Merge branch 'staging' into tooltips-vol1

public/index.html

@@ -3448,6 +3448,10 @@
                                         <span class="fa-solid fa-circle-question note-link-span"></span>
                                     </a>
                                 </label>
+                                <label class="checkbox_label" for="forbid_external_images" title="Disalow embedded images from other domains in chat messages.">
+                                    <input id="forbid_external_images" type="checkbox" />
+                                    <span data-i18n="Forbid External Images">Forbid External Images</span>
+                                </label>
                                 <label data-newbie-hidden class="checkbox_label" for="allow_name2_display">
                                     <input id="allow_name2_display" type="checkbox" />
                                     <span data-i18n="Allow {{char}}: in bot messages">Show {{char}}: in responses</span>
@@ -4388,7 +4392,7 @@
                                 </div>
                                 <div class="world_entry_form_control">
                                     <small class="textAlignCenter">Logic</small>
-                                    <select name="entryLogicType" class="widthFitContent margin0">
+                                    <select name="entryLogicType" class="text_pole widthFitContent margin0">
                                         <option value="0">AND ANY</option>
                                         <option value="3">AND ALL</option>
                                         <option value="1">NOT ALL</option>
@@ -4407,6 +4411,28 @@
                                     </div>
                                 </div>
                             </div>
+                            <div name="perEntryOverridesBlock" class="flex-container wide100p alignitemscenter">
+                                <div class="world_entry_form_control flex1">
+                                    <small class="textAlignCenter">Scan Depth</small>
+                                    <input class="text_pole" name="scanDepth" type="number" placeholder="Use global setting" max="100">
+                                </div>
+                                <div class="world_entry_form_control flex1">
+                                    <small class="textAlignCenter">Case-Sensitive</small>
+                                    <select name="caseSensitive" class="text_pole widthNatural margin0">
+                                        <option value="null">Use global setting</option>
+                                        <option value="true">Yes</option>
+                                        <option value="false">No</option>
+                                    </select>
+                                </div>
+                                <div class="world_entry_form_control flex1">
+                                    <small class="textAlignCenter">Match Whole Words</small>
+                                    <select name="matchWholeWords" class="text_pole widthNatural margin0">
+                                        <option value="null">Use global setting</option>
+                                        <option value="true">Yes</option>
+                                        <option value="false">No</option>
+                                    </select>
+                                </div>
+                            </div>
                             <div name="contentAndCharFilterBlock" class="world_entry_thin_controls flex2">
                                 <div class="world_entry_form_control flex1">
                                     <label for="content ">

public/script.js

@@ -296,6 +296,25 @@ DOMPurify.addHook('uponSanitizeAttribute', (_, data, config) => {
     }
 });
 
+DOMPurify.addHook('uponSanitizeElement', (node, _, config) => {
+    if (!config['MESSAGE_SANITIZE']) {
+        return;
+    }
+
+    switch (node.tagName) {
+        case 'IMG': {
+            const isExternalUrl = (url) => (url.indexOf('://') > 0 || url.indexOf('//') === 0) && !url.startsWith(window.location.origin);
+            const src = node.getAttribute('src');
+
+            if (power_user.forbid_external_images && isExternalUrl(src)) {
+                console.warn('External image blocked', src);
+                node.remove();
+            }
+        }
+        break;
+    }
+});
+
 // API OBJECT FOR EXTERNAL WIRING
 window['SillyTavern'] = {};
 
@@ -1447,6 +1466,7 @@ async function printMessages() {
 }
 
 async function clearChat() {
+    closeMessageEditor();
     count_view_mes = 0;
     extension_prompts = {};
     if (is_delete_mode) {

public/scripts/extensions/regex/editor.html

@@ -76,7 +76,7 @@
         </div>
 
         <div class="flex-container">
-            <div class="wi-enter-footer-text flex-container flexFlowColumn flexNoGap alignitemsstart">
+            <div class="flex1 wi-enter-footer-text flex-container flexFlowColumn flexNoGap alignitemsstart">
                 <small>Affects</small>
                 <div>
                     <label class="checkbox flex-container">
@@ -97,7 +97,7 @@
                     </label>
                 </div>
             </div>
-            <div class="wi-enter-footer-text flex-container flexFlowColumn flexNoGap alignitemsstart">
+            <div class="flex1 wi-enter-footer-text flex-container flexFlowColumn flexNoGap alignitemsstart">
                 <small>Other Options</small>
                 <label class="checkbox flex-container">
                     <input type="checkbox" name="disabled" />
@@ -120,13 +120,6 @@
                     <span data-i18n="Substitute Regex">Substitute Regex (?)</span>
                 </label>
             </div>
-            <div class="flex-container flexFlowColumn alignitemsstart">
-                <small>Replacement Strategy</small>
-                <select name="replace_strategy_select" class="margin0">
-                    <option value="0">Replace</option>
-                    <option value="1">Overlay (currently broken)</option>
-                </select>
-            </div>
         </div>
     </div>
 </div>

public/scripts/extensions/regex/engine.js

@@ -19,14 +19,6 @@ const regex_placement = {
     SLASH_COMMAND: 3,
 };
 
