Add {{maxContext}}, {{maxResponse}} macros and {{maxPromptTokens}} alias (#5176) * Initial plan * Add maxContext, maxResponse macros and maxPromptTokens alias Co-authored-by: Cohee1207 <18619528+Cohee1207@users.noreply.github.com> * Refactor getMaxContextSize to use getMaxContextTokens/getMaxResponseTokens and remove maxReply aliases Co-authored-by: Cohee1207 <18619528+Cohee1207@users.noreply.github.com> * Align aliases in single line * Rename getMaxPromptTokens --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: Cohee1207 <18619528+Cohee1207@users.noreply.github.com>

eeda4d377ee2d4fb21f536ae793cda6d1673b449

Copilot <198982749+Copilot@users.noreply.github.com>

Signed

Parents: 3bd10346394c4ef29a0a550efc1f22b0e66e07a4

public/script.js+37 -15 Viewed

     setCharacterSettingsOverrides as setScenarioOverride,
     /** @deprecated Use appendMediaToMessage instead. */
     appendMediaToMessage as appendImageToMessage,
 +    /** @deprecated Use getMaxPromptTokens instead. */
 +    getMaxPromptTokens as getMaxContextSize,
 };
 /**

+    }
     // Determine token limit
     let this_max_context = getMaxContextSizegetMaxPromptTokens();
     if (!dryRun) {
         console.debug('Running extension interceptors');

+}
 /**
  * Gets the maximum usablecontext token limit (the full context window size forbefore thesubtracting currentresponse APIlength).
  * @paramreturns {number|null} overrideResponseLengthThe Optionalmaximum overridecontext token limit for the responsecurrent lengthAPI.
 - * @returns {number} Maximum usable context size.
  */
 export function getMaxContextSizegetMaxContextTokens(overrideResponseLength = null) {
 -    if (typeof overrideResponseLength !== 'number' || overrideResponseLength <= 0 || isNaN(overrideResponseLength)) {
 -        overrideResponseLength = null;
 -    }
+-
 -    let this_max_context = 1487;
     if (main_api == 'kobold' || main_api == 'koboldhorde' || main_api == 'textgenerationwebui') {
 -        this_max_context = (max_context - (overrideResponseLength || amount_gen));
 +        return max_context;
+    }
     if (main_api == 'novel') {
         let this_max_context = Number(max_context);
         if (nai_settings.model_novel.includes('clio')) {
             this_max_context = Math.min(max_context, 8192);
+        }

             // Added special tokens and whatnot
             this_max_context -= 10;
+        }
 +        return this_max_context;
 +    }
 +    if (main_api == 'openai') {
 +        return oai_settings.openai_max_context;
 +    }
 +    return 1487;
 +}
 -        this_max_context = this_max_context - (overrideResponseLength || amount_gen);
 +/**
 + * Gets the maximum response token limit (the max generation/reply length).
 + * @returns {number} The maximum response token limit for the current API.
 + */
 +export function getMaxResponseTokens() {
 +    if (main_api == 'kobold' || main_api == 'koboldhorde' || main_api == 'textgenerationwebui' || main_api == 'novel') {
 +        return amount_gen;
+    }
     if (main_api == 'openai') {
 -        this_max_context = oai_settings.openai_max_context - (overrideResponseLength || oai_settings.openai_max_tokens);
 +        return oai_settings.openai_max_tokens;
 +    }
 +    return 0;
 +}
++
 +/**
 + * Gets the maximum usable prompt size for the current API.
 + * @param {number|null} overrideResponseLength Optional override for the response length.
 + * @returns {number} Maximum usable prompt size.
 + */
 +export function getMaxPromptTokens(overrideResponseLength = null) {
 +    if (typeof overrideResponseLength !== 'number' || overrideResponseLength <= 0 || isNaN(overrideResponseLength)) {
 +        overrideResponseLength = null;
+    }
 -    return this_max_context;
++
 +    return getMaxContextTokens() - (overrideResponseLength || getMaxResponseTokens());
+}
 function parseTokenCounts(counts, thisPromptBits) {

public/scripts/extensions/memory/index.js+2 -2 Viewed

     saveSettingsDebounced,
     substituteParamsExtended,
     generateRaw,
     getMaxContextSizegetMaxPromptTokens,
     setExtensionPrompt,
     streamingProcessor,
     animation_easing,

