build llama.cpp embedding url path by appending instead of overwriting (#5178) * build llama.cpp embedding url path by appending instead of overwriting When using a llama-swap frontend, this allows choosing the embedding model by path; for instance, with the secondary embedding endpoint URL configured to http://127.0.0.1:5001/upstream/emb-model to pick the "emb-model" model. With this approach, the same llama-swap instance can be used to serve both the main LLM and the embedding model from local llama-server commands. * Trim trailing v1 from URL --------- Co-authored-by: Cohee <18619528+Cohee1207@users.noreply.github.com>

ab8d7e518454ad9cc030d8f22c09614246f4695a

Wagner Bruna <wbruna@users.noreply.github.com>

Signed

src/vectors/llamacpp-vectors.js+3 -2 Viewed

 import fetch from 'node-fetch';
+import urlJoin from 'url-join';
 import { setAdditionalHeadersByType } from '../additional-headers.js';
 import { TEXTGEN_TYPES } from '../constants.js';
+import { trimV1 } from '../util.js';
 /**
  * Gets the vector for the given text from LlamaCpp

  * @returns {Promise<number[][]>} - The array of vectors for the texts
  */
 export async function getLlamaCppBatchVector(texts, apiUrl, directories) {
-    const url = new URL(apiUrl);
+    const url = new URL(urlJoin(trimV1(apiUrl), '/v1/embeddings'));
-    url.pathname = '/v1/embeddings';
     const headers = {};
     setAdditionalHeadersByType(headers, TEXTGEN_TYPES.LLAMACPP, apiUrl, directories);