-/**
- * @enum {number} How the regex script should replace the matched string
- */
-const regex_replace_strategy = {
-    REPLACE: 0,
-    OVERLAY: 1,
-};
-
 /**
  * Instantiates a regular expression from a string.
  * @param {string} input The input string.
@@ -153,86 +145,3 @@ function filterString(rawString, trimStrings, { characterOverride } = {}) {
 
     return finalString;
 }
-
-/**
- * Substitutes regex-specific and normal parameters
- * @param {string} rawString
- * @param {string} regexMatch
- * @param {RegexSubstituteParams} params The parameters to use for the regex substitution
- * @returns {string} The substituted string
- * @typedef {{characterOverride?: string, replaceStrategy?: number}} RegexSubstituteParams The parameters to use for the regex substitution
- */
-function substituteRegexParams(rawString, regexMatch, { characterOverride, replaceStrategy } = {}) {
-    let finalString = rawString;
-    finalString = substituteParams(finalString, undefined, characterOverride);
-
-    let overlaidMatch = regexMatch;
-    // TODO: Maybe move the for loops into a separate function?
-    if (replaceStrategy === regex_replace_strategy.OVERLAY) {
-        const splitReplace = finalString.split('{{match}}');
-
-        // There's a prefix
-        if (splitReplace[0]) {
-            // Fetch the prefix
-            const splicedPrefix = spliceSymbols(splitReplace[0], false);
-
-            // Sequentially remove all occurrences of prefix from start of split
-            const splitMatch = overlaidMatch.split(splicedPrefix);
-            let sliceNum = 0;
-            for (let index = 0; index < splitMatch.length; index++) {
-                if (splitMatch[index].length === 0) {
-                    sliceNum++;
-                } else {
-                    break;
-                }
-            }
-
-            overlaidMatch = splitMatch.slice(sliceNum, splitMatch.length).join(splicedPrefix);
-        }
-
-        // There's a suffix
-        if (splitReplace[1]) {
-            // Fetch the suffix
-            const splicedSuffix = spliceSymbols(splitReplace[1], true);
-
-            // Sequential removal of all suffix occurrences from end of split
-            const splitMatch = overlaidMatch.split(splicedSuffix);
-            let sliceNum = 0;
-            for (let index = splitMatch.length - 1; index >= 0; index--) {
-                if (splitMatch[index].length === 0) {
-                    sliceNum++;
-                } else {
-                    break;
-                }
-            }
-
-            overlaidMatch = splitMatch.slice(0, splitMatch.length - sliceNum).join(splicedSuffix);
-        }
-    }
-
-    // Only one match is replaced. This is by design
-    finalString = finalString.replace('{{match}}', overlaidMatch) || finalString.replace('{{match}}', regexMatch);
-
-    return finalString;
-}
-
-/**
- * Splices common sentence symbols and whitespace from the beginning and end of a string.
- * Using a for loop due to sequential ordering.
- * @param {string} rawString The raw string to splice
- * @param {boolean} isSuffix String is a suffix
- * @returns {string} The spliced string
- */
-function spliceSymbols(rawString, isSuffix) {
-    let offset = 0;
-
-    for (const ch of isSuffix ? rawString.split('').reverse() : rawString) {
-        if (ch.match(/[^\w.,?'!]/)) {
-            offset++;
-        } else {
-            break;
-        }
-    }
-
-    return isSuffix ? rawString.substring(0, rawString.length - offset) : rawString.substring(offset);
-}

public/scripts/extensions/regex/index.js

@@ -141,9 +141,6 @@ async function onRegexEditorOpenClick(existingId) {
             editorHtml
                 .find('input[name="substitute_regex"]')
                 .prop('checked', existingScript.substituteRegex ?? false);
-            editorHtml
-                .find('select[name="replace_strategy_select"]')
-                .val(existingScript.replaceStrategy ?? 0);
 
             existingScript.placement.forEach((element) => {
                 editorHtml
@@ -181,7 +178,6 @@ async function onRegexEditorOpenClick(existingId) {
             replaceString: editorHtml.find('.regex_replace_string').val(),
             trimStrings: String(editorHtml.find('.regex_trim_strings').val()).split('\n').filter((e) => e.length !== 0) || [],
             substituteRegex: editorHtml.find('input[name="substitute_regex"]').prop('checked'),
-            replaceStrategy: Number(editorHtml.find('select[name="replace_strategy_select"]').find(':selected').val()) ?? 0,
         };
         const rawTestString = String(editorHtml.find('#regex_test_input').val());
         const result = runRegexScript(testScript, rawTestString);
@@ -224,11 +220,6 @@ async function onRegexEditorOpenClick(existingId) {
                 editorHtml
                     .find('input[name="substitute_regex"]')
                     .prop('checked'),
-            replaceStrategy:
-                parseInt(editorHtml
-                    .find('select[name="replace_strategy_select"]')
-                    .find(':selected')
-                    .val()) ?? 0,
         };
 
         saveRegexScript(newRegexScript, existingScriptIndex);

public/scripts/extensions/vectors/index.js

@@ -1,5 +1,5 @@
 import { eventSource, event_types, extension_prompt_types, getCurrentChatId, getRequestHeaders, is_send_press, saveSettingsDebounced, setExtensionPrompt, substituteParams } from '../../../script.js';
-import { ModuleWorkerWrapper, extension_settings, getContext, renderExtensionTemplate } from '../../extensions.js';
+import { ModuleWorkerWrapper, extension_settings, getContext, modules, renderExtensionTemplate } from '../../extensions.js';
 import { collapseNewlines } from '../../power-user.js';
 import { SECRET_KEYS, secret_state } from '../../secrets.js';
 import { debounce, getStringHash as calculateHash, waitUntilCondition, onlyUnique, splitRecursive } from '../../utils.js';
@@ -152,8 +152,25 @@ async function synchronizeChat(batchSize = 5) {
 
         return newVectorItems.length - batchSize;
     } catch (error) {
+        /**
+         * Gets the error message for a given cause
+         * @param {string} cause Error cause key
+         * @returns {string} Error message
+         */
+        function getErrorMessage(cause) {
+            switch (cause) {
+                case 'api_key_missing':
+                    return 'API key missing. Save it in the "API Connections" panel.';
+                case 'extras_module_missing':
+                    return 'Extras API must provide an "embeddings" module.';
+                default:
+                    return 'Check server console for more details';
+            }
+        }
+
         console.error('Vectors: Failed to synchronize chat', error);
-        const message = error.cause === 'api_key_missing' ? 'API key missing. Save it in the "API Connections" panel.' : 'Check server console for more details';
+
+        const message = getErrorMessage(error.cause);
         toastr.error(message, 'Vectorization failed');
         return -1;
     } finally {
@@ -411,6 +428,18 @@ async function getSavedHashes(collectionId) {
     return hashes;
 }
 
