@ryanfw/prompt-orchestration-pipeline 1.2.6 → 1.2.7

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ryanfw/prompt-orchestration-pipeline",
-  "version": "1.2.6",
+  "version": "1.2.7",
   "description": "A Prompt-orchestration pipeline (POP) is a framework for building, running, and experimenting with complex chains of LLM tasks.",
   "type": "module",
   "main": "src/ui/server/index.ts",

package/src/config/__tests__/models.test.ts

@@ -15,7 +15,7 @@ import {
 } from "../models";
 import type { ModelConfigEntry } from "../models";
 
-const MODEL_COUNT = 41;
+const MODEL_COUNT = 48;
 const PROVIDER_COUNT = 8;
 
 describe("ModelAlias", () => {
@@ -110,8 +110,8 @@ describe("DEFAULT_MODEL_BY_PROVIDER", () => {
 });
 
 describe("aliasToFunctionName", () => {
-  it('converts "openai:gpt-5.2" to "gpt52"', () => {
-    expect(aliasToFunctionName("openai:gpt-5.2")).toBe("gpt52");
+  it('converts "openai:gpt-5.4" to "gpt54"', () => {
+    expect(aliasToFunctionName("openai:gpt-5.4")).toBe("gpt54");
   });
 
   it('converts "gemini:flash-2.5-lite" to "flash25Lite"', () => {
@@ -137,8 +137,8 @@ describe("aliasToFunctionName", () => {
 });
 
 describe("getProviderFromAlias", () => {
-  it('returns "openai" for "openai:gpt-5.2"', () => {
-    expect(getProviderFromAlias("openai:gpt-5.2")).toBe("openai");
+  it('returns "openai" for "openai:gpt-5.4"', () => {
+    expect(getProviderFromAlias("openai:gpt-5.4")).toBe("openai");
   });
 
   it("throws for non-string input", () => {
@@ -152,8 +152,8 @@ describe("getProviderFromAlias", () => {
 });
 
 describe("getModelFromAlias", () => {
-  it('returns "gpt-5.2" for "openai:gpt-5.2"', () => {
-    expect(getModelFromAlias("openai:gpt-5.2")).toBe("gpt-5.2");
+  it('returns "gpt-5.4" for "openai:gpt-5.4"', () => {
+    expect(getModelFromAlias("openai:gpt-5.4")).toBe("gpt-5.4");
   });
 
   it("handles multiple colons by rejoining segments after first", () => {
@@ -171,8 +171,8 @@ describe("getModelFromAlias", () => {
 });
 
 describe("getModelConfig", () => {
-  it('returns config for "openai:gpt-5.2" with provider "openai"', () => {
-    const config = getModelConfig("openai:gpt-5.2");
+  it('returns config for "openai:gpt-5.4" with provider "openai"', () => {
+    const config = getModelConfig("openai:gpt-5.4");
     expect(config).not.toBeNull();
     expect(config!.provider).toBe("openai");
   });
@@ -184,10 +184,20 @@ describe("getModelConfig", () => {
 
   it("returns null for removed aliases", () => {
     expect(getModelConfig("openai:gpt-4")).toBeNull();
+    expect(getModelConfig("openai:gpt-4o")).toBeNull();
+    expect(getModelConfig("openai:gpt-4o-mini")).toBeNull();
+    expect(getModelConfig("openai:gpt-5.2")).toBeNull();
+    expect(getModelConfig("openai:gpt-5.2-pro")).toBeNull();
+    expect(getModelConfig("openai:o3-mini")).toBeNull();
+    expect(getModelConfig("anthropic:haiku-4-6")).toBeNull();
+    expect(getModelConfig("deepseek:deepseek-chat")).toBeNull();
+    expect(getModelConfig("deepseek:deepseek-reasoner")).toBeNull();
     expect(getModelConfig("deepseek:r1")).toBeNull();
+    expect(getModelConfig("moonshot:kimi-k1.5")).toBeNull();
     expect(getModelConfig("moonshot:kimi-moonshot-v1-128k")).toBeNull();
     expect(getModelConfig("gemini:flash-3")).toBeNull();
     expect(getModelConfig("gemini:pro-3")).toBeNull();
+    expect(getModelConfig("zai:glm-5-code")).toBeNull();
     expect(getModelConfig("zai:glm-4-plus")).toBeNull();
   });
 });
@@ -201,8 +211,8 @@ describe("FUNCTION_NAME_BY_ALIAS", () => {
     expect(Object.isFrozen(FUNCTION_NAME_BY_ALIAS)).toBe(true);
   });
 
