fix(langchain): Fix serialization for initChatModel (#7222)

Author: jacoblee93

langchain/src/chat_models/tests/universal.int.test.ts

@@ -6,6 +6,7 @@ import { ChatPromptTemplate, PromptTemplate } from "@langchain/core/prompts";
 import { RunLogPatch, StreamEvent } from "@langchain/core/tracers/log_stream";
 import { AIMessageChunk } from "@langchain/core/messages";
 import { concat } from "@langchain/core/utils/stream";
+import { awaitAllCallbacks } from "@langchain/core/callbacks/promises";
 import { AgentExecutor, createReactAgent } from "../../agents/index.js";
 import { pull } from "../../hub.js";
 import { initChatModel } from "../universal.js";
@@ -32,7 +33,7 @@ const googleApiKey = process.env.GOOGLE_API_KEY;
 process.env.GOOGLE_API_KEY = "";
 
 test("Initialize non-configurable models", async () => {
-  const gpt4 = await initChatModel("gpt-4", {
+  const gpt4 = await initChatModel("gpt-4o-mini", {
     modelProvider: "openai",
     temperature: 0.25, // Funky temperature to verify it's being set properly.
     apiKey: openAIApiKey,
@@ -67,7 +68,7 @@ test("Create a partially configurable model with no default model", async () =>
 
   const gpt4Result = await configurableModel.invoke("what's your name", {
     configurable: {
-      model: "gpt-4",
+      model: "gpt-4o-mini",
       apiKey: openAIApiKey,
     },
   });
@@ -85,7 +86,7 @@ test("Create a partially configurable model with no default model", async () =>
 });
 
 test("Create a fully configurable model with a default model and a config prefix", async () => {
-  const configurableModelWithDefault = await initChatModel("gpt-4", {
+  const configurableModelWithDefault = await initChatModel("gpt-4o-mini", {
     modelProvider: "openai",
     configurableFields: "any",
     configPrefix: "foo",
@@ -155,7 +156,7 @@ test("Bind tools to a configurable model", async () => {
     }
   );
 
-  const configurableModel = await initChatModel("gpt-4", {
+  const configurableModel = await initChatModel("gpt-4o-mini", {
     configurableFields: ["model", "modelProvider", "apiKey"],
     temperature: 0,
   });
@@ -602,3 +603,38 @@ describe("Can call base runnable methods", () => {
     expect(result.tool_calls?.[0].name).toBe("GetWeather");
   });
 });
+
+describe("Serialization", () => {
+  it("does not contain additional fields", async () => {
+    const gpt4 = await initChatModel("gpt-4o-mini", {
+      modelProvider: "openai",
+      temperature: 0.25, // Funky temperature to verify it's being set properly.
+      apiKey: openAIApiKey,
+    });
+    let serializedRepresentation;
+    const res = await gpt4.invoke("foo", {
+      callbacks: [
+        {
+          handleChatModelStart(llm) {
+            serializedRepresentation = llm;
+          },
+        },
+      ],
+      configurable: { extra: "bar" },
+    });
+    await awaitAllCallbacks();
+    expect(res).toBeDefined();
+    const { ChatOpenAI } = await import("@langchain/openai");
+    expect(serializedRepresentation).toEqual(
+      JSON.parse(
+        JSON.stringify(
+          new ChatOpenAI({
+            model: "gpt-4o-mini",
+            temperature: 0.25,
+            apiKey: openAIApiKey,
+          })
+        )
+      )
+    );
+  });
+});

langchain/src/chat_models/universal.ts

@@ -73,50 +73,51 @@ async function _initChatModelHelper(
       `Unable to infer model provider for { model: ${model} }, please specify modelProvider directly.`
     );
   }
+  const { modelProvider: _unused, ...passedParams } = params;
 
