Create Chatbot with gpt-trainer API on New Modified Linked Reference from Roam Research API

Pipedream makes it easy to connect APIs for gpt-trainer, Roam Research and 2,800+ other apps remarkably fast.

Trigger workflow on

New Modified Linked Reference from the Roam Research API

Next, do this

Create Chatbot with the gpt-trainer API

No credit card required

▶

Watch us build a workflow

8 min

Watch now ➜

Trusted by 1,000,000+ developers from startups to Fortune 500 companies

Developers ♥ Pipedream

Getting Started#

This integration creates a workflow with a Roam Research trigger and gpt-trainer action. When you configure and deploy the workflow, it will run on Pipedream's servers 24x7 for free.

Select this integration
Configure the New Modified Linked Reference trigger
1. Connect your Roam Research account
2. Configure Polling Schedule
3. Select a Resource Type
4. Configure Page Title Or Block UID
Configure the Create Chatbot action
1. Connect your gpt-trainer account
2. Configure Name
3. Optional- Configure Prompt
4. Optional- Configure Temperature
5. Optional- Select a Model
6. Optional- Select a Visibility
7. Optional- Configure Rate Limit Message
8. Optional- Configure Show Citations
Deploy the workflow
Send a test event to validate your setup
Turn on the trigger

Details#

This integration uses pre-built, source-available components from Pipedream's GitHub repo. These components are developed by Pipedream and the community, and verified and maintained by Pipedream.

To contribute an update to an existing component or create a new component, create a PR on GitHub. If you're new to Pipedream component development, you can start with quickstarts for trigger span and action development, and then review the component API reference.

Trigger#

New Modified Linked Reference on Roam Research

Description:Emit new event for each new or modified linked reference in Roam Research.

Version:0.0.1

Key:roamresearch-new-modified-linked-reference

View on GitHub

Trigger Code#

import { DEFAULT_POLLING_SOURCE_TIMER_INTERVAL } from "@pipedream/platform";
import app from "../../roamresearch.app.mjs";
import utils from "../../common/utils.mjs";

export default {
  key: "roamresearch-new-modified-linked-reference",
  name: "New Modified Linked Reference",
  description: "Emit new event for each new or modified linked reference in Roam Research.",
  version: "0.0.1",
  type: "source",
  dedupe: "unique",
  props: {
    app,
    db: "$.service.db",
    timer: {
      type: "$.interface.timer",
      label: "Polling Schedule",
      description: "How often to poll the API",
      default: {
        intervalSeconds: DEFAULT_POLLING_SOURCE_TIMER_INTERVAL,
      },
    },
    resourceType: {
      propDefinition: [
        app,
        "resourceType",
      ],
    },
    pageOrBlock: {
      propDefinition: [
        app,
        "pageOrBlock",
      ],
    },
  },
  methods: {
    getResourcesName() {
      return "result.:block/_refs";
    },
    getResourcesFn() {
      return this.app.pull;
    },
    getResourcesFnArgs() {
      const {
        resourceType,
        pageOrBlock,
      } = this;

      const attribute = resourceType === "page"
        ? ":node/title"
        : ":block/uid";

      return {
        data: {
          selector: `[${attribute} :block/string :block/order :edit/time {:block/_refs ...}]`,
          eid: `[${attribute} "${pageOrBlock}"]`,
        },
      };
    },
    generateMeta(resource) {
      const ts = resource[":edit/time"];
      return {
        id: ts,
        summary: `Link Reference: ${resource[":block/string"]}`,
        ts,
      };
    },
    processResource(resource) {
      const meta = this.generateMeta(resource);
      this.$emit(resource, meta);
    },
  },
  async run() {
    const {
      getResourcesFn,
      getResourcesName,
      getResourcesFnArgs,
      processResource,
    } = this;

    const resourcesFn = getResourcesFn();
    const response = await resourcesFn(getResourcesFnArgs());
    const resources = utils.getNestedProperty(response, getResourcesName());

    if (!resources) {
      console.log("No resources found");
      return;
    }

    Array.from(resources)
      .reverse()
      .forEach(processResource);
  },
};

Trigger Configuration#

This component may be configured based on the props defined in the component code. Pipedream automatically prompts for input values in the UI and CLI.

Label	Prop	Type	Description
Roam Research	`app`	`app`	This component uses the Roam Research app.
N/A	`db`	`$.service.db`	This component uses `$.service.db` to maintain state between executions.
Polling Schedule	`timer`	`$.interface.timer`	How often to poll the API
Resource Type	`resourceType`	`string`	Select a value from the drop down menu:`pageblock`
Page Title Or Block UID	`pageOrBlock`	`string`	The page title of the block uid to get data for. Page title example: `My Page` and Block UID example: `ideWWvTgI`.