-  it("has correct value for openai:gpt-5.2", () => {
-    expect(FUNCTION_NAME_BY_ALIAS["openai:gpt-5.2"]).toBe("gpt52");
+  it("has correct value for openai:gpt-5.4", () => {
+    expect(FUNCTION_NAME_BY_ALIAS["openai:gpt-5.4"]).toBe("gpt54");
   });
 
   it("has correct value for gemini:flash-2.5-lite", () => {

package/src/config/models.ts

@@ -41,12 +41,12 @@ export const ModelAlias = Object.freeze({
   OPENAI_GPT_4_1: "openai:gpt-4-1",
   OPENAI_GPT_4_1_MINI: "openai:gpt-4-1-mini",
   OPENAI_GPT_4_1_NANO: "openai:gpt-4-1-nano",
-  OPENAI_GPT_4O: "openai:gpt-4o",
-  OPENAI_GPT_4O_MINI: "openai:gpt-4o-mini",
-  OPENAI_GPT_5_2: "openai:gpt-5.2",
-  OPENAI_GPT_5_2_PRO: "openai:gpt-5.2-pro",
+  OPENAI_GPT_5_4: "openai:gpt-5.4",
+  OPENAI_GPT_5_4_MINI: "openai:gpt-5.4-mini",
+  OPENAI_GPT_5_4_NANO: "openai:gpt-5.4-nano",
+  OPENAI_GPT_5_5: "openai:gpt-5.5",
+  OPENAI_GPT_5_5_PRO: "openai:gpt-5.5-pro",
   OPENAI_O3: "openai:o3",
-  OPENAI_O3_MINI: "openai:o3-mini",
   OPENAI_O4_MINI: "openai:o4-mini",
   // Anthropic
   ANTHROPIC_OPUS_4_5: "anthropic:opus-4-5",
@@ -54,33 +54,40 @@ export const ModelAlias = Object.freeze({
   ANTHROPIC_HAIKU_4_5: "anthropic:haiku-4-5",
   ANTHROPIC_OPUS_4_6: "anthropic:opus-4-6",
   ANTHROPIC_SONNET_4_6: "anthropic:sonnet-4-6",
-  ANTHROPIC_HAIKU_4_6: "anthropic:haiku-4-6",
+  ANTHROPIC_OPUS_4_7: "anthropic:opus-4-7",
   // Gemini
   GEMINI_FLASH_2_5: "gemini:flash-2.5",
   GEMINI_FLASH_2_5_LITE: "gemini:flash-2.5-lite",
   GEMINI_PRO_2_5: "gemini:pro-2.5",
-  // DeepSeek — cache miss prices
-  DEEPSEEK_CHAT: "deepseek:deepseek-chat",
-  DEEPSEEK_REASONER: "deepseek:deepseek-reasoner",
+  GEMINI_PRO_3_1_PREVIEW: "gemini:pro-3.1-preview",
+  GEMINI_FLASH_3_PREVIEW: "gemini:flash-3-preview",
+  GEMINI_FLASH_3_1_LITE_PREVIEW: "gemini:flash-3.1-lite-preview",
+  // DeepSeek
+  DEEPSEEK_V4_FLASH: "deepseek:v4-flash",
+  DEEPSEEK_V4_PRO: "deepseek:v4-pro",
   // Moonshot / Kimi
   MOONSHOT_KIMI_K2_5: "moonshot:kimi-k2.5",
-  MOONSHOT_KIMI_K1_5: "moonshot:kimi-k1.5",
+  MOONSHOT_KIMI_K2_6: "moonshot:kimi-k2.6",
   // Claude Code — subscription-based, zero token cost
   CLAUDE_CODE_SONNET: "claude-code:sonnet",
   CLAUDE_CODE_OPUS: "claude-code:opus",
   CLAUDE_CODE_HAIKU: "claude-code:haiku",
   // Z.ai
+  ZAI_GLM_5_1: "zai:glm-5-1",
   ZAI_GLM_5: "zai:glm-5",
-  ZAI_GLM_5_CODE: "zai:glm-5-code",
+  ZAI_GLM_5_TURBO: "zai:glm-5-turbo",
   ZAI_GLM_4_7: "zai:glm-4-7",
   ZAI_GLM_4_7_FLASH_X: "zai:glm-4-7-flash-x",
   ZAI_GLM_4_6: "zai:glm-4-6",
   ZAI_GLM_4_5: "zai:glm-4-5",
+  ZAI_GLM_4_5_X: "zai:glm-4-5-x",
   ZAI_GLM_4_5_AIR: "zai:glm-4-5-air",
