langchain 0.1.1 → 0.1.3

package/dist/agents/agent.cjs

@@ -82,7 +82,7 @@ function isAgentAction(input) {
 }
 /**
  * Class representing a single action agent which accepts runnables.
- * Extends the BaseSingleActionAgent class and provides methods for
+ * Extends the BaseMultiActionAgent class and provides methods for
  * planning agent actions with runnables.
  */
 class RunnableAgent extends BaseMultiActionAgent {

package/dist/agents/agent.d.ts

@@ -77,7 +77,7 @@ export declare abstract class BaseMultiActionAgent extends BaseAgent {
 }
 /**
  * Class representing a single action agent which accepts runnables.
- * Extends the BaseSingleActionAgent class and provides methods for
+ * Extends the BaseMultiActionAgent class and provides methods for
  * planning agent actions with runnables.
  */
 export declare class RunnableAgent extends BaseMultiActionAgent {

package/dist/agents/agent.js

@@ -76,7 +76,7 @@ function isAgentAction(input) {
 }
 /**
  * Class representing a single action agent which accepts runnables.
- * Extends the BaseSingleActionAgent class and provides methods for
+ * Extends the BaseMultiActionAgent class and provides methods for
  * planning agent actions with runnables.
  */
 export class RunnableAgent extends BaseMultiActionAgent {

package/dist/agents/openai_functions/index.cjs

@@ -2,7 +2,7 @@
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.createOpenAIFunctionsAgent = exports.OpenAIAgent = exports._formatIntermediateSteps = void 0;
 const runnables_1 = require("@langchain/core/runnables");
-const openai_1 = require("@langchain/openai");
+const function_calling_1 = require("@langchain/core/utils/function_calling");
 const messages_1 = require("@langchain/core/messages");
 const prompts_1 = require("@langchain/core/prompts");
 const agent_js_1 = require("../agent.cjs");
@@ -145,7 +145,7 @@ class OpenAIAgent extends agent_js_1.Agent {
         const llm = this.llmChain.llm;
         const valuesForPrompt = { ...newInputs };
         const valuesForLLM = {
-            functions: this.tools.map(openai_1.formatToOpenAIFunction),
+            functions: this.tools.map(function_calling_1.convertToOpenAIFunction),
         };
         const callKeys = "callKeys" in this.llmChain.llm ? this.llmChain.llm.callKeys : [];
         for (const key of callKeys) {
@@ -227,7 +227,7 @@ async function createOpenAIFunctionsAgent({ llm, tools, prompt, }) {
         ].join("\n"));
     }
     const llmWithTools = llm.bind({
-        functions: tools.map(openai_1.formatToOpenAIFunction),
+        functions: tools.map(function_calling_1.convertToOpenAIFunction),
     });
     const agent = runnables_1.RunnableSequence.from([
         runnables_1.RunnablePassthrough.assign({

package/dist/agents/openai_functions/index.js

@@ -1,5 +1,5 @@
 import { RunnablePassthrough, RunnableSequence, } from "@langchain/core/runnables";
-import { formatToOpenAIFunction, } from "@langchain/openai";
+import { convertToOpenAIFunction } from "@langchain/core/utils/function_calling";
 import { AIMessage, FunctionMessage, } from "@langchain/core/messages";
 import { ChatPromptTemplate, HumanMessagePromptTemplate, MessagesPlaceholder, SystemMessagePromptTemplate, } from "@langchain/core/prompts";
 import { Agent } from "../agent.js";
@@ -141,7 +141,7 @@ export class OpenAIAgent extends Agent {
         const llm = this.llmChain.llm;
         const valuesForPrompt = { ...newInputs };
         const valuesForLLM = {
-            functions: this.tools.map(formatToOpenAIFunction),
+            functions: this.tools.map(convertToOpenAIFunction),
         };
         const callKeys = "callKeys" in this.llmChain.llm ? this.llmChain.llm.callKeys : [];
         for (const key of callKeys) {
@@ -222,7 +222,7 @@ export async function createOpenAIFunctionsAgent({ llm, tools, prompt, }) {
         ].join("\n"));
     }
     const llmWithTools = llm.bind({
-        functions: tools.map(formatToOpenAIFunction),
+        functions: tools.map(convertToOpenAIFunction),
     });
     const agent = RunnableSequence.from([
         RunnablePassthrough.assign({