   try {
     switch (modelProviderCopy) {
       case "openai": {
         const { ChatOpenAI } = await import("@langchain/openai");
-        return new ChatOpenAI({ model, ...params });
+        return new ChatOpenAI({ model, ...passedParams });
       }
       case "anthropic": {
         const { ChatAnthropic } = await import("@langchain/anthropic");
-        return new ChatAnthropic({ model, ...params });
+        return new ChatAnthropic({ model, ...passedParams });
       }
       case "azure_openai": {
         const { AzureChatOpenAI } = await import("@langchain/openai");
-        return new AzureChatOpenAI({ model, ...params });
+        return new AzureChatOpenAI({ model, ...passedParams });
       }
       case "cohere": {
         const { ChatCohere } = await import("@langchain/cohere");
-        return new ChatCohere({ model, ...params });
+        return new ChatCohere({ model, ...passedParams });
       }
       case "google-vertexai": {
         const { ChatVertexAI } = await import("@langchain/google-vertexai");
-        return new ChatVertexAI({ model, ...params });
+        return new ChatVertexAI({ model, ...passedParams });
       }
       case "google-genai": {
         const { ChatGoogleGenerativeAI } = await import(
           "@langchain/google-genai"
         );
-        return new ChatGoogleGenerativeAI({ model, ...params });
+        return new ChatGoogleGenerativeAI({ model, ...passedParams });
       }
       case "ollama": {
         const { ChatOllama } = await import("@langchain/ollama");
-        return new ChatOllama({ model, ...params });
+        return new ChatOllama({ model, ...passedParams });
       }
       case "mistralai": {
         const { ChatMistralAI } = await import("@langchain/mistralai");
-        return new ChatMistralAI({ model, ...params });
+        return new ChatMistralAI({ model, ...passedParams });
       }
       case "groq": {
         const { ChatGroq } = await import("@langchain/groq");
-        return new ChatGroq({ model, ...params });
+        return new ChatGroq({ model, ...passedParams });
       }
       case "bedrock": {
         const { ChatBedrockConverse } = await import("@langchain/aws");
-        return new ChatBedrockConverse({ model, ...params });
+        return new ChatBedrockConverse({ model, ...passedParams });
       }
       case "fireworks": {
         const { ChatFireworks } = await import(
@@ -127,7 +128,7 @@ async function _initChatModelHelper(
           // @ts-ignore - Can not install as a proper dependency due to circular dependency
           "@langchain/community/chat_models/fireworks"
         );
-        return new ChatFireworks({ model, ...params });
+        return new ChatFireworks({ model, ...passedParams });
       }
       case "together": {
         const { ChatTogetherAI } = await import(
@@ -138,7 +139,7 @@ async function _initChatModelHelper(
           // @ts-ignore - Can not install as a proper dependency due to circular dependency
           "@langchain/community/chat_models/togetherai"
         );
-        return new ChatTogetherAI({ model, ...params });
+        return new ChatTogetherAI({ model, ...passedParams });
       }
       default: {
         const supported = _SUPPORTED_PROVIDERS.join(", ");
@@ -247,7 +248,10 @@ class _ConfigurableModel<
     if (fields.configurableFields === "any") {
       this._configurableFields = "any";
     } else {
-      this._configurableFields = fields.configurableFields ?? "any";
+      this._configurableFields = fields.configurableFields ?? [
+        "model",
+        "modelProvider",
+      ];
     }
 
     if (fields.configPrefix) {
@@ -786,12 +790,14 @@ export async function initChatModel<
     configPrefix: "",
     ...(fields ?? {}),
   };
-  let configurableFieldsCopy = configurableFields;
+  let configurableFieldsCopy = Array.isArray(configurableFields)
+    ? [...configurableFields]
+    : configurableFields;
 
-  if (!model && !configurableFieldsCopy) {
+  if (!model && configurableFieldsCopy === undefined) {
     configurableFieldsCopy = ["model", "modelProvider"];
   }
-  if (configPrefix && !configurableFieldsCopy) {
+  if (configPrefix && configurableFieldsCopy === undefined) {
     console.warn(
       `{ configPrefix: ${configPrefix} } has been set but no fields are configurable. Set ` +
         `{ configurableFields: [...] } to specify the model params that are ` +
@@ -802,7 +808,7 @@ export async function initChatModel<
   // eslint-disable-next-line @typescript-eslint/no-explicit-any
   const paramsCopy: Record<string, any> = { ...params };
 
-  if (!configurableFieldsCopy) {
+  if (configurableFieldsCopy === undefined) {
     return new _ConfigurableModel<RunInput, CallOptions>({
       defaultConfig: {
         ...paramsCopy,