   ZAI_GLM_4_5_AIR_X: "zai:glm-4-5-air-x",
-  // Alibaba (Qwen via DashScope)
+  // Alibaba (Qwen via DashScope, international/Singapore deployment)
   ALIBABA_QWEN3_MAX: "alibaba:qwen3-max",
+  ALIBABA_QWEN3_6_PLUS: "alibaba:qwen3.6-plus",
   ALIBABA_QWEN3_5_PLUS: "alibaba:qwen3.5-plus",
+  ALIBABA_QWEN3_5_FLASH: "alibaba:qwen3.5-flash",
   ALIBABA_QWEN_PLUS: "alibaba:qwen-plus",
   ALIBABA_QWEN_FLASH: "alibaba:qwen-flash",
   ALIBABA_QWQ_PLUS: "alibaba:qwq-plus",
@@ -112,41 +119,41 @@ const MODEL_CONFIG_RAW: Record<ModelAliasKey, ModelConfigEntry> = {
     tokenCostInPerMillion: 0.1,
     tokenCostOutPerMillion: 0.4,
   },
-  "openai:gpt-4o": {
+  "openai:gpt-5.4": {
     provider: "openai",
-    model: "gpt-4o",
+    model: "gpt-5.4",
     tokenCostInPerMillion: 2.5,
-    tokenCostOutPerMillion: 10,
+    tokenCostOutPerMillion: 15,
   },
-  "openai:gpt-4o-mini": {
+  "openai:gpt-5.4-mini": {
     provider: "openai",
-    model: "gpt-4o-mini",
-    tokenCostInPerMillion: 0.15,
-    tokenCostOutPerMillion: 0.6,
+    model: "gpt-5.4-mini",
+    tokenCostInPerMillion: 0.75,
+    tokenCostOutPerMillion: 4.5,
   },
-  "openai:gpt-5.2": {
+  "openai:gpt-5.4-nano": {
     provider: "openai",
-    model: "gpt-5.2",
-    tokenCostInPerMillion: 2,
-    tokenCostOutPerMillion: 10,
+    model: "gpt-5.4-nano",
+    tokenCostInPerMillion: 0.2,
+    tokenCostOutPerMillion: 1.25,
   },
-  "openai:gpt-5.2-pro": {
+  "openai:gpt-5.5": {
     provider: "openai",
-    model: "gpt-5.2-pro",
-    tokenCostInPerMillion: 15,
-    tokenCostOutPerMillion: 60,
+    model: "gpt-5.5",
+    tokenCostInPerMillion: 5,
+    tokenCostOutPerMillion: 30,
   },
-  "openai:o3": {
+  "openai:gpt-5.5-pro": {
     provider: "openai",
-    model: "o3",
-    tokenCostInPerMillion: 10,
-    tokenCostOutPerMillion: 40,
+    model: "gpt-5.5-pro",
+    tokenCostInPerMillion: 30,
+    tokenCostOutPerMillion: 180,
   },
-  "openai:o3-mini": {
+  "openai:o3": {
     provider: "openai",
-    model: "o3-mini",
-    tokenCostInPerMillion: 1.1,
-    tokenCostOutPerMillion: 4.4,
+    model: "o3",
+    tokenCostInPerMillion: 2,
+    tokenCostOutPerMillion: 8,
   },
   "openai:o4-mini": {
     provider: "openai",
@@ -185,13 +192,13 @@ const MODEL_CONFIG_RAW: Record<ModelAliasKey, ModelConfigEntry> = {
     tokenCostInPerMillion: 3,
     tokenCostOutPerMillion: 15,
   },
-  "anthropic:haiku-4-6": {
+  "anthropic:opus-4-7": {
     provider: "anthropic",
-    model: "claude-haiku-4-6",
-    tokenCostInPerMillion: 1,
-    tokenCostOutPerMillion: 5,
+    model: "claude-opus-4-7",
+    tokenCostInPerMillion: 5,
+    tokenCostOutPerMillion: 25,
