zack3d/big-agi
1
0
Fork 0
mirror of https://github.com/enricoros/big-AGI.git synced 2026-05-10 21:50:14 -07:00
Code Issues Packages Projects Releases Wiki Activity

Ollama: update dispatch, Fixes #678

Browse Source
This commit is contained in:
Enrico Ros
2024-12-10 22:00:15 -08:00
parent 8bc118c060
commit 24e8539b3e
1 changed files with 19 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/modules/aix/server/dispatch/chatGenerate/chatGenerate.dispatch.ts
+19 -9
View File
@@ -1,5 +1,6 @@
import { anthropicAccess } from '~/modules/llms/server/anthropic/anthropic.router';
import { geminiAccess } from '~/modules/llms/server/gemini/gemini.router';
import { ollamaAccess } from '~/modules/llms/server/ollama/ollama.router';
import { openAIAccess } from '~/modules/llms/server/openai/openai.router';
import type { AixAPI_Access, AixAPI_Model, AixAPIChatGenerate_Request } from '../../api/aix.wiretypes';
@@ -53,16 +54,25 @@ export function createChatGenerateDispatch(access: AixAPI_Access, model: AixAPI_
chatGenerateParse: createGeminiGenerateContentResponseParser(model.id, streaming),
};
/**
* Ollama has now an OpenAI compability layer for `chatGenerate` API, but still its own protocol for models listing.
* - as such, we 'cast' here to the dispatch to an OpenAI dispatch, while using Ollama access
* - we still use the ollama.router for the models listing and aministration APIs
*
* For reference we show the old code for body/demuxerFormat/chatGenerateParse also below
*/
case 'ollama':
throw new Error('Ollama is not supported in this context');
// return {
// request: {
// ...ollamaAccess(access, OLLAMA_PATH_CHAT),
// body: ollamaChatCompletionPayload(model, _hist, access.ollamaJson, streaming),
// },
// demuxerFormat: streaming ? 'json-nl' : null,
// chatGenerateParse: createDispatchParserOllama(),
// };
return {
request: {
...ollamaAccess(access, '/v1/chat/completions'), // use the OpenAI-compatible endpoint
// body: ollamaChatCompletionPayload(model, _hist, access.ollamaJson, streaming),
body: aixToOpenAIChatCompletions('openai', model, chatGenerate, access.ollamaJson, streaming),
},
// demuxerFormat: streaming ? 'json-nl' : null,
demuxerFormat: streaming ? 'sse' : null,
// chatGenerateParse: createDispatchParserOllama(),
chatGenerateParse: streaming ? createOpenAIChatCompletionsChunkParser() : createOpenAIChatCompletionsParserNS(),
};
case 'azure':
case 'deepseek':