use-chat-setting-schema.tsx

Overview

The use-chat-setting-schema.tsx file defines a custom React hook useChatSettingSchema that provides a comprehensive Zod validation schema for chat settings configuration in a web application. The schema encapsulates various chat-related settings including prompt configurations, knowledge base IDs, LLM (Large Language Model) parameters, reranking options, vector similarity weights, and metadata filters.

This schema is primarily used for form validation purposes, ensuring that user input for chat assistant settings meets the expected structure and constraints. It also integrates translation support for validation messages, adapting error messages dynamically based on localization.

Detailed Explanation

`useChatSettingSchema()`

Purpose

To create and return a Zod schema object representing the entire form structure for configuring chat assistant settings. This hook leverages other imported schemas and adds localization to validation messages.

Returns

ZodObject: A Zod schema object that validates chat settings form data.

Usage Example

import { useChatSettingSchema } from './use-chat-setting-schema';

function ChatSettingsForm() {
  const chatSettingSchema = useChatSettingSchema();

  // Example form data that needs validation
  const formData = {
    name: "My Assistant",
    icon: "robot",
    kb_ids: ["kb1", "kb2"],
    prompt_config: {
      quote: true,
      keyword: false,
      tts: true,
      system: "System prompt message",
      refine_multiturn: false,
      use_kg: true,
      parameters: [{ key: "param1", optional: false }],
      cross_languages: ["en", "fr"]
    },
    // other fields as per the schema...
  };

  try {
    chatSettingSchema.parse(formData);
    // Valid form data, proceed with submission
  } catch (e) {
    // Handle validation errors
  }

  return <div>/* form UI here */</div>;
}

Schema Structure

1. `promptConfigSchema`

Defines the structure of the prompt_config object within the form, which configures prompt-related settings for the chat assistant.

Property	Type	Description	Validation/Notes
`quote`	`boolean`	Whether quoting is enabled in prompts	Required
`keyword`	`boolean`	Whether keyword extraction is enabled	Required
`tts`	`boolean`	Whether text-to-speech is enabled	Required
`empty_response`	`string` (optional)	Optional string to use when response is empty	Optional
`prologue`	`string` (optional)	Optional prologue text prepended to prompts	Optional
`system`	`string`	System message text	Required, minimum 1 character; error message localized
`refine_multiturn`	`boolean`	Whether to refine multi-turn conversations	Required
`use_kg`	`boolean`	Whether to use knowledge graph	Required
`parameters`	`Array` of objects	Array of parameter objects with `key` and `optional` flag	Required; each object must have a string `key` and a boolean `optional`
`tavily_api_key`	`string` (optional)	Optional API key	Optional
`reasoning`	`boolean` (optional)	Optional flag for enabling reasoning	Optional
`cross_languages`	`Array<string>` (optional)	Optional array of language codes	Optional

2. `formSchema`

Top-level schema representing the entire chat assistant form.

Property	Type	Description	Validation/Notes
`name`	`string`	Name of the assistant	Required, min 1 character, localized error message
`icon`	`string`	Icon identifier for the assistant	Required
`description`	`string` (optional)	Optional description of the assistant	Optional
`kb_ids`	`Array<string>`	Array of knowledge base IDs associated with the assistant	Required, minimum length 0 (can be empty), localized message
`prompt_config`	`promptConfigSchema`	Nested prompt configuration object	Required
`rerankFormSchema`	Imported schema spread	Reranking related form schema	Spread into form schema
`llm_setting`	Object with `LlmSettingFieldSchema`	LLM (Large Language Model) specific settings	Required
`LlmSettingEnabledSchema`	Imported schema spread	Schema for enabling/disabling LLM settings	Spread into form schema
`llm_id`	`string` (optional)	Optional LLM identifier	Optional
`vectorSimilarityWeightSchema`	Imported schema spread	Schema for vector similarity weighting	Spread into form schema
`similarityThresholdSchema`	Imported schema spread	Schema for similarity threshold	Spread into form schema
`topnSchema`	Imported schema spread	Schema for top-N selection parameters	Spread into form schema
`MetadataFilterSchema`	Imported schema spread	Schema for metadata filtering	Spread into form schema

Important Implementation Details

Localization:
The hook uses useTranslate('chat') to get the translation function t scoped to the chat namespace. This function is used to localize validation error messages for fields like system and name.
Schema Composition:
The form schema is composed by merging multiple imported schemas (like rerankFormSchema, LlmSettingEnabledSchema, etc.) with locally defined schemas (promptConfigSchema). This modular approach allows for flexible, reusable validation logic.
Zod Library:
The schema uses the zod library for defining and enforcing type-safe validation rules. This ensures form data integrity before submission or processing.

Interaction with Other Parts of the System

Imported Schemas:
The file imports multiple schemas from other components:
- LlmSettingEnabledSchema, LlmSettingFieldSchema — from LLM settings components.
- MetadataFilterSchema — from metadata filter components.
- rerankFormSchema — from rerank feature components.
- similarityThresholdSchema, vectorSimilarityWeightSchema — from similarity slider components.
- topnSchema — from top N item selection components.
These imports indicate that the chat setting schema aggregates validation rules from various specialized modules, ensuring a unified form validation.
Translation Hook:
The hook useTranslate is imported from common hooks, enabling multilingual support for validation messages.
Usage Context:
This schema is likely used in chat assistant creation or configuration forms within the larger application, ensuring that the settings entered by users are valid and complete before being saved or used.

Visual Diagram

classDiagram
    class useChatSettingSchema {
        +t: function
        +promptConfigSchema: object
        +formSchema: object
        +useChatSettingSchema(): ZodObject
    }
    class promptConfigSchema {
        +quote: boolean
        +keyword: boolean
        +tts: boolean
        +empty_response?: string
        +prologue?: string
        +system: string
        +refine_multiturn: boolean
        +use_kg: boolean
        +parameters: Array<{key: string, optional: boolean}>
        +tavily_api_key?: string
        +reasoning?: boolean
        +cross_languages?: Array<string>
    }
    class formSchema {
        +name: string
        +icon: string
        +description?: string
        +kb_ids: Array<string>
        +prompt_config: promptConfigSchema
        +rerankFormSchema: object (spread)
        +llm_setting: object (LlmSettingFieldSchema)
        +LlmSettingEnabledSchema: object (spread)
        +llm_id?: string
        +vectorSimilarityWeightSchema: object (spread)
        +similarityThresholdSchema: object (spread)
        +topnSchema: object (spread)
        +MetadataFilterSchema: object (spread)
    }

    useChatSettingSchema --> promptConfigSchema : defines
    useChatSettingSchema --> formSchema : defines

Summary

use-chat-setting-schema.tsx provides a modular, extensible, and localized Zod schema hook to validate chat assistant settings form data. It integrates multiple feature-specific schemas, supports internationalization, and enforces strict validation on critical fields like system prompts, assistant name, and knowledge base associations. This hook is a key utility in ensuring robust and consistent configuration data for chat assistants within the application.