   },
-  // Gemini
+  // Gemini — base (≤200k context) pricing for tiered models
   "gemini:flash-2.5": {
     provider: "gemini",
     model: "gemini-2.5-flash",
@@ -210,31 +217,49 @@ const MODEL_CONFIG_RAW: Record<ModelAliasKey, ModelConfigEntry> = {
     tokenCostInPerMillion: 1.25,
     tokenCostOutPerMillion: 10,
   },
+  "gemini:pro-3.1-preview": {
+    provider: "gemini",
+    model: "gemini-3.1-pro-preview",
+    tokenCostInPerMillion: 2,
+    tokenCostOutPerMillion: 12,
+  },
+  "gemini:flash-3-preview": {
+    provider: "gemini",
+    model: "gemini-3-flash-preview",
+    tokenCostInPerMillion: 0.5,
+    tokenCostOutPerMillion: 3,
+  },
+  "gemini:flash-3.1-lite-preview": {
+    provider: "gemini",
+    model: "gemini-3.1-flash-lite-preview",
+    tokenCostInPerMillion: 0.25,
+    tokenCostOutPerMillion: 1.5,
+  },
   // DeepSeek — cache miss prices
-  "deepseek:deepseek-chat": {
+  "deepseek:v4-flash": {
     provider: "deepseek",
-    model: "deepseek-chat",
-    tokenCostInPerMillion: 0.28,
-    tokenCostOutPerMillion: 0.42,
+    model: "deepseek-v4-flash",
+    tokenCostInPerMillion: 0.14,
+    tokenCostOutPerMillion: 0.28,
   },
-  "deepseek:deepseek-reasoner": {
+  "deepseek:v4-pro": {
     provider: "deepseek",
-    model: "deepseek-reasoner",
-    tokenCostInPerMillion: 0.28,
-    tokenCostOutPerMillion: 0.42,
+    model: "deepseek-v4-pro",
+    tokenCostInPerMillion: 1.74,
+    tokenCostOutPerMillion: 3.48,
   },
-  // Moonshot / Kimi
+  // Moonshot / Kimi — cache miss prices
   "moonshot:kimi-k2.5": {
     provider: "moonshot",
     model: "kimi-k2.5",
-    tokenCostInPerMillion: 0.5,
+    tokenCostInPerMillion: 0.6,
     tokenCostOutPerMillion: 2.5,
   },
-  "moonshot:kimi-k1.5": {
+  "moonshot:kimi-k2.6": {
     provider: "moonshot",
-    model: "kimi-k1.5",
-    tokenCostInPerMillion: 0.5,
-    tokenCostOutPerMillion: 2.5,
+    model: "kimi-k2.6",
+    tokenCostInPerMillion: 0.95,
+    tokenCostOutPerMillion: 4,
   },
   // Claude Code — subscription-based, zero token cost
   "claude-code:sonnet": {
@@ -256,17 +281,23 @@ const MODEL_CONFIG_RAW: Record<ModelAliasKey, ModelConfigEntry> = {
     tokenCostOutPerMillion: 0,
   },
   // Z.ai
+  "zai:glm-5-1": {
+    provider: "zai",
+    model: "glm-5.1",
+    tokenCostInPerMillion: 1.4,
+    tokenCostOutPerMillion: 4.4,
+  },
   "zai:glm-5": {
     provider: "zai",
     model: "glm-5",
     tokenCostInPerMillion: 1,
     tokenCostOutPerMillion: 3.2,
   },
-  "zai:glm-5-code": {
+  "zai:glm-5-turbo": {
     provider: "zai",
-    model: "glm-5-code",
+    model: "glm-5-turbo",