+/**
+ * Add headers for the Extras API source.
+ * @param {object} headers Headers object
+ */
+function addExtrasHeaders(headers) {
+    console.log(`Vector source is extras, populating API URL: ${extension_settings.apiUrl}`);
+    Object.assign(headers, {
+        'X-Extras-Url': extension_settings.apiUrl,
+        'X-Extras-Key': extension_settings.apiKey,
+    });
+}
+
 /**
  * Inserts vector items into a collection
  * @param {string} collectionId - The collection to insert into
@@ -424,9 +453,18 @@ async function insertVectorItems(collectionId, items) {
         throw new Error('Vectors: API key missing', { cause: 'api_key_missing' });
     }
 
+    if (settings.source === 'extras' && !modules.includes('embeddings')) {
+        throw new Error('Vectors: Embeddings module missing', { cause: 'extras_module_missing' });
+    }
+
+    const headers = getRequestHeaders();
+    if (settings.source === 'extras') {
+        addExtrasHeaders(headers);
+    }
+
     const response = await fetch('/api/vector/insert', {
         method: 'POST',
-        headers: getRequestHeaders(),
+        headers: headers,
         body: JSON.stringify({
             collectionId: collectionId,
             items: items,
@@ -468,9 +506,14 @@ async function deleteVectorItems(collectionId, hashes) {
  * @returns {Promise<{ hashes: number[], metadata: object[]}>} - Hashes of the results
  */
 async function queryCollection(collectionId, searchText, topK) {
+    const headers = getRequestHeaders();
+    if (settings.source === 'extras') {
+        addExtrasHeaders(headers);
+    }
+
     const response = await fetch('/api/vector/query', {
         method: 'POST',
-        headers: getRequestHeaders(),
+        headers: headers,
         body: JSON.stringify({
             collectionId: collectionId,
             searchText: searchText,

public/scripts/extensions/vectors/settings.html

@@ -11,6 +11,7 @@
                 </label>
                 <select id="vectors_source" class="text_pole">
                     <option value="transformers">Local (Transformers)</option>
+                    <option value="extras">Extras</option>
                     <option value="openai">OpenAI</option>
                     <option value="palm">Google MakerSuite (PaLM)</option>
                     <option value="mistral">MistralAI</option>

public/scripts/openai.js

@@ -1160,6 +1160,7 @@ function tryParseStreamingError(response, decoded) {
         }
 
         checkQuotaError(data);
+        checkModerationError(data);
 
         if (data.error) {
             toastr.error(data.error.message || response.statusText, 'Chat Completion API');
@@ -1187,6 +1188,15 @@ function checkQuotaError(data) {
     }
 }
 
+function checkModerationError(data) {
+    const moderationError = data?.error?.message?.includes('requires moderation');
+    if (moderationError) {
+        const moderationReason = `Reasons: ${data?.error?.metadata?.reasons?.join(', ') ?? '(N/A)'}`;
+        const flaggedText = data?.error?.metadata?.flagged_input ?? '(N/A)';
+        toastr.info(flaggedText, moderationReason, { timeOut: 10000 });
+    }
+}
+
 async function sendWindowAIRequest(messages, signal, stream) {
     if (!('ai' in window)) {
         return showWindowExtensionError();
@@ -1688,6 +1698,7 @@ async function sendOpenAIRequest(type, messages, signal) {
         const data = await response.json();
 
         checkQuotaError(data);
+        checkModerationError(data);
 
         if (data.error) {
             toastr.error(data.error.message || response.statusText, 'API returned an error');

public/scripts/power-user.js

@@ -237,6 +237,7 @@ let power_user = {
     compact_input_area: true,
     auto_connect: false,
     auto_load_chat: false,
+    forbid_external_images: false,
 };
 
 let themes = [];
@@ -1529,6 +1530,7 @@ function loadPowerUserSettings(settings, data) {
     $('#reduced_motion').prop('checked', power_user.reduced_motion);
     $('#auto-connect-checkbox').prop('checked', power_user.auto_connect);
     $('#auto-load-chat-checkbox').prop('checked', power_user.auto_load_chat);
+    $('#forbid_external_images').prop('checked', power_user.forbid_external_images);
 
     for (const theme of themes) {
         const option = document.createElement('option');
@@ -3234,6 +3236,12 @@ $(document).ready(() => {
         saveSettingsDebounced();
     });
 
+    $('#forbid_external_images').on('input', function () {
+        power_user.forbid_external_images = !!$(this).prop('checked');
+        saveSettingsDebounced();
+        reloadCurrentChat();
+    });
+
     $(document).on('click', '#debug_table [data-debug-function]', function () {
         const functionId = $(this).data('debug-function');
         const functionRecord = debug_functions.find(f => f.functionId === functionId);

public/scripts/variables.js

@@ -664,7 +664,7 @@ function randValuesCallback(from, to, args) {
     if (args.round == 'floor') {
         return Math.floor(value);
     }
-    return value;
+    return String(value);
 }
 
 export function registerVariableCommands() {

public/scripts/world-info.js

@@ -70,6 +70,135 @@ const SORT_ORDER_KEY = 'world_info_sort_order';
 const METADATA_KEY = 'world_info';
 