package/dist/agents/openai_functions/output_parser.cjs

@@ -18,7 +18,7 @@ const types_js_1 = require("../types.cjs");
  *   modelName: "gpt-4",
  *   temperature: 0,
  * }).bind({
- *   functions: tools.map((tool) => formatToOpenAIFunction(tool)),
+ *   functions: tools.map((tool) => convertToOpenAIFunction(tool)),
  * });
  *
  * const runnableAgent = RunnableSequence.from([

package/dist/agents/openai_functions/output_parser.d.ts

@@ -22,7 +22,7 @@ export type FunctionsAgentAction = AgentAction & {
  *   modelName: "gpt-4",
  *   temperature: 0,
  * }).bind({
- *   functions: tools.map((tool) => formatToOpenAIFunction(tool)),
+ *   functions: tools.map((tool) => convertToOpenAIFunction(tool)),
  * });
  *
  * const runnableAgent = RunnableSequence.from([

package/dist/agents/openai_functions/output_parser.js

@@ -15,7 +15,7 @@ import { AgentActionOutputParser } from "../types.js";
  *   modelName: "gpt-4",
  *   temperature: 0,
  * }).bind({
- *   functions: tools.map((tool) => formatToOpenAIFunction(tool)),
+ *   functions: tools.map((tool) => convertToOpenAIFunction(tool)),
  * });
  *
  * const runnableAgent = RunnableSequence.from([

package/dist/agents/openai_tools/index.cjs

@@ -2,7 +2,7 @@
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.createOpenAIToolsAgent = exports.OpenAIToolsAgentOutputParser = void 0;
 const runnables_1 = require("@langchain/core/runnables");
-const openai_1 = require("@langchain/openai");
+const function_calling_1 = require("@langchain/core/utils/function_calling");
 const openai_tools_js_1 = require("../format_scratchpad/openai_tools.cjs");
 const output_parser_js_1 = require("../openai/output_parser.cjs");
 Object.defineProperty(exports, "OpenAIToolsAgentOutputParser", { enumerable: true, get: function () { return output_parser_js_1.OpenAIToolsAgentOutputParser; } });
@@ -69,7 +69,7 @@ async function createOpenAIToolsAgent({ llm, tools, prompt, }) {
             `Found ${JSON.stringify(prompt.inputVariables)} instead.`,
         ].join("\n"));
     }
-    const modelWithTools = llm.bind({ tools: tools.map(openai_1.formatToOpenAITool) });
+    const modelWithTools = llm.bind({ tools: tools.map(function_calling_1.convertToOpenAITool) });
     const agent = runnables_1.RunnableSequence.from([
         runnables_1.RunnablePassthrough.assign({
             agent_scratchpad: (input) => (0, openai_tools_js_1.formatToOpenAIToolMessages)(input.steps),

package/dist/agents/openai_tools/index.js

@@ -1,5 +1,5 @@
 import { RunnablePassthrough, RunnableSequence, } from "@langchain/core/runnables";
-import { formatToOpenAITool } from "@langchain/openai";
+import { convertToOpenAITool } from "@langchain/core/utils/function_calling";
 import { formatToOpenAIToolMessages } from "../format_scratchpad/openai_tools.js";
 import { OpenAIToolsAgentOutputParser, } from "../openai/output_parser.js";
 export { OpenAIToolsAgentOutputParser };
@@ -66,7 +66,7 @@ export async function createOpenAIToolsAgent({ llm, tools, prompt, }) {
             `Found ${JSON.stringify(prompt.inputVariables)} instead.`,
         ].join("\n"));
     }
-    const modelWithTools = llm.bind({ tools: tools.map(formatToOpenAITool) });
+    const modelWithTools = llm.bind({ tools: tools.map(convertToOpenAITool) });
     const agent = RunnableSequence.from([
         RunnablePassthrough.assign({
             agent_scratchpad: (input) => formatToOpenAIToolMessages(input.steps),

package/dist/agents/openai_tools/output_parser.cjs

@@ -24,7 +24,7 @@ const types_js_1 = require("../types.cjs");
  *   new ChatOpenAI({
  *     modelName: "gpt-3.5-turbo-1106",
  *     temperature: 0,
- *   }).bind({ tools: tools.map(formatToOpenAITool) }),
+ *   }).bind({ tools: tools.map(convertToOpenAITool) }),
  *   new OpenAIToolsAgentOutputParser(),
  * ]).withConfig({ runName: "OpenAIToolsAgent" });
  *

