // Listen for config changes at runtime
   pi.events.on(NEURALWATT_CONFIG_UPDATED_EVENT, (data: unknown) => {
-    enabled = (data as NeuralwattConfigUpdatedPayload).config.subBarIntegration;
+    enabled = (data as NeuralwattConfigUpdatedPayload).config.subBarIntegration
+      .enabled;
   });
 
   function isActive(): boolean {

package/{src/extensions → extensions}/sub-bar-integration/snapshot.ts

@@ -1,9 +1,9 @@
-import type { NeuralwattQuotas } from "../../types/quota-api";
+import type { NeuralwattQuotas } from "../../src/types/quota-api";
 import {
   percentCreditsRemaining,
   percentEnergyRemaining,
-} from "../../utils/quota-bar";
-import { formatKwh } from "../../utils/quota-format";
+} from "../../src/utils/quota-bar";
+import { formatKwh } from "../../src/utils/quota-format";
 
 interface RateWindow {
   label: string;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@aliou/pi-neuralwatt",
-  "version": "0.7.4",
+  "version": "0.8.0",
   "license": "MIT",
   "type": "module",
   "private": false,
@@ -15,10 +15,10 @@
   },
   "pi": {
     "extensions": [
-      "./src/extensions/provider/index.ts",
-      "./src/extensions/command-quotas/index.ts",
-      "./src/extensions/quota-warnings/index.ts",
-      "./src/extensions/sub-bar-integration/index.ts"
+      "./extensions/provider/index.ts",
+      "./extensions/command-quotas/index.ts",
+      "./extensions/quota-warnings/index.ts",
+      "./extensions/sub-bar-integration/index.ts"
     ],
     "video": "https://assets.aliou.me/github/aliou/pi-neuralwatt/demo.mp4"
   },
@@ -27,25 +27,28 @@
   },
   "files": [
     "src",
+    "extensions",
     "schema.json",
     "README.md",
-    "!src/**/*.test.ts"
+    "!src/**/*.test.ts",
+    "!extensions/**/*.test.ts"
   ],
   "dependencies": {
-    "@aliou/pi-utils-settings": "^0.17.0",
-    "@aliou/pi-utils-ui": "^0.4.0",
-    "@earendil-works/pi-ai": "$@earendil-works/pi-coding-agent"
+    "@aliou/pi-utils-settings": "^0.18.0",
+    "@aliou/pi-utils-ui": "^0.5.0"
   },
   "peerDependencies": {
+    "@earendil-works/pi-ai": "*",
     "@earendil-works/pi-coding-agent": "*",
     "@earendil-works/pi-tui": "*"
   },
   "devDependencies": {
-    "@aliou/biome-plugins": "^0.8.1",
+    "@aliou/biome-plugins": "^0.11.0",
     "@biomejs/biome": "^2.4.15",
     "@changesets/cli": "^2.27.11",
-    "@earendil-works/pi-coding-agent": "0.77.0",
-    "@earendil-works/pi-tui": "0.77.0",
+    "@earendil-works/pi-ai": "0.80.3",
+    "@earendil-works/pi-coding-agent": "0.80.3",
+    "@earendil-works/pi-tui": "0.80.3",
     "@types/node": "^25.0.10",
     "husky": "^9.1.7",
     "ts-json-schema-generator": "^2.4.0",
@@ -64,8 +67,8 @@
     "format": "biome check --write",
     "test": "vitest run",
     "test:watch": "vitest",
-    "gen:schema": "ts-json-schema-generator --path src/config.ts --type NeuralwattConfig --no-type-check -o schema.json",
-    "check:schema": "ts-json-schema-generator --path src/config.ts --type NeuralwattConfig --no-type-check -o /tmp/schema-check.json && diff -q schema.json /tmp/schema-check.json",
+    "gen:schema": "ts-json-schema-generator --path src/config/types.ts --type NeuralwattConfig --no-type-check -o schema.json",
+    "check:schema": "ts-json-schema-generator --path src/config/types.ts --type NeuralwattConfig --no-type-check -o /tmp/schema-check.json && diff -q schema.json /tmp/schema-check.json",
     "check:lockfile": "pnpm install --frozen-lockfile --ignore-scripts",
     "prepare": "[ -d .git ] && husky || true",
     "changeset": "changeset",
@@ -74,6 +77,9 @@
   },
   "packageManager": "pnpm@10.26.1",
   "peerDependenciesMeta": {
+    "@earendil-works/pi-ai": {
+      "optional": true
+    },
     "@earendil-works/pi-coding-agent": {
       "optional": true
     },

