DEV Community

Sanjeet Sahay
Sanjeet Sahay

Posted on • Updated on

axios/got faceoff: uploading a file to Amazon WorkDocs using Amazon S3 pre-signed url

Overview

If you are using Amazon WorkDocs as your managed cloud based content management and/or storage system, and if you are planning to automate tasks such as integrate it with other document/content storage systems, then you must have come across the use case of uploading a file. After some automation, you should be able to do this at scale. In a future post, I will share a detailed reference architecture on how to build such an integrated system.

The following sections demonstrate the various aspects of the app, starting with setting up a simple Node.js app. However, there are some prerequisites

Initialize npm project

I have used the following commands to initialize a new npm project

➜ mkdir workdocs-sample && cd workdocs-sample
➜ npm init
➜ npm install aws-sdk axios form-data got
➜ touch index.js
Enter fullscreen mode Exit fullscreen mode

After initialization, my folder structure looks like this:

➜  workdocs-sample ls
da-quiz-storage-result.pdf 
index.js                   
node_modules               
package-lock.json          
package.json               
yarn.lock
Enter fullscreen mode Exit fullscreen mode

Initialize the WorkDocs client

Setup AWS credentials in index.js. For more information, read best practices to use AWS credentials in your development environment

const AWS = require("aws-sdk");
const credentials = new AWS.SharedIniFileCredentials({ profile: "default" });
AWS.config.credentials = credentials;
Enter fullscreen mode Exit fullscreen mode

In addition to that, you'll need the following declarations

const got = require("got");
const fs = require("fs");
const FormData = require("form-data");
const workdocs = new AWS.WorkDocs();
Enter fullscreen mode Exit fullscreen mode

Finally, initialize the WorkDocs client

const workdocs = new AWS.WorkDocs();
Enter fullscreen mode Exit fullscreen mode

Steps to upload a file

To upload a file to a WorkDocs folder you need the following:

  • a folder ID to upload
    • to get the root folder ID, you need to make a call to describeUsers API
    • if you have created new folders at the root, then you need to call describeFolderContents with the root folder ID
  • call initiateDocumentVersionUpload with the folder ID, name of the file, and optionally, a content type. It returns an Amazon S3 pre-signed upload url, document ID, and a version ID among other things
  • use got to upload the file to the returned uploadUrl
  • call updateDocumentVersion with document ID, version ID, and set VersionStatus to ACTIVE

Get the root folder ID

Every user has a root folder which can contain one or more children - nothing fancy, just the usual nested folder structure. The root folder has an ID that can used to create folders inside it. Using the describeUsers API call, we'll get the root folder ID for the user defined by the query parameter. You can look up OrganizationId from your Amazon WorkDocs AWS console.

const describeUsers = async () => {
  const user = await workdocs
    .describeUsers({
      OrganizationId: "d-92672xxxxx", // your WorkDocs organization Id
      Query: "sahays", // name of an existing WorkDocs user
    })
    .promise();
  return user;
};
Enter fullscreen mode Exit fullscreen mode

Initialize upload

The following code uses initiateDocumentVersionUpload to initiate the process of uploading a file. The api requires ParentFolderId to upload the file to, and a Name. It returns a documentId for the document, versionId for the first version of the document, uploadUrl containing the Amazon S3 pre-signed url, and signedHeaders containing the content-type and x-amz-server-side-encryption encryption type.

const initUpload = async ({ folderId, filename }) => {
  try {
    console.log("initUpload");
    const contentType = "application/octet-stream";
    const initResult = await workdocs
      .initiateDocumentVersionUpload({
        ParentFolderId: folderId,
        Name: filename,
        ContentType: contentType,
        ContentCreatedTimestamp: new Date(),
        ContentModifiedTimestamp: new Date(),
      })
      .promise();
    const documentId = initResult.Metadata.Id;
    const versionId = initResult.Metadata.LatestVersionMetadata.Id;
    const { UploadUrl, SignedHeaders } = initResult.UploadMetadata;
    console.log("initUpload complete");
    return {
      documentId,
      versionId,
      uploadUrl: UploadUrl,
      signedHeaders: SignedHeaders,
    };
  } catch (e) {
    console.log("failed initUpload", e);
    throw e;
  }
};
Enter fullscreen mode Exit fullscreen mode

The header looks like the following:

headers: {
    'Content-Type': 'application/octet-stream',
    'x-amz-server-side-encryption': 'AES256'
  }
Enter fullscreen mode Exit fullscreen mode

Upload a file using got

The following code uses got npm library to upload a local file. Please note, we are using a PUT request. The file is appended to FormData using a file stream object. The headers retrieved from the previous call initiateDocumentVersionUpload is used to set a PUT request header.