package/dist/agents/openai_tools/output_parser.d.ts

@@ -32,7 +32,7 @@ export type ToolsAgentStep = AgentStep & {
  *   new ChatOpenAI({
  *     modelName: "gpt-3.5-turbo-1106",
  *     temperature: 0,
- *   }).bind({ tools: tools.map(formatToOpenAITool) }),
+ *   }).bind({ tools: tools.map(convertToOpenAITool) }),
  *   new OpenAIToolsAgentOutputParser(),
  * ]).withConfig({ runName: "OpenAIToolsAgent" });
  *

package/dist/agents/openai_tools/output_parser.js

@@ -21,7 +21,7 @@ import { AgentMultiActionOutputParser } from "../types.js";
  *   new ChatOpenAI({
  *     modelName: "gpt-3.5-turbo-1106",
  *     temperature: 0,
- *   }).bind({ tools: tools.map(formatToOpenAITool) }),
+ *   }).bind({ tools: tools.map(convertToOpenAITool) }),
  *   new OpenAIToolsAgentOutputParser(),
  * ]).withConfig({ runName: "OpenAIToolsAgent" });
  *

package/dist/cache/cloudflare_kv.cjs

@@ -17,6 +17,7 @@ Object.defineProperty(exports, "__esModule", { value: true });
 const entrypoint_deprecation_js_1 = require("../util/entrypoint_deprecation.cjs");
 /* #__PURE__ */ (0, entrypoint_deprecation_js_1.logVersion010MigrationWarning)({
     oldEntrypointName: "cache/cloudflare_kv",
-    newEntrypointName: "caches/cloudflare_kv",
+    newEntrypointName: "",
+    newPackageName: "@langchain/cloudflare",
 });
 __exportStar(require("@langchain/community/caches/cloudflare_kv"), exports);

package/dist/cache/cloudflare_kv.js

@@ -1,6 +1,7 @@
 import { logVersion010MigrationWarning } from "../util/entrypoint_deprecation.js";
 /* #__PURE__ */ logVersion010MigrationWarning({
     oldEntrypointName: "cache/cloudflare_kv",
-    newEntrypointName: "caches/cloudflare_kv",
+    newEntrypointName: "",
+    newPackageName: "@langchain/cloudflare",
 });
 export * from "@langchain/community/caches/cloudflare_kv";

package/dist/chains/history_aware_retriever.cjs

@@ -24,7 +24,7 @@ const output_parsers_1 = require("@langchain/core/output_parsers");
  * const rephrasePrompt = await pull("langchain-ai/chat-langchain-rephrase");
  * const llm = new ChatOpenAI({});
  * const retriever = ...
- * const historyAwareRetrieverChain = await createHistoryAwareRetriever({
+ * const chain = await createHistoryAwareRetriever({
  *   llm,
  *   retriever,
  *   rephrasePrompt,

package/dist/chains/history_aware_retriever.d.ts

@@ -41,7 +41,7 @@ export type CreateHistoryAwareRetrieverParams = {
  * const rephrasePrompt = await pull("langchain-ai/chat-langchain-rephrase");
  * const llm = new ChatOpenAI({});
  * const retriever = ...
- * const historyAwareRetrieverChain = await createHistoryAwareRetriever({
+ * const chain = await createHistoryAwareRetriever({
  *   llm,
  *   retriever,
  *   rephrasePrompt,

package/dist/chains/history_aware_retriever.js

@@ -21,7 +21,7 @@ import { StringOutputParser } from "@langchain/core/output_parsers";
  * const rephrasePrompt = await pull("langchain-ai/chat-langchain-rephrase");
  * const llm = new ChatOpenAI({});
  * const retriever = ...
- * const historyAwareRetrieverChain = await createHistoryAwareRetriever({
+ * const chain = await createHistoryAwareRetriever({
  *   llm,
  *   retriever,
  *   rephrasePrompt,

