From ab0d666f0376f183446898f35a88000fb336bd2f Mon Sep 17 00:00:00 2001
From: Elliot Kang <kkang2097@gmail.com>
Date: Sat, 30 Sep 2023 11:46:54 -0700
Subject: [PATCH] fixed imports, moved llmStrem example
---
{examples => apps/simple}/llmStream.ts | 37 ++++++++++++--------------
1 file changed, 17 insertions(+), 20 deletions(-)
rename {examples => apps/simple}/llmStream.ts (53%)
diff --git a/examples/llmStream.ts b/apps/simple/llmStream.ts
similarity index 53%
rename from examples/llmStream.ts
rename to apps/simple/llmStream.ts
index ffaa495b6..4cf0ff366 100644
--- a/examples/llmStream.ts
+++ b/apps/simple/llmStream.ts
@@ -1,6 +1,6 @@
-import * as tiktoken from "tiktoken-node";
-import { ChatMessage, OpenAI } from "../packages/core/src/llm/LLM";
-import {SimpleChatEngine } from "../packages/core/src/ChatEngine";
+import {ChatMessage, OpenAI, SimpleChatEngine} from "llamaindex";
+import { stdin as input, stdout as output } from "node:process";
+import readline from "node:readline/promises";
async function main() {
const query: string = `
@@ -13,9 +13,7 @@ Where is Istanbul?
var accumulated_result: string = "";
var total_tokens: number = 0;
- //Callback stuff, like logging token usage.
- //GPT 3.5 Turbo uses CL100K_Base encodings, check your LLM to see which tokenizer it uses.
- const encoding = tiktoken.getEncoding("cl100k_base");
+ //TODO: Add callbacks later
//Stream Complete
//Note: Setting streaming flag to true or false will auto-set your return type to
@@ -31,25 +29,24 @@ Where is Istanbul?
accumulated_result += part;
}
- const correct_total_tokens: number =
- encoding.encode(accumulated_result).length;
-
- console.log(accumulated_result);
- //Check if our stream token counter works
- console.log(
- `Output token total using tokenizer on accumulated output: ${correct_total_tokens}`,
- );
-
accumulated_result = "";
const chatEngine: SimpleChatEngine = new SimpleChatEngine();
- const chatStream = await chatEngine.chat(query, undefined, true);
- for await (const part of chatStream){
- console.log(part);
- accumulated_result += part;
+
+ const rl = readline.createInterface({ input, output });
+ while (true) {
+ const query = await rl.question("Query: ");
+
+ if (!query) {
+ break;
}
- console.log(accumulated_result);
+ const chatStream = await chatEngine.chat(query, undefined, true);
+ for await (const part of chatStream){
+ process.stdout.write(part);
+ // accumulated_result += part;
+ }
+ }
}
main();
--
GitLab