 const DEFAULT_DEPTH = 4;
+const MAX_SCAN_DEPTH = 100;
+
+/**
+ * Represents a scanning buffer for one evaluation of World Info.
+ */
+class WorldInfoBuffer {
+    // Typedef area
+    /** @typedef {{scanDepth?: number, caseSensitive?: boolean, matchWholeWords?: boolean}} WIScanEntry The entry that triggered the scan */
+    // End typedef area
+
+    /**
+     * @type {string[]} Array of messages sorted by ascending depth
+     */
+    #depthBuffer = [];
+
+    /**
+     * @type {string[]} Array of strings added by recursive scanning
+     */
+    #recurseBuffer = [];
+
+    /**
+     * @type {number} The skew of the global scan depth. Used in "min activations"
+     */
+    #skew = 0;
+
+    /**
+     * Initialize the buffer with the given messages.
+     * @param {string[]} messages Array of messages to add to the buffer
+     */
+    constructor(messages) {
+        this.#initDepthBuffer(messages);
+    }
+
+    /**
+     * Populates the buffer with the given messages.
+     * @param {string[]} messages Array of messages to add to the buffer
+     * @returns {void} Hardly seen nothing down here
+     */
+    #initDepthBuffer(messages) {
+        for (let depth = 0; depth < MAX_SCAN_DEPTH; depth++) {
+            if (messages[depth]) {
+                this.#depthBuffer[depth] = messages[depth].trim();
+            }
+        }
+    }
+
+    /**
+     * Gets a string that respects the case sensitivity setting
+     * @param {string} str The string to transform
+     * @param {WIScanEntry} entry The entry that triggered the scan
+     * @returns {string} The transformed string
+    */
+    #transformString(str, entry) {
+        const caseSensitive = entry.caseSensitive ?? world_info_case_sensitive;
+        return caseSensitive ? str : str.toLowerCase();
+    }
+
+    /**
+     * Gets all messages up to the given depth + recursion buffer.
+     * @param {WIScanEntry} entry The entry that triggered the scan
+     * @returns {string} A slice of buffer until the given depth (inclusive)
+     */
+    get(entry) {
+        let depth = entry.scanDepth ?? (world_info_depth + this.#skew);
+
+        if (depth < 0) {
+            console.error(`Invalid WI scan depth ${depth}. Must be >= 0`);
+            return '';
+        }
+
+        if (depth > MAX_SCAN_DEPTH) {
+            console.warn(`Invalid WI scan depth ${depth}. Truncating to ${MAX_SCAN_DEPTH}`);
+            depth = MAX_SCAN_DEPTH;
+        }
+
+        let result = this.#depthBuffer.slice(0, depth).join('\n');
+
+        if (this.#recurseBuffer.length > 0) {
+            result += '\n' + this.#recurseBuffer.join('\n');
+        }
+
+        return this.#transformString(result, entry);
+    }
+
+    /**
+     * Matches the given string against the buffer.
+     * @param {string} haystack The string to search in
+     * @param {string} needle The string to search for
+     * @param {WIScanEntry} entry The entry that triggered the scan
+     * @returns {boolean} True if the string was found in the buffer
+     */
+    matchKeys(haystack, needle, entry) {
+        const transformedString = this.#transformString(needle, entry);
+        const matchWholeWords = entry.matchWholeWords ?? world_info_match_whole_words;
+
+        if (matchWholeWords) {
+            const keyWords = transformedString.split(/\s+/);
+
+            if (keyWords.length > 1) {
+                return haystack.includes(transformedString);
+            }
+            else {
+                const regex = new RegExp(`\\b${escapeRegex(transformedString)}\\b`);
+                if (regex.test(haystack)) {
+                    return true;
+                }
+            }
+        } else {
+            return haystack.includes(transformedString);
+        }
+
+        return false;
+    }
+
+    /**
+     * Adds a message to the recursion buffer.
+     * @param {string} message The message to add
+     */
+    addRecurse(message) {
+        this.#recurseBuffer.push(message);
+    }
+
+    /**
+     * Adds an increment to depth skew.
+     */
+    addSkew() {
+        this.#skew++;
+    }
+}
 
 export function getWorldInfoSettings() {
     return {
@@ -790,6 +919,9 @@ const originalDataKeyMap = {
     'key': 'keys',
     'keysecondary': 'secondary_keys',
     'selective': 'selective',
+    'matchWholeWords': 'extensions.match_whole_words',
+    'caseSensitive': 'extensions.case_sensitive',
+    'scanDepth': 'extensions.scan_depth',
 };
 
 function setOriginalDataValue(data, uid, key, value) {
@@ -1167,7 +1299,7 @@ function getWorldEntry(name, data, entry) {
     probabilityInput.data('uid', entry.uid);
     probabilityInput.on('input', function () {
         const uid = $(this).data('uid');
-        const value = parseInt($(this).val());
+        const value = Number($(this).val());
 
         data.entries[uid].probability = !isNaN(value) ? value : null;
 
@@ -1370,6 +1502,57 @@ function getWorldEntry(name, data, entry) {
         updateEditor(navigation_option.previous);
     });
 