package/dist/chains/openai_functions/index.d.ts

@@ -2,4 +2,4 @@ export { createExtractionChain, createExtractionChainFromZod, } from "./extracti
 export { type TaggingChainOptions, createTaggingChain, createTaggingChainFromZod, } from "./tagging.js";
 export { type OpenAPIChainOptions, createOpenAPIChain } from "./openapi.js";
 export { type StructuredOutputChainInput, createStructuredOutputChain, createStructuredOutputChainFromZod, } from "./structured_output.js";
-export { createStructuredOutputRunnable, createOpenAIFnRunnable, } from "./base.js";
+export { type CreateStructuredOutputRunnableConfig, createStructuredOutputRunnable, type CreateOpenAIFnRunnableConfig, createOpenAIFnRunnable, } from "./base.js";

package/dist/chains/sql_db/sql_db_prompt.cjs

@@ -1,6 +1,6 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
-exports.SQL_SAP_HANA_PROMPT = exports.SQL_MSSQL_PROMPT = exports.SQL_MYSQL_PROMPT = exports.SQL_SQLITE_PROMPT = exports.SQL_POSTGRES_PROMPT = exports.DEFAULT_SQL_DATABASE_PROMPT = void 0;
+exports.SQL_ORACLE_PROMPT = exports.SQL_SAP_HANA_PROMPT = exports.SQL_MSSQL_PROMPT = exports.SQL_MYSQL_PROMPT = exports.SQL_SQLITE_PROMPT = exports.SQL_POSTGRES_PROMPT = exports.DEFAULT_SQL_DATABASE_PROMPT = void 0;
 /* eslint-disable spaced-comment */
 const prompts_1 = require("@langchain/core/prompts");
 exports.DEFAULT_SQL_DATABASE_PROMPT = new prompts_1.PromptTemplate({
@@ -120,3 +120,22 @@ Only use the following tables:
 Question: {input}`,
     inputVariables: ["dialect", "table_info", "input", "top_k"],
 });
+exports.SQL_ORACLE_PROMPT = new prompts_1.PromptTemplate({
+    template: `You are a ORACLE expert. Given an input question, first create a syntactically correct ORACLE query to run, then look at the results of the query and return the answer to the input question.
+Unless the user specifies in the question a specific number of examples to obtain, query for at most {top_k} results using the ROWNUM clause as per ORACLE. You can order the results to return the most informative data in the database.
+Never query for all columns from a table. You must query only the columns that are needed to answer the question.
+Pay attention to use only the column names you can see in the tables below. Be careful to not query for columns that do not exist. Also, pay attention to which column is in which table.
+
+Use the following format:
+
+Question: "Question here"
+SQLQuery: "SQL Query to run"
+SQLResult: "Result of the SQLQuery"
+Answer: "Final answer here"
+
+Only use the following tables:
+{table_info}
+
+Question: {input}`,
+    inputVariables: ["dialect", "table_info", "input", "top_k"],
+});

package/dist/chains/sql_db/sql_db_prompt.d.ts

@@ -35,3 +35,9 @@ export declare const SQL_SAP_HANA_PROMPT: PromptTemplate<{
     dialect: any;
     table_info: any;
 }, any>;
+export declare const SQL_ORACLE_PROMPT: PromptTemplate<{
+    input: any;
+    top_k: any;
+    dialect: any;
+    table_info: any;
+}, any>;

package/dist/chains/sql_db/sql_db_prompt.js

@@ -117,3 +117,22 @@ Only use the following tables:
 Question: {input}`,
     inputVariables: ["dialect", "table_info", "input", "top_k"],
 });
+export const SQL_ORACLE_PROMPT = /*#__PURE__*/ new PromptTemplate({
+    template: `You are a ORACLE expert. Given an input question, first create a syntactically correct ORACLE query to run, then look at the results of the query and return the answer to the input question.
+Unless the user specifies in the question a specific number of examples to obtain, query for at most {top_k} results using the ROWNUM clause as per ORACLE. You can order the results to return the most informative data in the database.
+Never query for all columns from a table. You must query only the columns that are needed to answer the question.
+Pay attention to use only the column names you can see in the tables below. Be careful to not query for columns that do not exist. Also, pay attention to which column is in which table.
+
+Use the following format:
+
+Question: "Question here"
+SQLQuery: "SQL Query to run"
+SQLResult: "Result of the SQLQuery"
+Answer: "Final answer here"
+
+Only use the following tables:
+{table_info}
+
+Question: {input}`,
+    inputVariables: ["dialect", "table_info", "input", "top_k"],
+});

