rixty/SillyTavern
1
0
Fork 0
mirror of https://github.com/SillyTavern/SillyTavern.git synced 2025-06-05 21:59:27 +02:00
Code Issues Packages Projects Releases Wiki Activity

Add ollama and llamacpp as vector sources

Browse Source
This commit is contained in:
Cohee
2024-05-28 22:54:50 +03:00
parent c858fccc5f
commit 2b3dfc5ae2
6 changed files with 286 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

88
public/scripts/extensions/vectors/index.js
View File

@@ -25,6 +25,7 @@ import { getDataBankAttachments, getFileAttachment } from '../../chats.js';
import { debounce, getStringHash as calculateHash, waitUntilCondition, onlyUnique, splitRecursive } from '../../utils.js';
import { debounce_timeout } from '../../constants.js';
import { getSortedEntries } from '../../world-info.js';
import { textgen_types, textgenerationwebui_settings } from '../../textgen-settings.js';
const MODULE_NAME = 'vectors';
@@ -38,6 +39,8 @@ const settings = {
togetherai_model: 'togethercomputer/m2-bert-80M-32k-retrieval',
openai_model: 'text-embedding-ada-002',
cohere_model: 'embed-english-v3.0',
ollama_model: 'mxbai-embed-large',
ollama_keep: false,
summarize: false,
summarize_sent: false,
summary_source: 'main',
@@ -272,6 +275,10 @@ async function synchronizeChat(batchSize = 5) {
switch (cause) {
case 'api_key_missing':
return 'API key missing. Save it in the "API Connections" panel.';
case 'api_url_missing':
return 'API URL missing. Save it in the "API Connections" panel.';
case 'api_model_missing':
return 'Vectorization Source Model is required, but not set.';
case 'extras_module_missing':
return 'Extras API must provide an "embeddings" module.';
default:
@@ -637,6 +644,12 @@ function getVectorHeaders() {
case 'cohere':
addCohereHeaders(headers);
break;
case 'ollama':
addOllamaHeaders(headers);
break;
case 'llamacpp':
addLlamaCppHeaders(headers);
break;
default:
break;
}
@@ -685,6 +698,28 @@ function addCohereHeaders(headers) {
});
}
/**
* Add headers for the Ollama API source.
* @param {object} headers Header object
*/
function addOllamaHeaders(headers) {
Object.assign(headers, {
'X-Ollama-Model': extension_settings.vectors.ollama_model,
'X-Ollama-URL': textgenerationwebui_settings.server_urls[textgen_types.OLLAMA],
'X-Ollama-Keep': !!extension_settings.vectors.ollama_keep,
});
}
/**
* Add headers for the LlamaCpp API source.
* @param {object} headers Header object
*/
function addLlamaCppHeaders(headers) {
Object.assign(headers, {
'X-LlamaCpp-URL': textgenerationwebui_settings.server_urls[textgen_types.LLAMACPP],
});
}
/**
* Inserts vector items into a collection
* @param {string} collectionId - The collection to insert into
@@ -692,18 +727,7 @@ function addCohereHeaders(headers) {
* @returns {Promise<void>}
*/
async function insertVectorItems(collectionId, items) {
if (settings.source === 'openai' && !secret_state[SECRET_KEYS.OPENAI] ||
settings.source === 'palm' && !secret_state[SECRET_KEYS.MAKERSUITE] ||
settings.source === 'mistral' && !secret_state[SECRET_KEYS.MISTRALAI] ||
settings.source === 'togetherai' && !secret_state[SECRET_KEYS.TOGETHERAI] ||
settings.source === 'nomicai' && !secret_state[SECRET_KEYS.NOMICAI] ||
settings.source === 'cohere' && !secret_state[SECRET_KEYS.COHERE]) {
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' });
}
throwIfSourceInvalid();
const headers = getVectorHeaders();
@@ -722,6 +746,33 @@ async function insertVectorItems(collectionId, items) {
}
}
/**
* Throws an error if the source is invalid (missing API key or URL, or missing module)
*/
function throwIfSourceInvalid() {
if (settings.source === 'openai' && !secret_state[SECRET_KEYS.OPENAI] ||
settings.source === 'palm' && !secret_state[SECRET_KEYS.MAKERSUITE] ||
settings.source === 'mistral' && !secret_state[SECRET_KEYS.MISTRALAI] ||
settings.source === 'togetherai' && !secret_state[SECRET_KEYS.TOGETHERAI] ||
settings.source === 'nomicai' && !secret_state[SECRET_KEYS.NOMICAI] ||
settings.source === 'cohere' && !secret_state[SECRET_KEYS.COHERE]) {
throw new Error('Vectors: API key missing', { cause: 'api_key_missing' });
}
if (settings.source === 'ollama' && !textgenerationwebui_settings.server_urls[textgen_types.OLLAMA] ||
settings.source === 'llamacpp' && !textgenerationwebui_settings.server_urls[textgen_types.LLAMACPP]) {
throw new Error('Vectors: API URL missing', { cause: 'api_url_missing' });
}
if (settings.source === 'ollama' && !settings.ollama_model) {
throw new Error('Vectors: API model missing', { cause: 'api_model_missing' });
}
if (settings.source === 'extras' && !modules.includes('embeddings')) {
throw new Error('Vectors: Embeddings module missing', { cause: 'extras_module_missing' });
}
}
/**
* Deletes vector items from a collection
* @param {string} collectionId - The collection to delete from
@@ -870,6 +921,8 @@ function toggleSettings() {
$('#together_vectorsModel').toggle(settings.source === 'togetherai');
$('#openai_vectorsModel').toggle(settings.source === 'openai');
$('#cohere_vectorsModel').toggle(settings.source === 'cohere');
$('#ollama_vectorsModel').toggle(settings.source === 'ollama');
$('#llamacpp_vectorsModel').toggle(settings.source === 'llamacpp');
$('#nomicai_apiKey').toggle(settings.source === 'nomicai');
}
@@ -1154,6 +1207,17 @@ jQuery(async () => {
Object.assign(extension_settings.vectors, settings);
saveSettingsDebounced();
});
$('#vectors_ollama_model').val(settings.ollama_model).on('input', () => {
$('#vectors_modelWarning').show();
settings.ollama_model = String($('#vectors_ollama_model').val());
Object.assign(extension_settings.vectors, settings);
saveSettingsDebounced();
});
$('#vectors_ollama_keep').prop('checked', settings.ollama_keep).on('input', () => {
settings.ollama_keep = $('#vectors_ollama_keep').prop('checked');
Object.assign(extension_settings.vectors, settings);
saveSettingsDebounced();
});
$('#vectors_template').val(settings.template).on('input', () => {
settings.template = String($('#vectors_template').val());
Object.assign(extension_settings.vectors, settings);