+    // scan depth
+    const scanDepthInput = template.find('input[name="scanDepth"]');
+    scanDepthInput.data('uid', entry.uid);
+    scanDepthInput.on('input', function () {
+        const uid = $(this).data('uid');
+        const isEmpty = $(this).val() === '';
+        const value = Number($(this).val());
+
+        // Clamp if necessary
+        if (value < 0) {
+            $(this).val(0).trigger('input');
+            return;
+        }
+
+        if (value > MAX_SCAN_DEPTH) {
+            $(this).val(MAX_SCAN_DEPTH).trigger('input');
+            return;
+        }
+
+        data.entries[uid].scanDepth = !isEmpty && !isNaN(value) && value >= 0 && value < MAX_SCAN_DEPTH ? Math.floor(value) : null;
+        setOriginalDataValue(data, uid, 'extensions.scan_depth', data.entries[uid].scanDepth);
+        saveWorldInfo(name, data);
+    });
+    scanDepthInput.val(entry.scanDepth ?? null).trigger('input');
+
+    // case sensitive select
+    const caseSensitiveSelect = template.find('select[name="caseSensitive"]');
+    caseSensitiveSelect.data('uid', entry.uid);
+    caseSensitiveSelect.on('input', function () {
+        const uid = $(this).data('uid');
+        const value = $(this).val();
+
+        data.entries[uid].caseSensitive = value === 'null' ? null : value === 'true';
+        setOriginalDataValue(data, uid, 'extensions.case_sensitive', data.entries[uid].caseSensitive);
+        saveWorldInfo(name, data);
+    });
+    caseSensitiveSelect.val((entry.caseSensitive === null || entry.caseSensitive === undefined) ? 'null' : entry.caseSensitive ? 'true' : 'false').trigger('input');
+
+    // match whole words select
+    const matchWholeWordsSelect = template.find('select[name="matchWholeWords"]');
+    matchWholeWordsSelect.data('uid', entry.uid);
+    matchWholeWordsSelect.on('input', function () {
+        const uid = $(this).data('uid');
+        const value = $(this).val();
+
+        data.entries[uid].matchWholeWords = value === 'null' ? null : value === 'true';
+        setOriginalDataValue(data, uid, 'extensions.match_whole_words', data.entries[uid].matchWholeWords);
+        saveWorldInfo(name, data);
+    });
+    matchWholeWordsSelect.val((entry.matchWholeWords === null || entry.matchWholeWords === undefined) ? 'null' : entry.matchWholeWords ? 'true' : 'false').trigger('input');
+
     template.find('.inline-drawer-content').css('display', 'none'); //entries start collapsed
 
     function updatePosOrdDisplay(uid) {
@@ -1428,6 +1611,9 @@ const newEntryTemplate = {
     useProbability: true,
     depth: DEFAULT_DEPTH,
     group: '',
+    scanDepth: null,
+    caseSensitive: null,
+    matchWholeWords: null,
 };
 
 function createWorldInfoEntry(name, data, fromSlashCommand = false) {
@@ -1585,11 +1771,6 @@ async function createNewWorldInfo(worldInfoName) {
     }
 }
 
-// Gets a string that respects the case sensitivity setting
-function transformString(str) {
-    return world_info_case_sensitive ? str : str.toLowerCase();
-}
-
 async function getCharacterLore() {
     const character = characters[this_chid];
     const name = character?.name;
@@ -1711,11 +1892,10 @@ async function getSortedEntries() {
 
 async function checkWorldInfo(chat, maxContext) {
     const context = getContext();
-    const messagesToLookBack = world_info_depth * 2 || 1;
+    const buffer = new WorldInfoBuffer(chat);
 
     // Combine the chat
-    let textToScan = chat.slice(0, messagesToLookBack).join('');
+    let minActivationMsgIndex = world_info_depth; // tracks chat index to satisfy `world_info_min_activations`
-    let minActivationMsgIndex = messagesToLookBack; // tracks chat index to satisfy `world_info_min_activations`
 
     // Add the depth or AN if enabled
     // Put this code here since otherwise, the chat reference is modified
@@ -1723,14 +1903,11 @@ async function checkWorldInfo(chat, maxContext) {
         if (context.extensionPrompts[key]?.scan) {
             const prompt = getExtensionPromptByName(key);
             if (prompt) {
-                textToScan = `${prompt}\n${textToScan}`;
+                buffer.addRecurse(prompt);
             }
         }
     }
 
-    // Transform the resulting string
-    textToScan = transformString(textToScan);
-
     let needsToScan = true;
     let token_budget_overflowed = false;
     let count = 0;
@@ -1809,10 +1986,11 @@ async function checkWorldInfo(chat, maxContext) {
 
                 primary: for (let key of entry.key) {
                     const substituted = substituteParams(key);
+                    const textToScan = buffer.get(entry);
 
                     console.debug(`${entry.uid}: ${substituted}`);
 
-                    if (substituted && matchKeys(textToScan, substituted.trim())) {
+                    if (substituted && buffer.matchKeys(textToScan, substituted.trim(), entry)) {
                         console.debug(`WI UID ${entry.uid} found by primary match: ${substituted}.`);
 
                         //selective logic begins
@@ -1826,7 +2004,7 @@ async function checkWorldInfo(chat, maxContext) {
                             let hasAllMatch = true;
                             secondary: for (let keysecondary of entry.keysecondary) {
                                 const secondarySubstituted = substituteParams(keysecondary);
-                                const hasSecondaryMatch = secondarySubstituted && matchKeys(textToScan, secondarySubstituted.trim());
+                                const hasSecondaryMatch = secondarySubstituted && buffer.matchKeys(textToScan, secondarySubstituted.trim(), entry);
                                 console.debug(`WI UID:${entry.uid}: Filtering for secondary keyword - "${secondarySubstituted}".`);
 
                                 if (hasSecondaryMatch) {
@@ -1926,9 +2104,8 @@ async function checkWorldInfo(chat, maxContext) {
                 .filter(x => !failedProbabilityChecks.has(x))
                 .filter(x => !x.preventRecursion)
                 .map(x => x.content).join('\n');
-            const currentlyActivatedText = transformString(text);
+            buffer.addRecurse(text);
-            textToScan = (currentlyActivatedText + '\n' + textToScan);
+            allActivatedText = (text + '\n' + allActivatedText);
-            allActivatedText = (currentlyActivatedText + '\n' + allActivatedText);
         }
 