package/dist/chat_models/cloudflare_workersai.cjs

@@ -17,5 +17,7 @@ Object.defineProperty(exports, "__esModule", { value: true });
 const entrypoint_deprecation_js_1 = require("../util/entrypoint_deprecation.cjs");
 /* #__PURE__ */ (0, entrypoint_deprecation_js_1.logVersion010MigrationWarning)({
     oldEntrypointName: "chat_models/cloudflare_workersai",
+    newEntrypointName: "",
+    newPackageName: "@langchain/cloudflare",
 });
 __exportStar(require("@langchain/community/chat_models/cloudflare_workersai"), exports);

package/dist/chat_models/cloudflare_workersai.js

@@ -1,5 +1,7 @@
 import { logVersion010MigrationWarning } from "../util/entrypoint_deprecation.js";
 /* #__PURE__ */ logVersion010MigrationWarning({
     oldEntrypointName: "chat_models/cloudflare_workersai",
+    newEntrypointName: "",
+    newPackageName: "@langchain/cloudflare",
 });
 export * from "@langchain/community/chat_models/cloudflare_workersai";

package/dist/embeddings/cloudflare_workersai.cjs

@@ -17,5 +17,7 @@ Object.defineProperty(exports, "__esModule", { value: true });
 const entrypoint_deprecation_js_1 = require("../util/entrypoint_deprecation.cjs");
 /* #__PURE__ */ (0, entrypoint_deprecation_js_1.logVersion010MigrationWarning)({
     oldEntrypointName: "embeddings/cloudflare_workersai",
+    newEntrypointName: "",
+    newPackageName: "@langchain/cloudflare",
 });
 __exportStar(require("@langchain/community/embeddings/cloudflare_workersai"), exports);

package/dist/embeddings/cloudflare_workersai.js

@@ -1,5 +1,7 @@
 import { logVersion010MigrationWarning } from "../util/entrypoint_deprecation.js";
 /* #__PURE__ */ logVersion010MigrationWarning({
     oldEntrypointName: "embeddings/cloudflare_workersai",
+    newEntrypointName: "",
+    newPackageName: "@langchain/cloudflare",
 });
 export * from "@langchain/community/embeddings/cloudflare_workersai";

package/dist/evaluation/agents/trajectory.d.ts

@@ -6,7 +6,7 @@ import { ChatGeneration, Generation } from "@langchain/core/outputs";
 import { BasePromptTemplate } from "@langchain/core/prompts";
 import { Callbacks, BaseCallbackConfig } from "@langchain/core/callbacks/manager";
 import { BaseChatModel } from "@langchain/core/language_models/chat_models";
-import { AgentTrajectoryEvaluator, EvalOutputType, LLMEvalChainInput, LLMTrajectoryEvaluatorArgs } from "../base.js";
+import { AgentTrajectoryEvaluator, EvalOutputType, LLMEvalChainInput, LLMTrajectoryEvaluatorArgs, type ExtractLLMCallOptions } from "../base.js";
 /**
  * A parser for the output of the TrajectoryEvalChain.
  */
@@ -51,5 +51,5 @@ export declare class TrajectoryEvalChain extends AgentTrajectoryEvaluator {
      */
     getAgentTrajectory(steps: AgentStep[]): string;
     formatReference(reference?: string): string;
-    _evaluateAgentTrajectory(args: LLMTrajectoryEvaluatorArgs, callOptions: this["llm"]["CallOptions"], config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
+    _evaluateAgentTrajectory(args: LLMTrajectoryEvaluatorArgs, callOptions: ExtractLLMCallOptions<this["llm"]>, config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
 }