package/schema.json

@@ -3,6 +3,32 @@
   "$schema": "http://json-schema.org/draft-07/schema#",
   "definitions": {
     "NeuralwattConfig": {
+      "additionalProperties": false,
+      "properties": {
+        "$schema": {
+          "description": "$schema URL for editor autocomplete.",
+          "type": "string"
+        },
+        "provider": {
+          "$ref": "#/definitions/NeuralwattProviderConfig",
+          "description": "Provider/model behavior."
+        },
+        "quotaCommand": {
+          "$ref": "#/definitions/NeuralwattQuotaCommandConfig",
+          "description": "Quota command feature."
+        },
+        "quotaWarnings": {
+          "$ref": "#/definitions/NeuralwattQuotaWarningsConfig",
+          "description": "Quota warning feature."
+        },
+        "subBarIntegration": {
+          "$ref": "#/definitions/NeuralwattSubBarIntegrationConfig",
+          "description": "Sub-bar/status-bar integration feature."
+        }
+      },
+      "type": "object"
+    },
+    "NeuralwattProviderConfig": {
       "additionalProperties": false,
       "properties": {
         "includeHiddenModels": {
@@ -12,16 +38,34 @@
         "includeLegacyModelIds": {
           "description": "Include legacy Neuralwatt model IDs in the model picker.",
           "type": "boolean"
-        },
-        "quotaCommand": {
+        }
+      },
+      "type": "object"
+    },
+    "NeuralwattQuotaCommandConfig": {
+      "additionalProperties": false,
+      "properties": {
+        "enabled": {
           "description": "Show the quota command (/neuralwatt:quota).",
           "type": "boolean"
-        },
-        "quotaWarnings": {
+        }
+      },
+      "type": "object"
+    },
+    "NeuralwattQuotaWarningsConfig": {
+      "additionalProperties": false,
+      "properties": {
+        "enabled": {
           "description": "Show quota warnings when credits or energy are low.",
           "type": "boolean"
-        },
-        "subBarIntegration": {
+        }
+      },
+      "type": "object"
+    },
+    "NeuralwattSubBarIntegrationConfig": {
+      "additionalProperties": false,
+      "properties": {
+        "enabled": {
           "description": "Show usage in the sub-bar / status bar.",
           "type": "boolean"
         }

package/src/config/defaults.ts

@@ -0,0 +1,17 @@
+import type { ResolvedNeuralwattConfig } from "./types";
+
+export const DEFAULT_CONFIG: ResolvedNeuralwattConfig = {
+  provider: {
+    includeLegacyModelIds: false,
+    includeHiddenModels: false,
+  },
+  quotaCommand: {
+    enabled: true,
+  },
+  quotaWarnings: {
+    enabled: true,
+  },
+  subBarIntegration: {
+    enabled: true,
+  },
+};

package/src/config/index.ts

@@ -0,0 +1,9 @@
+export { DEFAULT_CONFIG } from "./defaults";
+export { configLoader } from "./loader";
+export { migrations } from "./migration";
+export type {
+  NeuralwattConfig,
+  NeuralwattRawConfig,
+  PreviousNeuralwattConfig,
+  ResolvedNeuralwattConfig,
+} from "./types";