Trigger Authentication#

Roam Research uses API keys for authentication. When you connect your Roam Research account, Pipedream securely stores the keys so you can easily authenticate to Roam Research APIs in both code and no-code steps.

About Roam Research#

A note-taking tool for networked thought.

Action#

Create Chatbot on gpt-trainer

Description:Creates a new chatbot that belongs to the authenticated user. [See the documentation](https://guide.gpt-trainer.com/api-reference/chatbots/create)

Version:0.0.1

Key:gpt_trainer-create-chatbot

View on GitHub

gpt-trainer Overview#

The gpt-trainer API is a tool designed to train, run, and manage custom GPT-2 and GPT-3 models. It provides endpoints for submitting training data, starting the training process, and generating text from the trained model. With Pipedream's serverless integration platform, you can automate workflows that interact with the gpt-trainer API. You can trigger workflows using webhooks, schedule them, or even run them in response to events from other apps. Integrate the gpt-trainer API with other services on Pipedream to create powerful applications such as automated content creation, personalized messaging, or AI-driven data analysis.

Action Code#

import app from "../../gpt_trainer.app.mjs";

export default {
  key: "gpt_trainer-create-chatbot",
  name: "Create Chatbot",
  description: "Creates a new chatbot that belongs to the authenticated user. [See the documentation](https://guide.gpt-trainer.com/api-reference/chatbots/create)",
  version: "0.0.1",
  type: "action",
  props: {
    app,
    name: {
      type: "string",
      label: "Name",
      description: "The name of the chatbot",
    },
    prompt: {
      type: "string",
      label: "Prompt",
      description: "The prompt for the chatbot",
      optional: true,
    },
    temperature: {
      type: "string",
      label: "Temperature",
      description: "The temperature setting for the chatbot",
      optional: true,
    },
    model: {
      type: "string",
      label: "Model",
      description: "The model of the chatbot",
      optional: true,
      default: "gpt-3.5-turbo",
      options: [
        "gpt-3.5-turbo",
        "gpt-3.5-turbo-16k",
        "gpt-4",
      ],
    },
    visibility: {
      type: "string",
      label: "Visibility",
      description: "The visibility of the chatbot",
      optional: true,
      default: "private",
      options: [
        "public",
        "private",
        "hybrid",
      ],
    },
    rateLimitMessage: {
      type: "string",
      label: "Rate Limit Message",
      description: "The rate limit message for the chatbot. Eg. `Too many messages in a row.`",
      optional: true,
    },
    showCitations: {
      type: "boolean",
      label: "Show Citations",
      description: "Whether the chatbot should show citations",
      optional: true,
    },
  },
  methods: {
    parseFloat(value) {
      const parsed = parseFloat(value);
      return isFinite(parsed)
        ? parsed
        : 0;
    },
    createChatbot(args = {}) {
      return this.app.post({
        path: "/chatbot/create",
        ...args,
      });
    },
  },
  async run({ $ }) {
    const {
      createChatbot,
      parseFloat,
      name,
      prompt,
      temperature,
      model,
      visibility,
      rateLimitMessage,
      showCitations,
    } = this;

    const response = await createChatbot({
      $,
      data: {
        name,
        temperature: parseFloat(temperature),
        prompt,
        model,
        visibility,
        show_citations: showCitations,
        rate_limit_message: rateLimitMessage,
      },
    });

    $.export("$summary", `Successfully created chatbot with UUID \`${response.uuid}\`.`);
    return response;
  },
};

Action Configuration#

This component may be configured based on the props defined in the component code. Pipedream automatically prompts for input values in the UI.

Label	Prop	Type	Description
gpt-trainer	`app`	`app`	This component uses the gpt-trainer app.
Name	`name`	`string`	The name of the chatbot
Prompt	`prompt`	`string`	The prompt for the chatbot
Temperature	`temperature`	`string`	The temperature setting for the chatbot
Model	`model`	`string`	Select a value from the drop down menu:`gpt-3.5-turbogpt-3.5-turbo-16kgpt-4`
Visibility	`visibility`	`string`	Select a value from the drop down menu:`publicprivatehybrid`
Rate Limit Message	`rateLimitMessage`	`string`	The rate limit message for the chatbot. Eg. `Too many messages in a row.`
Show Citations	`showCitations`	`boolean`	Whether the chatbot should show citations

Action Authentication#

gpt-trainer uses API keys for authentication. When you connect your gpt-trainer account, Pipedream securely stores the keys so you can easily authenticate to gpt-trainer APIs in both code and no-code steps.