package/dist/evaluation/base.d.ts

@@ -8,6 +8,7 @@ import { BaseChain, LLMChain, LLMChainInput } from "../chains/index.js";
  */
 export interface LLMEvalChainInput<T extends EvalOutputType = EvalOutputType, L extends BaseLanguageModelInterface = BaseLanguageModelInterface> extends LLMChainInput<T, L> {
 }
+export type ExtractLLMCallOptions<LanguageModelInterface> = LanguageModelInterface extends BaseLanguageModelInterface<any, infer CallOptions> ? CallOptions : never;
 /**
  * Compare two sets for equality
  *
@@ -113,7 +114,7 @@ export declare abstract class LLMStringEvaluator<T extends EvalOutputType = Eval
      * @param callOptions
      * @param config
      */
-    abstract _evaluateStrings(args: StringEvaluatorArgs, callOptions?: this["llm"]["CallOptions"], config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
+    abstract _evaluateStrings(args: StringEvaluatorArgs, callOptions?: ExtractLLMCallOptions<this["llm"]>, config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
     /**
      * Evaluate Chain or LLM output, based on optional input and label.
      * @returns The evaluation results containing the score or value. It is recommended that the dictionary contain the following keys:
@@ -124,7 +125,7 @@ export declare abstract class LLMStringEvaluator<T extends EvalOutputType = Eval
      * @param callOptions
      * @param config
      */
-    evaluateStrings(args: StringEvaluatorArgs, callOptions?: this["llm"]["CallOptions"], config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
+    evaluateStrings(args: StringEvaluatorArgs, callOptions?: ExtractLLMCallOptions<this["llm"]>, config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
 }
 /**
  * Grade, tag, or otherwise evaluate predictions relative to their inputs
@@ -194,7 +195,7 @@ export declare abstract class LLMPairwiseStringEvaluator extends LLMEvalChain {
      * @param config
      * @return A dictionary containing the preference, scores, and/or other information.
      */
-    abstract _evaluateStringPairs(args: LLMPairwiseStringEvaluatorArgs, callOptions?: this["llm"]["CallOptions"], config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
+    abstract _evaluateStringPairs(args: LLMPairwiseStringEvaluatorArgs, callOptions?: ExtractLLMCallOptions<this["llm"]>, config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
     /**
      * Evaluate the output string pairs.
      * @param args
@@ -202,7 +203,7 @@ export declare abstract class LLMPairwiseStringEvaluator extends LLMEvalChain {
      * @param config
      * @return A dictionary containing the preference, scores, and/or other information.
      */
-    evaluateStringPairs(args: LLMPairwiseStringEvaluatorArgs, callOptions?: this["llm"]["CallOptions"], config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
+    evaluateStringPairs(args: LLMPairwiseStringEvaluatorArgs, callOptions?: ExtractLLMCallOptions<this["llm"]>, config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
 }
 /**
  * Interface for evaluating agent trajectories.
@@ -220,7 +221,7 @@ export declare abstract class AgentTrajectoryEvaluator extends LLMEvalChain {
      * @param callOptions
      * @param config
      */
-    abstract _evaluateAgentTrajectory(args: LLMTrajectoryEvaluatorArgs, callOptions?: this["llm"]["CallOptions"], config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
+    abstract _evaluateAgentTrajectory(args: LLMTrajectoryEvaluatorArgs, callOptions?: ExtractLLMCallOptions<this["llm"]>, config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
     /**
      * Evaluate a trajectory.
      * @return The evaluation result.
@@ -228,5 +229,5 @@ export declare abstract class AgentTrajectoryEvaluator extends LLMEvalChain {
      * @param callOptions
      * @param config
      */
-    evaluateAgentTrajectory(args: LLMTrajectoryEvaluatorArgs, callOptions?: this["llm"]["CallOptions"], config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
+    evaluateAgentTrajectory(args: LLMTrajectoryEvaluatorArgs, callOptions?: ExtractLLMCallOptions<this["llm"]>, config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
 }

package/dist/evaluation/comparison/pairwise.d.ts

@@ -4,7 +4,7 @@ import { ChainValues } from "@langchain/core/utils/types";
 import { ChatGeneration, Generation } from "@langchain/core/outputs";
 import { BasePromptTemplate } from "@langchain/core/prompts";
 import { Callbacks, BaseCallbackConfig } from "@langchain/core/callbacks/manager";
-import { EvalOutputType, LLMEvalChainInput, LLMPairwiseStringEvaluator, LLMPairwiseStringEvaluatorArgs } from "../base.js";
+import { EvalOutputType, LLMEvalChainInput, LLMPairwiseStringEvaluator, LLMPairwiseStringEvaluatorArgs, type ExtractLLMCallOptions } from "../base.js";
 import { CriteriaLike } from "../criteria/criteria.js";
 /**
  * A parser for the output of the PairwiseStringEvalChain.
@@ -36,7 +36,7 @@ export declare class PairwiseStringEvalChain extends LLMPairwiseStringEvaluator
      */
     static fromLLM(llm: BaseLanguageModelInterface, criteria?: CriteriaLike, chainOptions?: Partial<Omit<LLMEvalChainInput, "llm">>): Promise<PairwiseStringEvalChain>;
     _prepareOutput(result: ChainValues): any;
-    _evaluateStringPairs(args: LLMPairwiseStringEvaluatorArgs, callOptions: this["llm"]["CallOptions"], config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
+    _evaluateStringPairs(args: LLMPairwiseStringEvaluatorArgs, callOptions: ExtractLLMCallOptions<this["llm"]>, config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
 }
 /**
  * A chain for comparing two outputs, such as the outputs

