@arizeai/phoenix-mcp

import { getTemplateVariables } from "../template"; import { ClassificationChoicesMap, CreateClassificationEvaluatorArgs, EvaluatorFn, PromptTemplate, WithPromptTemplate, } from "../types"; import { ObjectMapping } from "../types/data"; import { remapObject } from "../utils/objectMappingUtils"; import { createClassifierFn } from "./createClassifierFn"; import { LLMEvaluator } from "./LLMEvaluator"; import { LanguageModel } from "ai"; /** * An LLM evaluator that performs evaluation via classification */ export class ClassificationEvaluator<RecordType extends Record<string, unknown>> extends LLMEvaluator<RecordType> implements WithPromptTemplate { readonly evaluatorFn: EvaluatorFn<RecordType>; readonly promptTemplate: PromptTemplate; /** * A dynamically computed set of prompt template variables */ private _promptTemplateVariables: string[] | undefined; /** * The model to use for classification */ readonly model: LanguageModel; /** * The choices to classify the example into */ readonly choices: ClassificationChoicesMap; constructor(args: CreateClassificationEvaluatorArgs<RecordType>) { super(args); this.promptTemplate = args.promptTemplate; this.model = args.model; this.choices = args.choices; this.evaluatorFn = createClassifierFn<RecordType>({ ...args, }); } evaluate = (example: RecordType) => { return this.evaluatorFn( this.inputMapping ? remapObject<RecordType>(example, this.inputMapping) : example ); }; /** * List out the prompt template variables needed to perform evaluation */ get promptTemplateVariables(): string[] { // Use dynamic programming to see if it's computed already if (!Array.isArray(this._promptTemplateVariables)) { this._promptTemplateVariables = getTemplateVariables({ template: this.promptTemplate, }); } // Give a copy of the variables return [...this._promptTemplateVariables]; } /** * Binds the input mapping to the evaluator. It makes a a copy of the evaluator and returns it. */ bindInputMapping( inputMapping: ObjectMapping<RecordType> ): ClassificationEvaluator<RecordType> { return new ClassificationEvaluator({ ...this, inputMapping, }); } }