     tokenCostInPerMillion: 1.2,
-    tokenCostOutPerMillion: 5,
+    tokenCostOutPerMillion: 4,
   },
   "zai:glm-4-7": {
     provider: "zai",
@@ -292,6 +323,12 @@ const MODEL_CONFIG_RAW: Record<ModelAliasKey, ModelConfigEntry> = {
     tokenCostInPerMillion: 0.6,
     tokenCostOutPerMillion: 2.2,
   },
+  "zai:glm-4-5-x": {
+    provider: "zai",
+    model: "glm-4.5-x",
+    tokenCostInPerMillion: 2.2,
+    tokenCostOutPerMillion: 8.9,
+  },
   "zai:glm-4-5-air": {
     provider: "zai",
     model: "glm-4.5-air",
@@ -304,30 +341,42 @@ const MODEL_CONFIG_RAW: Record<ModelAliasKey, ModelConfigEntry> = {
     tokenCostInPerMillion: 1.1,
     tokenCostOutPerMillion: 4.5,
   },
-  // Alibaba (Qwen via DashScope)
+  // Alibaba (Qwen via DashScope, international/Singapore deployment, base tier)
   "alibaba:qwen3-max": {
     provider: "alibaba",
     model: "qwen3-max",
-    tokenCostInPerMillion: 0.359,
-    tokenCostOutPerMillion: 1.434,
+    tokenCostInPerMillion: 1.2,
+    tokenCostOutPerMillion: 6,
+  },
+  "alibaba:qwen3.6-plus": {
+    provider: "alibaba",
+    model: "qwen3.6-plus",
+    tokenCostInPerMillion: 0.276,
+    tokenCostOutPerMillion: 1.651,
   },
   "alibaba:qwen3.5-plus": {
     provider: "alibaba",
     model: "qwen3.5-plus",
-    tokenCostInPerMillion: 0.115,
-    tokenCostOutPerMillion: 0.688,
+    tokenCostInPerMillion: 0.4,
+    tokenCostOutPerMillion: 2.4,
+  },
+  "alibaba:qwen3.5-flash": {
+    provider: "alibaba",
+    model: "qwen3.5-flash",
+    tokenCostInPerMillion: 0.1,
+    tokenCostOutPerMillion: 0.4,
   },
   "alibaba:qwen-plus": {
     provider: "alibaba",
     model: "qwen-plus",
-    tokenCostInPerMillion: 0.115,
-    tokenCostOutPerMillion: 0.287,
+    tokenCostInPerMillion: 0.4,
+    tokenCostOutPerMillion: 1.2,
   },
   "alibaba:qwen-flash": {
     provider: "alibaba",
     model: "qwen-flash",
-    tokenCostInPerMillion: 0.022,
-    tokenCostOutPerMillion: 0.216,
+    tokenCostInPerMillion: 0.05,
+    tokenCostOutPerMillion: 0.4,
   },
   "alibaba:qwq-plus": {
     provider: "alibaba",
@@ -338,14 +387,14 @@ const MODEL_CONFIG_RAW: Record<ModelAliasKey, ModelConfigEntry> = {
   "alibaba:qwen3-coder-plus": {
     provider: "alibaba",
     model: "qwen3-coder-plus",
-    tokenCostInPerMillion: 0.574,
-    tokenCostOutPerMillion: 2.294,
+    tokenCostInPerMillion: 1,
+    tokenCostOutPerMillion: 5,
   },
   "alibaba:qwen3-coder-flash": {
     provider: "alibaba",
     model: "qwen3-coder-flash",
-    tokenCostInPerMillion: 0.144,
-    tokenCostOutPerMillion: 0.574,
+    tokenCostInPerMillion: 0.3,
+    tokenCostOutPerMillion: 1.5,
   },
 };
 