package/dist/evaluation/criteria/criteria.d.ts

@@ -4,7 +4,7 @@ import { ChainValues } from "@langchain/core/utils/types";
 import { ChatGeneration, Generation } from "@langchain/core/outputs";
 import { BasePromptTemplate } from "@langchain/core/prompts";
 import { Callbacks, BaseCallbackConfig } from "@langchain/core/callbacks/manager";
-import { EvalOutputType, LLMEvalChainInput, LLMStringEvaluator, StringEvaluatorArgs } from "../base.js";
+import { EvalOutputType, LLMEvalChainInput, LLMStringEvaluator, StringEvaluatorArgs, type ExtractLLMCallOptions } from "../base.js";
 import { ConstitutionalPrinciple } from "../../chains/constitutional_ai/constitutional_principle.js";
 /**
  * A Criteria to evaluate.
@@ -61,7 +61,7 @@ export declare class CriteriaEvalChain extends LLMStringEvaluator {
      * @param result
      */
     _prepareOutput(result: ChainValues): any;
-    _evaluateStrings(args: StringEvaluatorArgs, callOptions: this["llm"]["CallOptions"], config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
+    _evaluateStrings(args: StringEvaluatorArgs, callOptions: ExtractLLMCallOptions<this["llm"]>, config?: Callbacks | BaseCallbackConfig): Promise<ChainValues>;
 }
 /**
  * Criteria evaluation chain that requires references.

package/dist/experimental/chat_models/anthropic_functions.cjs

@@ -5,7 +5,7 @@ const fast_xml_parser_1 = require("fast-xml-parser");
 const messages_1 = require("@langchain/core/messages");
 const chat_models_1 = require("@langchain/core/language_models/chat_models");
 const prompts_1 = require("@langchain/core/prompts");
-const openai_1 = require("@langchain/openai");
+const function_calling_1 = require("@langchain/core/utils/function_calling");
 const anthropic_js_1 = require("../../chat_models/anthropic.cjs");
 const TOOL_SYSTEM_PROMPT = 
 /* #__PURE__ */
@@ -71,7 +71,7 @@ class AnthropicFunctions extends chat_models_1.BaseChatModel {
         let functionCall;
         if (options.tools) {
             // eslint-disable-next-line no-param-reassign
-            options.functions = (options.functions ?? []).concat(options.tools.map(openai_1.formatToOpenAIFunction));
+            options.functions = (options.functions ?? []).concat(options.tools.map(function_calling_1.convertToOpenAIFunction));
         }
         if (options.functions !== undefined && options.functions.length > 0) {
             const content = await TOOL_SYSTEM_PROMPT.format({

package/dist/experimental/chat_models/anthropic_functions.js

@@ -2,7 +2,7 @@ import { XMLParser } from "fast-xml-parser";
 import { AIMessage, SystemMessage, } from "@langchain/core/messages";
 import { BaseChatModel, } from "@langchain/core/language_models/chat_models";
 import { PromptTemplate } from "@langchain/core/prompts";
-import { formatToOpenAIFunction } from "@langchain/openai";
+import { convertToOpenAIFunction } from "@langchain/core/utils/function_calling";
 import { ChatAnthropic, DEFAULT_STOP_SEQUENCES, } from "../../chat_models/anthropic.js";
 const TOOL_SYSTEM_PROMPT = 
 /* #__PURE__ */
@@ -68,7 +68,7 @@ export class AnthropicFunctions extends BaseChatModel {
         let functionCall;
         if (options.tools) {
             // eslint-disable-next-line no-param-reassign
-            options.functions = (options.functions ?? []).concat(options.tools.map(formatToOpenAIFunction));
+            options.functions = (options.functions ?? []).concat(options.tools.map(convertToOpenAIFunction));
         }
         if (options.functions !== undefined && options.functions.length > 0) {
             const content = await TOOL_SYSTEM_PROMPT.format({