         // world_info_min_activations
@@ -1941,8 +2118,8 @@ async function checkWorldInfo(chat, maxContext) {
                 ) || (minActivationMsgIndex >= chat.length);
                 if (!over_max) {
                     needsToScan = true;
-                    textToScan = transformString(chat.slice(minActivationMsgIndex, minActivationMsgIndex + 1).join(''));
                     minActivationMsgIndex += 1;
+                    buffer.addSkew();
                 }
             }
         }
@@ -2069,29 +2246,6 @@ function filterByInclusionGroups(newEntries, allActivatedEntries) {
     }
 }
 
-function matchKeys(haystack, needle) {
-    const transformedString = transformString(needle);
-
-    if (world_info_match_whole_words) {
-        const keyWords = transformedString.split(/\s+/);
-
-        if (keyWords.length > 1) {
-            return haystack.includes(transformedString);
-        }
-        else {
-            const regex = new RegExp(`\\b${escapeRegex(transformedString)}\\b`);
-            if (regex.test(haystack)) {
-                return true;
-            }
-        }
-
-    } else {
-        return haystack.includes(transformedString);
-    }
-
-    return false;
-}
-
 function convertAgnaiMemoryBook(inputObj) {
     const outputObj = { entries: {} };
 
@@ -2210,6 +2364,9 @@ function convertCharacterBook(characterBook) {
             depth: entry.extensions?.depth ?? DEFAULT_DEPTH,
             selectiveLogic: entry.extensions?.selectiveLogic ?? world_info_logic.AND_ANY,
             group: entry.extensions?.group ?? '',
+            scanDepth: entry.extensions?.scan_depth ?? null,
+            caseSensitive: entry.extensions?.case_sensitive ?? null,
+            matchWholeWords: entry.extensions?.match_whole_words ?? null,
         };
     });
 
@@ -2245,7 +2402,7 @@ export function checkEmbeddedWorld(chid) {
         const checkKey = `AlertWI_${characters[chid].avatar}`;
         const worldName = characters[chid]?.data?.extensions?.world;
         if (!localStorage.getItem(checkKey) && (!worldName || !world_names.includes(worldName))) {
-            localStorage.setItem(checkKey, 1);
+            localStorage.setItem(checkKey, 'true');
 
             if (power_user.world_import_dialog) {
                 const html = `<h3>This character has an embedded World/Lorebook.</h3>

public/style.css

@@ -538,7 +538,6 @@ hr {
     background-color: var(--SmartThemeChatTintColor);
     -webkit-backdrop-filter: blur(var(--SmartThemeBlurStrength));
     text-shadow: 0px 0px calc(var(--shadowWidth) * 1px) var(--SmartThemeShadowColor);
-    scrollbar-width: thin;
     flex-direction: column;
     z-index: 30;
 }
@@ -979,7 +978,6 @@ textarea {
     font-size: var(--mainFontSize);
     font-family: "Noto Sans", "Noto Color Emoji", sans-serif;
     padding: 5px 10px;
-    scrollbar-width: thin;
     max-height: 90vh;
     max-height: 90svh;
 }
@@ -3125,7 +3123,6 @@ a {
     box-shadow: none;
     border-radius: 10px;
     overflow: hidden;
-    scrollbar-width: thin;
     flex-flow: column;
     min-width: 100px;
 }

src/embedding.js

@@ -1,6 +1,7 @@
 const TASK = 'feature-extraction';
 
 /**
+ * Gets the vectorized text in form of an array of numbers.
  * @param {string} text - The text to vectorize
  * @returns {Promise<number[]>} - The vectorized text in form of an array of numbers
  */
@@ -12,6 +13,20 @@ async function getTransformersVector(text) {
     return vector;
 }
 
+/**
+ * Gets the vectorized texts in form of an array of arrays of numbers.
+ * @param {string[]} texts - The texts to vectorize
+ * @returns {Promise<number[][]>} - The vectorized texts in form of an array of arrays of numbers
+ */
+async function getTransformersBatchVector(texts) {
+    const result = [];
+    for (const text of texts) {
+        result.push(await getTransformersVector(text));
+    }
+    return result;
+}
+
 module.exports = {
     getTransformersVector,
+    getTransformersBatchVector,
 };

src/endpoints/backends/chat-completions.js

@@ -831,7 +831,7 @@ router.post('/generate', jsonParser, function (request, response) {
                 let json = await fetchResponse.json();
                 response.send(json);
                 console.log(json);
-                console.log(json?.choices[0]?.message);
+                console.log(json?.choices?.[0]?.message);
             } else if (fetchResponse.status === 429 && retries > 0) {
                 console.log(`Out of quota, retrying in ${Math.round(timeout / 1000)}s`);
                 setTimeout(() => {

src/endpoints/characters.js

@@ -388,6 +388,9 @@ function convertWorldInfoToCharacterBook(name, entries) {
                 selectiveLogic: entry.selectiveLogic ?? 0,
                 group: entry.group ?? '',
                 prevent_recursion: entry.preventRecursion ?? false,
+                scan_depth: entry.scanDepth ?? null,
+                match_whole_words: entry.matchWholeWords ?? null,
+                case_sensitive: entry.caseSensitive ?? null,
             },
         };
 

src/endpoints/vectors.js

@@ -7,16 +7,19 @@ const { jsonParser } = require('../express-common');
 /**
  * Gets the vector for the given text from the given source.
  * @param {string} source - The source of the vector
+ * @param {Object} sourceSettings - Settings for the source, if it needs any
  * @param {string} text - The text to get the vector for
  * @returns {Promise<number[]>} - The vector for the text
  */
-async function getVector(source, text) {
+async function getVector(source, sourceSettings, text) {
     switch (source) {
         case 'mistral':
         case 'openai':
             return require('../openai-vectors').getOpenAIVector(text, source);
         case 'transformers':
             return require('../embedding').getTransformersVector(text);
+        case 'extras':
+            return require('../extras-vectors').getExtrasVector(text, sourceSettings.extrasUrl, sourceSettings.extrasKey);
         case 'palm':
             return require('../makersuite-vectors').getMakerSuiteVector(text);
     }
@@ -24,6 +27,29 @@ async function getVector(source, text) {
     throw new Error(`Unknown vector source ${source}`);
 }
 
