@botmock/export

Overview

This package is used to export data into development platforms from Botmock projects.

Install

The package is available to install on npm.

npm install @botmock/export

Guides

The following guides assumes a TypeScript environment with LTS Node.js installed.

Note: You can bootstrap such an environment by running npx @botmock/cli

Note: Tokens and ids are created and accessed using the Botmock dashboard

• Platform exporter guide
• Custom exporter guide

Platform Exporter

@botmock/export ships with a number of platform exporters, which generate file structures necessary for importing Botmock projects in popular development platforms.

Listed below are the named exports corresponding to development platforms currently supported by @botmock/export.

• BotFrameworkExporter
• DialogflowExporter
• LexExporter
• LuisExporter
• RasaExporter
• SkillsKitExporter
• WatsonExporter

Each can be accessed by through the following pattern.

import { BotFrameworkExporter } from "@botmock/export";

Let's walk through using the package's DialogflowExporter.

We begin by adding an entry point to index.ts.

// index.ts
import {
  FileWriter,
  Kind,
  ProjectReference,
  DialogflowExporter,
} from "@botmock/export";

const projectReference: ProjectReference = {
  teamId: "YOUR_TEAM_ID",
  projectId: "YOUR_PROJECT_ID",
  boardId: "YOUR_BOARD_ID",
};

async function main(): Promise<void> {
  const exporter = new DialogflowExporter({ token: "YOUR_TOKEN" });
}

main().catch((err: Error) => {
  console.error(err);
});

The main function creates a new instance of the DialogflowExporter, and expects a valid token.

If we add to this file, we can generate the file structure that the Dialogflow dashboard requires in order to restore a project.

// index.ts
async function main(): Promise<void> {
  const exporter = new DialogflowExporter({ token: "YOUR_TOKEN" });
  // The `data` property of the object returned by this method can be used by `FileWriter`, as we will see in the next step.
  const { data } = await exporter.exportProjectUnderDataTransformations({ projectReference });

  // Calling `writeAllResourcesToFiles` with `data` writes the files specificed in `data` to the filesystem.
  const writeResult = await (new FileWriter({ directoryRoot: "./output" })).writeAllResourcesToFiles({ data });
  if (writeResult.kind !== Kind.OK) {
    console.error(writeResult.value);
  } else {
    console.log("done");
  }
}

Running ts-node index.ts should generate the .json files that the Dialogflow dashboard requries to restore a project.

In the case that there is an error writing the files, an error value will be written to the console.

Custom Exporter

Let's walk through building a custom exporter by extending BaseExporter in the package.

Extending the BaseExporter is the approach taken by all platform exporters in the package.

We begin by adding an entry point to index.ts.

// index.ts
const projectReference: ProjectReference = {
  teamId: "YOUR_TEAM_ID",
  projectId: "YOUR_PROJECT_ID",
  boardId: "YOUR_BOARD_ID",
};

async function main() { }

main().catch(err => {
  console.error(err);
});

Now, in a file named `exporter.ts

// exporter.ts
import {
  BaseExporter,
  Resources,
  DataTransformation,
} from "@botmock/export";

export default class CustomExporter extends BaseExporter {
  #rootTransformation = (resources: Resources): DataTransformation => {
    return {
      // Suppose our platform requires a file named `response-structure.json`
      filename: "response-structure.json",
      data: {},
    }
  };
  dataTransformations = new Map([
    ["./path/to/output", this.#rootTransformation],
  ]);
}

exporter.ts now exports CustomExporter, which has dataTransformation and #rootTransformation.

dataTransformation

This public property must be a Map. It is responsible for mapping relative output paths to functions that can turn Botmock project data into a specific object format.

#rootTransformation

This private field must return an object (or an array of objects) with the following structure.

{
  // `filename` must be the name of a file with a `.json`, `.yml`, or `.md` extension.
  filename: string;
  // `data` can be an object that will later be mapped to `.json`, `.yml`, or `.md` file contents.
  data: object;
}

Let's see how to make the #rootTransformation do something useful.

The resources parameter that is passed to this function is an object with the following properties.

Each of the following keys contains an object with payloads for each of the key names beloning to the project referenced in index.ts.

{
  project,
  team,
  board,
  intents,
  entities,
  variables,
}

In the transformation function itself, we have access to this entire resources object, to say how the Botmock data should be transformed for the use of our platform.

Below, we assume our platform expects a nested object structure in response-structure.json.