package/dist/experimental/chat_models/ollama_functions.d.ts

@@ -49,6 +49,7 @@ export declare class OllamaFunctions extends BaseChatModel<ChatOllamaFunctionsCa
             num_gpu: number | undefined;
             num_gqa: number | undefined;
             num_keep: number | undefined;
+            num_predict: number | undefined;
             num_thread: number | undefined;
             penalize_newline: boolean | undefined;
             presence_penalty: number | undefined;

package/dist/experimental/prompts/custom_format.cjs

@@ -0,0 +1,68 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.CustomFormatPromptTemplate = void 0;
+const prompts_1 = require("@langchain/core/prompts");
+class CustomFormatPromptTemplate extends prompts_1.PromptTemplate {
+    static lc_name() {
+        return "CustomPromptTemplate";
+    }
+    constructor(input) {
+        super(input);
+        Object.defineProperty(this, "lc_serializable", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: false
+        });
+        Object.defineProperty(this, "templateValidator", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: void 0
+        });
+        Object.defineProperty(this, "renderer", {
+            enumerable: true,
+            configurable: true,
+            writable: true,
+            value: void 0
+        });
+        Object.assign(this, input);
+        if (this.validateTemplate && this.templateValidator !== undefined) {
+            let totalInputVariables = this.inputVariables;
+            if (this.partialVariables) {
+                totalInputVariables = totalInputVariables.concat(Object.keys(this.partialVariables));
+            }
+            this.templateValidator(this.template, totalInputVariables);
+        }
+    }
+    /**
+     * Load prompt template from a template
+     */
+    static fromTemplate(template, { customParser, ...rest }) {
+        const names = new Set();
+        const nodes = customParser(template);
+        for (const node of nodes) {
+            if (node.type === "variable") {
+                names.add(node.name);
+            }
+        }
+        // eslint-disable-next-line @typescript-eslint/ban-types
+        return new this({
+            // eslint-disable-next-line @typescript-eslint/no-explicit-any
+            inputVariables: [...names],
+            template,
+            customParser,
+            ...rest,
+        });
+    }
+    /**
+     * Formats the prompt template with the provided values.
+     * @param values The values to be used to format the prompt template.
+     * @returns A promise that resolves to a string which is the formatted prompt.
+     */
+    async format(values) {
+        const allValues = await this.mergePartialAndUserVariables(values);
+        return this.renderer(this.template, allValues);
+    }
+}
+exports.CustomFormatPromptTemplate = CustomFormatPromptTemplate;

package/dist/experimental/prompts/custom_format.d.ts

@@ -0,0 +1,24 @@
+import type { InputValues } from "@langchain/core/utils/types";
+import { type ParsedFStringNode, PromptTemplate, type PromptTemplateInput, TypedPromptInputValues } from "@langchain/core/prompts";
+export type CustomFormatPromptTemplateInput<RunInput extends InputValues> = Omit<PromptTemplateInput<RunInput, string>, "templateFormat"> & {
+    customParser: (template: string) => ParsedFStringNode[];
+    templateValidator?: (template: string, inputVariables: string[]) => boolean;
+    renderer: (template: string, values: InputValues) => string;
+};
+export declare class CustomFormatPromptTemplate<RunInput extends InputValues = any, PartialVariableName extends string = any> extends PromptTemplate<RunInput, PartialVariableName> {
+    static lc_name(): string;
+    lc_serializable: boolean;
+    templateValidator?: (template: string, inputVariables: string[]) => boolean;
+    renderer: (template: string, values: InputValues) => string;
+    constructor(input: CustomFormatPromptTemplateInput<RunInput>);
+    /**
+     * Load prompt template from a template
+     */
+    static fromTemplate<RunInput extends InputValues = Record<string, any>>(template: string, { customParser, ...rest }: Omit<CustomFormatPromptTemplateInput<RunInput>, "template" | "inputVariables">): CustomFormatPromptTemplate<RunInput extends Symbol ? never : RunInput, any>;
+    /**
+     * Formats the prompt template with the provided values.
+     * @param values The values to be used to format the prompt template.
+     * @returns A promise that resolves to a string which is the formatted prompt.
+     */
+    format(values: TypedPromptInputValues<RunInput>): Promise<string>;
+}