+/**
+ * Gets the vector for the given text batch from the given source.
+ * @param {string} source - The source of the vector
+ * @param {Object} sourceSettings - Settings for the source, if it needs any
+ * @param {string[]} texts - The array of texts to get the vector for
+ * @returns {Promise<number[][]>} - The array of vectors for the texts
+ */
+async function getBatchVector(source, sourceSettings, texts) {
+    switch (source) {
+        case 'mistral':
+        case 'openai':
+            return require('../openai-vectors').getOpenAIBatchVector(texts, source);
+        case 'transformers':
+            return require('../embedding').getTransformersBatchVector(texts);
+        case 'extras':
+            return require('../extras-vectors').getExtrasBatchVector(texts, sourceSettings.extrasUrl, sourceSettings.extrasKey);
+        case 'palm':
+            return require('../makersuite-vectors').getMakerSuiteBatchVector(texts);
+    }
+
+    throw new Error(`Unknown vector source ${source}`);
+}
+
 /**
  * Gets the index for the vector collection
  * @param {string} collectionId - The collection ID
@@ -45,19 +71,20 @@ async function getIndex(collectionId, source, create = true) {
  * Inserts items into the vector collection
  * @param {string} collectionId - The collection ID
  * @param {string} source - The source of the vector
+ * @param {Object} sourceSettings - Settings for the source, if it needs any
  * @param {{ hash: number; text: string; index: number; }[]} items - The items to insert
  */
-async function insertVectorItems(collectionId, source, items) {
+async function insertVectorItems(collectionId, source, sourceSettings, items) {
     const store = await getIndex(collectionId, source);
 
     await store.beginUpdate();
 
-    for (const item of items) {
+    const vectors = await getBatchVector(source, sourceSettings, items.map(x => x.text));
-        const text = item.text;
+
-        const hash = item.hash;
+    for (let i = 0; i < items.length; i++) {
-        const index = item.index;
+        const item = items[i];
-        const vector = await getVector(source, text);
+        const vector = vectors[i];
-        await store.upsertItem({ vector: vector, metadata: { hash, text, index } });
+        await store.upsertItem({ vector: vector, metadata: { hash: item.hash, text: item.text, index: item.index } });
     }
 
     await store.endUpdate();
@@ -101,13 +128,14 @@ async function deleteVectorItems(collectionId, source, hashes) {
  * Gets the hashes of the items in the vector collection that match the search text
  * @param {string} collectionId - The collection ID
  * @param {string} source - The source of the vector
+ * @param {Object} sourceSettings - Settings for the source, if it needs any
  * @param {string} searchText - The text to search for
  * @param {number} topK - The number of results to return
  * @returns {Promise<{hashes: number[], metadata: object[]}>} - The metadata of the items that match the search text
  */
-async function queryCollection(collectionId, source, searchText, topK) {
+async function queryCollection(collectionId, source, sourceSettings, searchText, topK) {
     const store = await getIndex(collectionId, source);
-    const vector = await getVector(source, searchText);
+    const vector = await getVector(source, sourceSettings, searchText);
 
     const result = await store.queryItems(vector, topK);
     const metadata = result.map(x => x.item.metadata);
@@ -115,6 +143,28 @@ async function queryCollection(collectionId, source, searchText, topK) {
     return { metadata, hashes };
 }
 
+/**
+ * Extracts settings for the vectorization sources from the HTTP request headers.
+ * @param {string} source - Which source to extract settings for.
+ * @param {object} request - The HTTP request object.
+ * @returns {object} - An object that can be used as `sourceSettings` in functions that take that parameter.
+ */
+function getSourceSettings(source, request) {
+    // Extras API settings to connect to the Extras embeddings provider
+    let extrasUrl = '';
+    let extrasKey = '';
+    if (source === 'extras') {
+        extrasUrl = String(request.headers['x-extras-url']);
+        extrasKey = String(request.headers['x-extras-key']);
+    }
+
+    const sourceSettings = {
+        extrasUrl: extrasUrl,
+        extrasKey: extrasKey
+    };
+    return sourceSettings;
+}
+
 const router = express.Router();
 
 router.post('/query', jsonParser, async (req, res) => {
@@ -127,8 +177,9 @@ router.post('/query', jsonParser, async (req, res) => {
         const searchText = String(req.body.searchText);
         const topK = Number(req.body.topK) || 10;
         const source = String(req.body.source) || 'transformers';
+        const sourceSettings = getSourceSettings(source, req);
 
-        const results = await queryCollection(collectionId, source, searchText, topK);
+        const results = await queryCollection(collectionId, source, sourceSettings, searchText, topK);
         return res.json(results);
     } catch (error) {
         console.error(error);
@@ -145,8 +196,9 @@ router.post('/insert', jsonParser, async (req, res) => {
         const collectionId = String(req.body.collectionId);
         const items = req.body.items.map(x => ({ hash: x.hash, text: x.text, index: x.index }));
         const source = String(req.body.source) || 'transformers';
+        const sourceSettings = getSourceSettings(source, req);
 