@@ -397,13 +446,13 @@ export function getModelConfig(alias: string): ModelConfigEntry | null {
 
 export const DEFAULT_MODEL_BY_PROVIDER: Readonly<Record<ProviderName, ModelAliasKey>> =
   Object.freeze({
-    openai: "openai:gpt-5.2",
+    openai: "openai:gpt-5.4",
     anthropic: "anthropic:sonnet-4-6",
     gemini: "gemini:flash-2.5",
-    deepseek: "deepseek:deepseek-chat",
-    moonshot: "moonshot:kimi-k2.5",
+    deepseek: "deepseek:v4-flash",
+    moonshot: "moonshot:kimi-k2.6",
     "claude-code": "claude-code:sonnet",
-    zai: "zai:glm-5",
+    zai: "zai:glm-5-1",
     alibaba: "alibaba:qwen3-max",
   } as const);
 

package/src/core/__tests__/lifecycle-policy.test.ts

@@ -0,0 +1,78 @@
+import { describe, test, expect } from "bun:test";
+import { decideTransition } from "../lifecycle-policy";
+
+describe("decideTransition", () => {
+  test("start with pending task and ready dependencies returns ok", () => {
+    const result = decideTransition({
+      op: "start",
+      taskState: "pending",
+      dependenciesReady: true,
+    });
+    expect(result).toEqual({ ok: true });
+  });
+
+  test("start with pending task and unready dependencies is blocked by dependencies", () => {
+    const result = decideTransition({
+      op: "start",
+      taskState: "pending",
+      dependenciesReady: false,
+    });
+    expect(result).toEqual({
+      ok: false,
+      code: "unsupported_lifecycle",
+      reason: "dependencies",
+    });
+  });
+
+  test("restart from done state returns ok regardless of dependencies", () => {
+    const result = decideTransition({
+      op: "restart",
+      taskState: "done",
+      dependenciesReady: false,
+    });
+    expect(result).toEqual({ ok: true });
+  });
+
+  test("restart from running state is blocked by policy", () => {
+    const result = decideTransition({
+      op: "restart",
+      taskState: "running",
+      dependenciesReady: true,
+    });
+    expect(result).toEqual({
+      ok: false,
+      code: "unsupported_lifecycle",
+      reason: "policy",
+    });
+  });
+
+  test("throws for invalid op", () => {
+    expect(() =>
+      decideTransition({
+        op: "stop" as unknown as "start",
+        taskState: "pending",
+        dependenciesReady: true,
+      })
+    ).toThrow(/Invalid op/);
+  });
+
+  test("throws for non-string taskState", () => {
+    expect(() =>
+      decideTransition({
+        op: "start",
+        taskState: 42 as unknown as string,
+        dependenciesReady: true,
+      })
+    ).toThrow(/Invalid taskState/);
+  });
+
+  test("throws for non-boolean dependenciesReady", () => {
+    expect(() =>
+      decideTransition({
+        op: "start",
+        taskState: "pending",
+        dependenciesReady: "yes" as unknown as boolean,
+      })
+    ).toThrow(/Invalid dependenciesReady/);
+  });
+});