package/src/config/loader.ts

@@ -0,0 +1,74 @@
+import { buildSchemaUrl, ConfigLoader } from "@aliou/pi-utils-settings";
+import packageJson from "../../package.json";
+import { DEFAULT_CONFIG } from "./defaults";
+import { migrations } from "./migration";
+import type { NeuralwattRawConfig, ResolvedNeuralwattConfig } from "./types";
+
+type ConfigRecord = Record<string, unknown>;
+
+type MaybeEnabled = { enabled?: boolean };
+type MaybeProvider = {
+  includeLegacyModelIds?: boolean;
+  includeHiddenModels?: boolean;
+};
+
+function featureEnabled(value: unknown, fallback: boolean): boolean {
+  if (typeof value === "boolean") return value;
+  if (value && typeof value === "object") {
+    const { enabled } = value as MaybeEnabled;
+    if (typeof enabled === "boolean") return enabled;
+  }
+  return fallback;
+}
+
+function normalizeResolvedConfig(
+  resolved: ResolvedNeuralwattConfig,
+): ResolvedNeuralwattConfig {
+  const record = resolved as unknown as ConfigRecord;
+  const provider =
+    record.provider && typeof record.provider === "object"
+      ? (record.provider as MaybeProvider)
+      : {};
+
+  return {
+    provider: {
+      includeLegacyModelIds:
+        provider.includeLegacyModelIds ??
+        (typeof record.includeLegacyModelIds === "boolean"
+          ? record.includeLegacyModelIds
+          : DEFAULT_CONFIG.provider.includeLegacyModelIds),
+      includeHiddenModels:
+        provider.includeHiddenModels ??
+        (typeof record.includeHiddenModels === "boolean"
+          ? record.includeHiddenModels
+          : DEFAULT_CONFIG.provider.includeHiddenModels),
+    },
+    quotaCommand: {
+      enabled: featureEnabled(
+        record.quotaCommand,
+        DEFAULT_CONFIG.quotaCommand.enabled,
+      ),
+    },
+    quotaWarnings: {
+      enabled: featureEnabled(
+        record.quotaWarnings,
+        DEFAULT_CONFIG.quotaWarnings.enabled,
+      ),
+    },
+    subBarIntegration: {
+      enabled: featureEnabled(
+        record.subBarIntegration,
+        DEFAULT_CONFIG.subBarIntegration.enabled,
+      ),
+    },
+  };
+}
+
+export const configLoader = new ConfigLoader<
+  NeuralwattRawConfig,
+  ResolvedNeuralwattConfig
+>("neuralwatt", DEFAULT_CONFIG, {
+  migrations,
+  schemaUrl: buildSchemaUrl("@aliou/pi-neuralwatt", packageJson.version),
+  afterMerge: normalizeResolvedConfig,
+});

package/src/config/migration/01-disable-legacy-model-ids-by-default.ts

@@ -0,0 +1,44 @@
+import type { Migration } from "@aliou/pi-utils-settings";
+import type { NeuralwattConfig, PreviousNeuralwattConfig } from "../types";
+
+type MigrationConfig = PreviousNeuralwattConfig | NeuralwattConfig;
+
+type MutableConfigRecord = Record<string, unknown>;
+
+function hasOwn(record: MutableConfigRecord, key: string): boolean {
+  return Object.hasOwn(record, key);
+}
+
+function hasNestedConfig(config: MigrationConfig): boolean {
+  return Boolean(
+    ("provider" in config && config.provider) ||
+      (config.quotaCommand && typeof config.quotaCommand === "object") ||
+      (config.quotaWarnings && typeof config.quotaWarnings === "object") ||
+      (config.subBarIntegration &&
+        typeof config.subBarIntegration === "object"),
+  );
+}
+
+function isPreviousConfigWithoutLegacyDefault(
+  config: MigrationConfig,
+): config is PreviousNeuralwattConfig {
+  return (
+    !hasNestedConfig(config) &&
+    !hasOwn(config as MutableConfigRecord, "includeLegacyModelIds")
+  );
+}
+
+export const disableLegacyModelIdsByDefaultMigration: Migration<MigrationConfig> =
+  {
+    name: "disable-legacy-model-ids-by-default",
+    shouldRun: isPreviousConfigWithoutLegacyDefault,
+    message:
+      "[neuralwatt] legacy model IDs (ids including the provider and the quantization) are disabled by default. You can enable them with /neuralwatt:settings.",
+    run: (config) => {
+      const previous = config as PreviousNeuralwattConfig;
+      return {
+        ...previous,
+        includeLegacyModelIds: false,
+      };
+    },
+  };