27
public/scripts/extensions/vectors/settings.html
View File

@@ -12,14 +12,37 @@
<select id="vectors_source" class="text_pole">
<option value="cohere">Cohere</option>
<option value="extras">Extras</option>
<option value="palm">Google MakerSuite (PaLM)</option>
<option value="palm">Google MakerSuite</option>
<option value="llamacpp">llama.cpp</option>
<option value="transformers">Local (Transformers)</option>
<option value="ollama">Ollama</option>
<option value="openai">OpenAI</option>
<option value="mistral">MistralAI</option>
<option value="nomicai">NomicAI</option>
<option value="openai">OpenAI</option>
<option value="togetherai">TogetherAI</option>
</select>
</div>
<div class="flex-container flexFlowColumn" id="ollama_vectorsModel">
<label for="vectors_ollama_model">
Vectorization Model
</label>
<input id="vectors_ollama_model" class="text_pole" type="text" placeholder="Model tag, e.g. llama3" />
<label for="vectors_ollama_keep" class="checkbox_label" title="When checked, the model will not be unloaded after use.">
<input id="vectors_ollama_keep" type="checkbox" />
<span>Keep model in memory</span>
</label>
<i>
Hint: Download models and set the URL in the API connection settings.
</i>
</div>
<div class="flex-container flexFlowColumn" id="llamacpp_vectorsModel">
<span>
The server MUST be started with the <code>--embedding</code> flag to use this feature!
</span>
<i>
Hint: Set the URL in the API connection settings.
</i>
</div>
<div class="flex-container flexFlowColumn" id="openai_vectorsModel">
<label for="vectors_openai_model">
Vectorization Model