-        await insertVectorItems(collectionId, source, items);
+        await insertVectorItems(collectionId, source, sourceSettings, items);
         return res.sendStatus(200);
     } catch (error) {
         console.error(error);

src/extras-vectors.js

@@ -0,0 +1,78 @@
+const fetch = require('node-fetch').default;
+
+/**
+ * Gets the vector for the given text from SillyTavern-extras
+ * @param {string[]} texts - The array of texts to get the vectors for
+ * @param {string} apiUrl - The Extras API URL
+ * @param {string} apiKey - The Extras API key, or empty string if API key not enabled
+ * @returns {Promise<number[][]>} - The array of vectors for the texts
+ */
+async function getExtrasBatchVector(texts, apiUrl, apiKey) {
+    return getExtrasVectorImpl(texts, apiUrl, apiKey);
+}
+
+/**
+ * Gets the vector for the given text from SillyTavern-extras
+ * @param {string} text - The text to get the vector for
+ * @param {string} apiUrl - The Extras API URL
+ * @param {string} apiKey - The Extras API key, or empty string if API key not enabled
+ * @returns {Promise<number[]>} - The vector for the text
+ */
+async function getExtrasVector(text, apiUrl, apiKey) {
+    return getExtrasVectorImpl(text, apiUrl, apiKey);
+}
+
+/**
+ * Gets the vector for the given text from SillyTavern-extras
+ * @param {string|string[]} text - The text or texts to get the vector(s) for
+ * @param {string} apiUrl - The Extras API URL
+ * @param {string} apiKey - The Extras API key, or empty string if API key not enabled *
+ * @returns {Promise<Array>} - The vector for a single text if input is string, or the array of vectors for multiple texts if input is string[]
+ */
+async function getExtrasVectorImpl(text, apiUrl, apiKey) {
+    let url;
+    try {
+        url = new URL(apiUrl);
+        url.pathname = '/api/embeddings/compute';
+    }
+    catch (error) {
+        console.log('Failed to set up Extras API call:', error);
+        console.log('Extras API URL given was:', apiUrl);
+        throw error;
+    }
+
+    const headers = {
+        'Content-Type': 'application/json',
+    };
+
+    // Include the Extras API key, if enabled
+    if (apiKey && apiKey.length > 0) {
+        Object.assign(headers, {
+            'Authorization': `Bearer ${apiKey}`,
+        });
+    }
+
+    const response = await fetch(url, {
+        method: 'POST',
+        headers: headers,
+        body: JSON.stringify({
+            text: text,  // The backend accepts {string|string[]} for one or multiple text items, respectively.
+        }),
+    });
+
+    if (!response.ok) {
+        const text = await response.text();
+        console.log('Extras request failed', response.statusText, text);
+        throw new Error('Extras request failed');
+    }
+
+    const data = await response.json();
+    const vector = data.embedding;  // `embedding`: number[] (one text item), or number[][] (multiple text items).
+
+    return vector;
+}
+
+module.exports = {
+    getExtrasVector,
+    getExtrasBatchVector,
+};

src/makersuite-vectors.js

@@ -1,6 +1,17 @@
 const fetch = require('node-fetch').default;
 const { SECRET_KEYS, readSecret } = require('./endpoints/secrets');
 
+/**
+ * Gets the vector for the given text from gecko model
+ * @param {string[]} texts - The array of texts to get the vector for
+ * @returns {Promise<number[][]>} - The array of vectors for the texts
+ */
+async function getMakerSuiteBatchVector(texts) {
+    const promises = texts.map(text => getMakerSuiteVector(text));
+    const vectors = await Promise.all(promises);
+    return vectors;
+}
+
 /**
  * Gets the vector for the given text from PaLM gecko model
  * @param {string} text - The text to get the vector for
@@ -40,4 +51,5 @@ async function getMakerSuiteVector(text) {
 
 module.exports = {
     getMakerSuiteVector,
+    getMakerSuiteBatchVector,
 };

src/openai-vectors.js

@@ -3,7 +3,7 @@ const { SECRET_KEYS, readSecret } = require('./endpoints/secrets');
 
 const SOURCES = {
     'mistral': {
-        secretKey: SECRET_KEYS.MISTRAL,
+        secretKey: SECRET_KEYS.MISTRALAI,
         url: 'api.mistral.ai',
         model: 'mistral-embed',
     },
@@ -15,12 +15,12 @@ const SOURCES = {
 };
 
 /**
- * Gets the vector for the given text from an OpenAI compatible endpoint.
+ * Gets the vector for the given text batch from an OpenAI compatible endpoint.
- * @param {string} text - The text to get the vector for
+ * @param {string[]} texts - The array of texts to get the vector for
  * @param {string} source - The source of the vector
- * @returns {Promise<number[]>} - The vector for the text
+ * @returns {Promise<number[][]>} - The array of vectors for the texts
  */
-async function getOpenAIVector(text, source) {
+async function getOpenAIBatchVector(texts, source) {
     const config = SOURCES[source];
 
     if (!config) {
@@ -43,7 +43,7 @@ async function getOpenAIVector(text, source) {
             Authorization: `Bearer ${key}`,
         },
         body: JSON.stringify({
-            input: text,
+            input: texts,
             model: config.model,
         }),
     });
@@ -55,16 +55,31 @@ async function getOpenAIVector(text, source) {
     }
 
     const data = await response.json();
-    const vector = data?.data[0]?.embedding;
 
-    if (!Array.isArray(vector)) {
+    if (!Array.isArray(data?.data)) {
         console.log('API response was not an array');
         throw new Error('API response was not an array');
     }
 
-    return vector;
+    // Sort data by x.index to ensure the order is correct
+    data.data.sort((a, b) => a.index - b.index);
+
+    const vectors = data.data.map(x => x.embedding);
+    return vectors;
+}
+
+/**
+ * Gets the vector for the given text from an OpenAI compatible endpoint.
+ * @param {string} text - The text to get the vector for
+ * @param {string} source - The source of the vector
+ * @returns {Promise<number[]>} - The vector for the text
+ */
+async function getOpenAIVector(text, source) {
+    const vectors = await getOpenAIBatchVector([text], source);
+    return vectors[0];
 }
 
 module.exports = {
     getOpenAIVector,
+    getOpenAIBatchVector,
 };