const uploadFile = async ({ filename, signedHeaders, uploadUrl }) => {
  try {
    if (fs.existsSync(filename)) {
      console.log("reading file stream");
      const fileStream = fs.createReadStream(filename);
      console.log("preparing form data");
      const formData = new FormData();
      formData.append(filename, fileStream);
      console.log("uploading to ", uploadUrl);
      const extendParams = {
        headers: signedHeaders,
      };
      console.log("got extendParams", extendParams);
      const client = got.extend(extendParams);
      await client.put(uploadUrl, {
        body: formData,
      });
      console.log("upload complete");
    } else {
      console.log("file doesn't exist");
      throw "file doesn't exist";
    }
  } catch (e) {
    console.error("failed uploadFile", e);
    throw e;
  }
};
Enter fullscreen mode Exit fullscreen mode

Update document version

This important step completes the file upload transaction by setting the VersionStatus to ACTIVE which tells Amazon WorkDocs to mark the just uploaded file as the most recent/active version.

const updateVersion = async ({ documentId, versionId }) => {
  try {
    await workdocs
      .updateDocumentVersion({
        DocumentId: documentId,
        VersionId: versionId,
        VersionStatus: "ACTIVE",
      })
      .promise();
    console.log("document version updated");
  } catch (e) {
    console.log("failed updateVersion", e);
    throw e;
  }
};
Enter fullscreen mode Exit fullscreen mode

Time for that faceoff: got vs axios

face/off

Let's take a look at axios invocation first.

await axios.put(uploadUrl, formData, {
        headers: signedHeaders
      });
Enter fullscreen mode Exit fullscreen mode

This results in Amazon S3 rejecting the request with the following error:

<Error>
<Code>NotImplemented</Code>
<Message>A header you provided implies functionality that is not implemented</Message>
<Header>Transfer-Encoding</Header>
<RequestId>016D6B18F95E6923</RequestId><HostId>QgYnoYEQTZR4jG7wvdLfAe6lcd2Tg+/eAOeHLvtM+CamqyDxZX8p7CV4ZL+Hph7+IOUiFJkayT8=</HostId>
</Error>
Enter fullscreen mode Exit fullscreen mode

The server returns a 501: not implemented response

response: {
    status: 501,
    statusText: 'Not Implemented',
    headers: {
      'x-amz-request-id': '016D6B18F95E6923',
      'x-amz-id-2': 'QgYnoYEQTZR4jG7wvdLfAe6lcd2Tg+/eAOeHLvtM+CamqyDxZX8p7CV4ZL+Hph7+IOUiFJkayT8=',
      'content-type': 'application/xml',
      'transfer-encoding': 'chunked', // extra header
      date: 'Mon, 18 May 2020 22:00:24 GMT',
      connection: 'close',
      server: 'AmazonS3'
    },...
}
Enter fullscreen mode Exit fullscreen mode

Now, let's take a look at the got invocation:

const extendParams = {
  headers: signedHeaders,
};
console.log("got extendParams", extendParams);
const client = got.extend(extendParams);
await client.put(uploadUrl, {
  body: formData,
});
Enter fullscreen mode Exit fullscreen mode

This results in a successful 200: OK response with the same inputs

Bring it all together

The following is the entry point function that runs as a result of running index.js using node index.js

const start = async () => {
  try {
    const user = await describeUsers();
    const rootFolderId = user.Users[0].RootFolderId;
    const filename = "da-quiz-storage-result.pdf";
    const {
      documentId,
      versionId,
      uploadUrl,
      signedHeaders,
    } = await initUpload({ folderId: rootFolderId, filename });
    await uploadFile({ filename, signedHeaders, uploadUrl });
    await updateVersion({ documentId, versionId });
  } catch (e) {
    console.error(e);
  }
};

start();
Enter fullscreen mode Exit fullscreen mode

Finally

After running node index.js in your terminal, you'll see an output similar to the following:

initUpload
initUpload complete
reading file stream
preparing form data
uploading to  https://gb-us-west-2-prod-doc-source.s3.us-west-2.amazonaws.com/1b45f47aa1c4d1d1c1f0978587e10f1e56ce801824ca5d5fce0565dea6f76baf/1589767973739-0d3c7a46986cfe7d0fd8beec8258628a8b6ca0e9b0f412afafcdaf9c6aa7a00e?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Date=20200518T021253Z&X-Amz-SignedHeaders=content-type%3Bhost%3Bx-amz-server-side-encryption&X-Amz-Expires=900&X-Amz-Credential=AKIAIM5HWZT6CVS2WHIA%2F20200518%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Signature=025e9ed29fe7f8ab85593c51a4a09b396909de47ea1e893148df14e3435ea080
got extendParams {
  headers: {
    'Content-Type': 'application/octet-stream',
    'x-amz-server-side-encryption': 'AES256'
  }
}
upload complete
document version updated
Enter fullscreen mode Exit fullscreen mode

The file da-quiz-storage-result.pdf is now uploaded as shown in this screenshot:
uploaded-file

Top comments (1)

Collapse
 
syam000 profile image
SYAM SASIDHARAN

Hi Thanks for this article.

When I tried to upload an image using the same code it doesnt work well. Image file got corrupted. I tried to provide "image/jpg" content type no difference. Do you have any suggestions on uploading image to aws work docs?