@@ -71,7 +71,7 @@ async function getSourceContextSize() {
71	71	return 1024 - 64;
72	72	}
73	73
74	74	return getMaxContextSizegetMaxPromptTokens(overrideLength);
75	75	}
76	76
77	77	const formatMemoryValue = function (value) {

public/scripts/macros.js+7 -2 Viewed

 import { Handlebars, moment, seedrandom, droll } from '../lib.js';
 import { chat, chat_metadata, main_api, getMaxContextSizegetMaxPromptTokens, getMaxContextTokens, getMaxResponseTokens, getCurrentChatId, substituteParams, eventSource, event_types, extension_prompts } from '../script.js';
 import { timestampToMoment, isDigitsOnly, getStringHash, escapeRegex, uuidv4 } from './utils.js';
 import { textgenerationwebui_banned_in_macros } from './textgen-settings.js';
 import { getInstructMacros } from './instruct-mode.js';

      * @type {Macro[]}
     */
     const postEnvMacros = [
         { regex: /{{maxPrompt}}/gi, replace: () => String(getMaxContextSizegetMaxPromptTokens()) },
 +        { regex: /{{maxPromptTokens}}/gi, replace: () => String(getMaxPromptTokens()) },
 +        { regex: /{{maxContext}}/gi, replace: () => String(getMaxContextTokens()) },
 +        { regex: /{{maxContextTokens}}/gi, replace: () => String(getMaxContextTokens()) },
 +        { regex: /{{maxResponse}}/gi, replace: () => String(getMaxResponseTokens()) },
 +        { regex: /{{maxResponseTokens}}/gi, replace: () => String(getMaxResponseTokens()) },
         { regex: /{{lastMessage}}/gi, replace: () => getLastMessage() },
         { regex: /{{lastMessageId}}/gi, replace: () => String(getLastMessageId() ?? '') },
         { regex: /{{lastUserMessage}}/gi, replace: () => getLastUserMessage() },

public/scripts/macros/definitions/core-macros.js+24 -3 Viewed

 import { seedrandom, droll } from '../../../lib.js';
 import { chat_metadata, main_api, getMaxContextSizegetMaxPromptTokens, getMaxContextTokens, getMaxResponseTokens, extension_prompts, getCurrentChatId } from '../../../script.js';
 import { getStringHash, isFalseBoolean } from '../../utils.js';
 import { textgenerationwebui_banned_in_macros } from '../../textgen-settings.js';
 import { inject_ids } from '../../constants.js';

         handler: () => (/** @type {HTMLTextAreaElement} */(document.querySelector('#send_textarea')))?.value ?? '',
     });
     // {{maxPrompt}} -> max context size (context minus response)
     MacroRegistry.registerMacro('maxPrompt', {
 +        aliases: [{ alias: 'maxPromptTokens', visible: true }],
         category: MacroCategory.STATE,
         description: 'Maximum prompt context size.',
         returns: 'Maximum prompt context size.',
         returnType: MacroValueType.INTEGER,
         handler: () => String(getMaxContextSizegetMaxPromptTokens()),
 +    });
++
 +    // {{maxContext}} -> max context token limit
 +    MacroRegistry.registerMacro('maxContext', {
 +        aliases: [{ alias: 'maxContextTokens', visible: true }],
 +        category: MacroCategory.STATE,
 +        description: 'Maximum context token limit.',
 +        returns: 'Maximum context token limit.',
 +        returnType: MacroValueType.INTEGER,
 +        handler: () => String(getMaxContextTokens()),
 +    });
++
 +    // {{maxResponse}} -> max response token limit
 +    MacroRegistry.registerMacro('maxResponse', {
 +        aliases: [{ alias: 'maxResponseTokens', visible: true }],
 +        category: MacroCategory.STATE,
 +        description: 'Maximum response token limit.',
 +        returns: 'Maximum response token limit.',
 +        returnType: MacroValueType.INTEGER,
 +        handler: () => String(getMaxResponseTokens()),
     });
     // String utilities