// exporter.ts
#rootTransformation = (resources: Resources): DataTransformation => {
  const {
    project,
    board,
    intents,
    entities,
  } = resources;
  return {
    filename: "response-structure.json",
    data: {
      name: resources.project.name,
      responses: Object.values(board.board.messages).reduce((responses, response) => {
        // Because the response block data is kept in a twice-nested object, the following code unpacks the underlying data to a flat array, for use in our platform.
        const responseDataPerLocale = Object.values(response);
        const responseDataPerPlatform = responseDataPerLocale.flatMap(data => Object.values(data).flatMap(value => value.blocks))
        return {
          ...responses,
          ...responseDataPerPlatform,
        }
      }, {})
    },
  }
};

Jumping back to index.ts, we can now do the following.

// index.ts
import {
  FileWriter,
  Kind,
  ProjectReference,
} from "@botmock/export";
import { default as CustomExporter } from "./exporter.ts";

const projectReference: ProjectReference = {
  teamId: "YOUR_TEAM_ID",
  projectId: "YOUR_PROJECT_ID",
  boardId: "YOUR_BOARD_ID",
};

async function main() {
  const exporter = new CustomExporter({ token: "YOUR_TOKEN" });
  // The `data` property of the object returned by this method can be used by `FileWriter`, as we will see in the next step.
  const { data } = await exporter.exportProjectUnderDataTransformations({ projectReference });
}

main().catch(err => {
  console.error(err);
});

Calling exporter.exportProjectUnderDataTransformations({ projectReference }) allows us to have access to the data object, which can be used to write files to the local filesystem in a predictable way.

Let's see how FileWriter works.

// index.ts
async function main() {
  const exporter = new CustomExporter({ token: "YOUR_TOKEN" });
  const { data } = await exporter.exportProjectUnderDataTransformations({ projectReference });

  // Calling `writeAllResourcesToFiles` with `data` from the previous step writes the files specificed in `data` to the filesystem.
  const writeResult = await (new FileWriter()).writeAllResourcesToFiles({ data });
  // `writeResult` is an object with `kind` and `value` properties based on the success of the writes.
  if (writeResult.kind !== Kind.OK) {
    console.error(writeResult.value);
  } else {
    console.log("done");
  }
}

Running ts-node index.ts should generate the single response-structure.json we outlined in our CustomExporter.

API

BaseExporter

class extended by platform exporters and custom exporters.

exportProjectUnderDataTransformations

Returns Promise which resolves an object containing the data property consumable by FileWriter.

data

The data property returned by exportProjectUnderDataTransformations has the following shape.

Key name	Type	Description
[directoryPath]	array	an array of objects containing { filename: "file.json", data: {}, }, where data property in this object is a description of file contents that should belong in filename.

This means that data property can look like the following.

{
  './': [
    { filename: 'some-project.json', data: {} },
  ]
}

exportProjectUnderDataTransformations takes a single options object as an argument.

The options object contains the following.

Key name	Type	Description
projectReference	object	object containing key, value pairs for "teamId", "projectId", and optionally, "boardId"

dataTransformations

Required property on any BaseExporter. This property should be a Map that describes how Botmock project resources should be transformed for a given path.

import { BaseExporter, Resources, DataTransformation } from "@botmock/export";

class MyExporter extends BaseExporter {
  #rootTransformation = (resources: Resources): DataTransformation => {
    return {
      filename: "file.json",
      data: {
        intents: resources.intents.map(intent => Object.values(intent.utterances)),
      }
    }
  };
  dataTransformations = new Map([
    ["../../some/relative/path", this.#rootTransformation],
  ]);
}

FileWriter

class for parsing data property generated by exportProjectUnderDataTransformations and writing the files specified there to the correct locations in the local filesystem.

It can optionally take an options object, which has the following form.

Key name	Type	Description
directoryRoot	string	relative path to a path from which all paths in the dataTransformation should be joined
jsonWriteOptions	object	object which is passed down to the fs-extra writeJSON method options

import { FileWriter } from "@botmock/export";
import { MyExporter } from "./my-exporter";

(async () => {
  const exporter = new MyExporter({ token: "YOUR_TOKEN" });
  const { data } = await exporter.exportProjectUnderDataTransformations({ projectReference });

  const writer = new FileWriter();
})();

writeAllResourcesToFiles

Returns a Promise which resolves an object describing the success or failure of the file writing.

It takes a options argument with the following structure.

Key name	Type	Description
data	object	data property

import { FileWriter } from "@botmock/export";
import { MyExporter } from "./my-exporter";

(async () => {
  const exporter = new MyExporter({ token: "YOUR_TOKEN" });
  const { data } = await exporter.exportProjectUnderDataTransformations({ projectReference });

  const writer = new FileWriter();
  const writeResult = await writer.writeAllResourcesToFiles({ data });
})();