Save an audiofile from Google Text-to-Speech to Firebase Storage using Google Cloud Storage?

Question

We're trying to get an audiofile from Google Text-to-Speech and save it to Firebase Storage, using a Google Cloud Function. The documentation for Google Text-to-Speech show how to get an audiofile and save it locally:

// Performs the Text-to-Speech request
const [response] = await client.synthesizeSpeech(request);
// Write the binary audio content to a local file
const writeFile = util.promisify(fs.writeFile);
await writeFile('output.mp3', response.audioContent, 'binary');
console.log('Audio content written to file: output.mp3');

This results in an error message Error: EROFS: read-only file system. Google Cloud Storage doesn't allow writing files locally.

Using Firebase Storage bucket.upload() has a few problems:

   const destinationPath = 'Audio/Spanish' + filename.ogg;
   // Performs the Text-to-Speech request
   const [response] = await client.synthesizeSpeech(request);
   // response.audioContent is the downloaded file
   await bucket.upload(response.audioContent, {
      destination: destinationPath
   ));

The error message is TypeError: Path must be a string. The first parameter of bucket.upload() is The fully qualified path to the file you wish to upload to your bucket. and is expected to be a string so response.audioContent doesn't work.

The documentation for bucket.upload() suggests that destination: destinationPath is where we should put the path to the Firebase Storage location. Is this correct?

How do we take the audiofile from Google Text-to-Speech (response.audioContent) and save it as a string that bucket.upload() can access? Or should we use something else instead of bucket.upload()?

Here's our full cloud function:

exports.Google_T2S = functions.firestore.document('Users/{userID}/Spanish/T2S_Request').onUpdate((change, context) => {
  if (change.after.data().word != undefined) {

    // Performs the Text-to-Speech request
    async function test() {
      try {
        const word = change.after.data().word; // the text
        const longLanguage = 'Spanish';
        const audioFormat = '.mp3';
        // copied from https://cloud.google.com/text-to-speech/docs/quickstart-client-libraries#client-libraries-usage-nodejs
        const fs = require('fs');
        const util = require('util');
        const textToSpeech = require('@google-cloud/text-to-speech'); // Imports the Google Cloud client library
        const client = new textToSpeech.TextToSpeechClient(); // Creates a client

        let myWordFile = word.replace(/ /g,"_"); // replace spaces with underscores in the file name
        myWordFile = myWordFile.toLowerCase(); // convert the file name to lower case
        myWordFile = myWordFile + audioFormat; // append .mp3 to the file name;

        // copied from https://cloud.google.com/blog/products/gcp/use-google-cloud-client-libraries-to-store-files-save-entities-and-log-data
        const {Storage} = require('@google-cloud/storage');
        const storage = new Storage();
        const bucket = storage.bucket('myProject-cd99d.appspot.com');
        const destinationPath = 'Audio/Spanish/' + myWordFile;

        const request = { // Construct the request
          input: {text: word},
          // Select the language and SSML Voice Gender (optional)
          voice: {languageCode: 'es-ES', ssmlGender: 'FEMALE'},
          // Select the type of audio encoding
          audioConfig: {audioEncoding: 'MP3'},
        };

        const [response] = await client.synthesizeSpeech(request);
        // Write the binary audio content to a local file
        const writeFile = util.promisify(fs.writeFile);
        await writeFile('output.mp3', response.audioContent, 'binary');
        console.log('Audio content written to file: output.mp3')
        // response.audioContent is the downloaded file

        await bucket.upload(response.audioContent, {
          destination: destinationPath
        });
      }
      catch (error) {
        console.error(error);
      }
    }
    test();
  } // close if
  return 0; // prevents an error message "Function returned undefined, expected Promise or value"
});

Answer 1

file.save() was the answer. util.promisify was unnecessary, and causes an error message about original something. Here's the finished cloud function:

exports.Google_T2S = functions.firestore.document('Users/{userID}/Spanish/T2S_Request').onUpdate((change, context) => {
  if (change.after.data().word != undefined) {

    async function textToSpeechRequest() {
      try {
        const word = change.after.data().word; // the text
        const longLanguage = 'Spanish';
        const audioFormat = '.mp3';
        // copied from https://cloud.google.com/text-to-speech/docs/quickstart-client-libraries#client-libraries-usage-nodejs
        const util = require('util');
        const textToSpeech = require('@google-cloud/text-to-speech'); // Imports the Google Cloud client library
        const client = new textToSpeech.TextToSpeechClient(); // Creates a client

        let myWordFile = word.replace(/ /g,"_"); // replace spaces with underscores in the file name
        myWordFile = myWordFile.toLowerCase(); // convert the file name to lower case
        myWordFile = myWordFile + audioFormat; // append .mp3 to the file name;

        // copied from https://cloud.google.com/blog/products/gcp/use-google-cloud-client-libraries-to-store-files-save-entities-and-log-data
        const {Storage} = require('@google-cloud/storage');
        const storage = new Storage();
        const bucket = storage.bucket('myProject-cd99d.appspot.com');
        var file = bucket.file('Audio/Spanish/' + myWordFile);

        const request = { // Construct the request
          input: {text: word},
          // Select the language and SSML Voice Gender (optional)
          voice: {languageCode: 'es-ES', ssmlGender: 'FEMALE'},
          // Select the type of audio encoding
          audioConfig: {audioEncoding: 'MP3'},
        };

        const options = { // construct the file to write
          metadata: {
            contentType: 'audio/mpeg',
            metadata: {
              source: 'Google Text-to-Speech'
            }
          }
        };

        // copied from https://cloud.google.com/text-to-speech/docs/quickstart-client-libraries#client-libraries-usage-nodejs
        const [response] = await client.synthesizeSpeech(request);
        // Write the binary audio content to a local file
        // response.audioContent is the downloaded file
        return await file.save(response.audioContent, options)
        .then(function() {
          console.log("File written to Firebase Storage.")
        })
        .catch(function(error) {
          console.error(error);
        });
      );
    } // close try
    catch (error) {
      console.error(error);
    } // close catch
  } // close async
  textToSpeechRequest();
} // close if
}); // close Google_T2S

We're getting an error TypeError: [ERR_INVALID_ARG_TYPE]: The "original" argument must be of type function at Object.promisify. This error doesn't appear to effect the cloud function.

To reiterate the stuff that didn't work, fs.createWriteStream didn't work because Google Cloud Functions can't handle Node file system commands. Instead, Google Cloud Functions have their own methods that wrap the Node file system commands. bucket.upload() will upload a local file to a bucket, but the path to the local file has to be a string, not a buffer or a stream coming from an API. file.save() is documented as

Write arbitrary data to a file.

This is a convenience method which wraps File#createWriteStream.

That's what I want! If there's one thing about my data, it's arbitrary. Or maybe contrary by nature. After that we just had to straighten out the contentType (audio/mpeg, not mp3) and the file path.