Multiple Chains Example (LangChain Port)

[allanbunch]

Multiple Chains

A Node-RED flow can be used to perform a sequence of dependent model inferences.

This example is a port of LangChain's Multiple chains cookbook entry.

Download: Multiple Chains Example Flow JSON

For this example, Prompt Template 1 is configured to use OpenAI's gpt-3.5-turbo model. Prompt Template 2 is configured to use Ollama backed by Stability AI's amazing Stable LM 2 Zephyr 1.6B small language model (SLM). You'll want to configure these and their corresponding AI API nodes according to your specific needs.
Note: This is a simplified example of a prompt architecture technique I call Mixture of Model Outputs.

Instructions

1. Configure the AI API nodes as needed.
2. Set the Topics node's msg.topic values to whatever you like--Prompt Template 2 is configured to present the final response using your specified topic.language.

---

LangChain JS Equivalent

import { PromptTemplate } from "@langchain/core/prompts";
import { RunnableSequence } from "@langchain/core/runnables";
import { StringOutputParser } from "@langchain/core/output_parsers";
import { ChatAnthropic } from "@langchain/anthropic";

const prompt1 = PromptTemplate.fromTemplate(
  `What is the city {person} is from? Only respond with the name of the city.`
);
const prompt2 = PromptTemplate.fromTemplate(
  `What country is the city {city} in? Respond in {language}.`
);

const model = new ChatAnthropic({});

const chain = prompt1.pipe(model).pipe(new StringOutputParser());

const combinedChain = RunnableSequence.from([
  {
    city: chain,
    language: (input) => input.language,
  },
  prompt2,
  model,
  new StringOutputParser(),
]);

const result = await combinedChain.invoke({
  person: "Obama",
  language: "German",
});

console.log(result);

/*
  Chicago befindet sich in den Vereinigten Staaten.
*/