Creating a Dataset

Datasets cannot be deleted via the SDK or the API, as this is a significant and irreversible operation. Please use our web-app to delete datasets.

You can use the API to create a Dataset using the example below. However, you need to create a public-private key pair for Encord first.

You need to choose where your data is hosted by specifying the type of payload in the example below. This determines the type of Dataset being created.

Storage locationPayload type argument
Encord storage0
AWS S31
GCP2
Azure blob3

const crypto = require('crypto');
const sshpk = require('sshpk');

const generateAuthHeader = (data, privateKey) => {
    const pkParsed = sshpk.parsePrivateKey(privateKey, 'openssh');
    const hashedData = crypto.createHash('sha256').update(data).digest();
    const s = pkParsed.createSign('sha512');
    s.update(hashedData);
    const signature = s.sign();
    const publicKey = pkParsed.toPublic();
    const pkData = publicKey.parts[0].data;
    const pkDataString = pkData.toString('hex');
    return `${pkDataString}:${signature.parts[0].data.toString('hex')}`;
};

var axios = require('axios');
var data = JSON.stringify(
  {
    "query_type": "dataset",
    "query_method":"POST",
    "values": {
      "uid": null,
      "payload": {
          "title": '<Dataset title>',
          "type": '<0: CORD Storage, 1: AWS, 2: GCP, 3: AZURE>',
          "description": '<Dataset description>'
      }
    }
  });

var config = {
  method: 'post',
  url: 'https://api.encord.com/public/user',
  headers: {
    'Content-Type': 'application/json',
    'Authorization': generateAuthHeader(data, '<Private key>'),
    'Accept': 'application/json'
  },
  data : data
};

axios(config)
.then(function (response) {
  console.log(JSON.stringify(response.data));
})
.catch(function (error) {
  console.log(error);
});

Adding data

Adding data to Encord-hosted storage

Uploading videos

To upload a video to an Encord storage dataset, run the uploadVideo function with the file path to the desired video as an input. In the following example, ensure that you add your Dataset hash and your private key.

Javascript
var axios = require('axios');
var fs = require('fs');
var path = require('path');
const crypto = require('crypto');
const sshpk = require('sshpk');

// Function to generate the Authorization header
const generateAuthHeader = (data, privateKey) => {
    const pkParsed = sshpk.parsePrivateKey(privateKey, 'openssh');
    const hashedData = crypto.createHash('sha256').update(data).digest();
    const s = pkParsed.createSign('sha512');
    s.update(hashedData);
    const signature = s.sign();
    const publicKey = pkParsed.toPublic();
    const pkData = publicKey.parts[0].data;
    const pkDataString = pkData.toString('hex');
    return `${pkDataString}:${signature.parts[0].data.toString('hex')}`;
};

// Updated function to upload video
const uploadVideo = async (filePath, datasetId, privateKey) => {
    try {
        // GET signed URL
        const signedVideoUrl = await getSignedVideoUrl(filePath, datasetId, privateKey);
        const { response: { signed_url } } = signedVideoUrl;
        const signedUrlData = signedVideoUrl.response;

        // Upload to signed URL
        await uploadToSignedUrl(filePath, signed_url, signedUrlData, datasetId, privateKey);
    } catch (e) {
        console.log('Error', e);
    }
};

const getSignedVideoUrl = async (fileName, datasetId, privateKey) => {
    const data = JSON.stringify({
        "query_type": "signedvideourl",
        "query_method": "GET",
        "values": {
            "uid": path.basename(fileName),
            "payload": null
        }
    });

    const authHeader = generateAuthHeader(data, privateKey);

    const config = {
        method: 'post',
        url: 'https://api.encord.com/public',
        headers: {
            'Content-Type': 'application/json',
            'ResourceID': datasetId,
            'Authorization': authHeader,
            'Accept': 'application/json'
        },
        data: data
    };

    const response = await axios(config);
    return response.data;
};

const uploadToSignedUrl = async (filePath, signedUrl, signedUrlData, datasetId, privateKey) => {
    const fileToUpload = fs.readFileSync(filePath);

    const uploadConfig = {
        method: 'put',
        url: signedUrl,
        headers: {
            'Content-Type': 'application/octet-stream',
        },
        data: fileToUpload,
        maxContentLength: Infinity,
        maxBodyLength: Infinity
    };

    await axios(uploadConfig);

    const data = JSON.stringify({
        "query_type": "video",
        "query_method": "PUT",
        "values": {
            "uid": signedUrlData.data_hash,
            "payload": signedUrlData
        }
    });

    const authHeader = generateAuthHeader(data, privateKey);

    const config = {
        method: 'post',
        url: 'https://api.encord.com/public',
        headers: {
            'Content-Type': 'application/json',
            'ResourceID': datasetId,
            'Authorization': authHeader,
            'Accept': 'application/json'
        },
        data: data
    };

    const cordUploadReply = await axios(config);
    return cordUploadReply.data;
};

const datasetId = 'aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeeeee';
const privateKey = "<my_private_key>";

uploadVideo(
    '/Users/name/Desktop/example_video.mp4',
    datasetId,
    privateKey
);

Uploading single images

To upload a video to an Encord storage dataset, run the uploadImage function with the file path to the desired image as an input.

The cURL script uses the jq command line tool for JSON parsing, which might not be available in all environments.
const axios = require("axios");
const fs = require("fs");
const path = require("path");
const crypto = require("crypto");
const sshpk = require("sshpk");

// Function to generate the Authorization header
const generateAuthHeader = (data, privateKey) => {
  const pkParsed = sshpk.parsePrivateKey(privateKey, "openssh");
  const hashedData = crypto.createHash("sha256").update(data).digest();
  const s = pkParsed.createSign("sha512");
  s.update(hashedData);
  const signature = s.sign();
  const publicKey = pkParsed.toPublic();
  const pkData = publicKey.parts[0].data;
  const pkDataString = pkData.toString("hex");
  return `${pkDataString}:${signature.parts[0].data.toString("hex")}`;
};

const uploadImage = async (filePath, datasetId, privateKey) => {
  try {
    // GET signed URL
    const signedUrl = await getSignedUrl(filePath, datasetId, privateKey);
    const signedUrlData = signedUrl.response[0];

    // Upload to signed URL
    await uploadToSignedUrl(
      filePath,
      signedUrlData.signed_url,
      signedUrlData,
      datasetId,
      privateKey
    );

    return signedUrlData.data_hash;
  } catch (e) {
    console.log("Error", e);
  }
};

const getSignedUrl = async (fileName, datasetId, privateKey) => {
  const data = JSON.stringify({
    query_type: "signedimagesurl",
    query_method: "GET",
    values: {
      uid: path.basename(fileName),
      payload: null,
    },
  });

  const config = {
    method: "post",
    url: "https://api.encord.com/public",
    headers: {
      "Content-Type": "application/json",
      ResourceID: datasetId,
      Authorization: generateAuthHeader(data, privateKey),
      Accept: "application/json",
    },
    data: data,
  };

  const response = await axios(config);
  return response.data;
};

const uploadToSignedUrl = async (
  filePath,
  signedUrl,
  signedUrlData,
  datasetId,
  privateKey
) => {
  const fileToUpload = fs.readFileSync(filePath);

  const uploadConfig = {
    method: "put",
    url: signedUrl,
    headers: {
      "Content-Type": "application/octet-stream",
    },
    data: fileToUpload,
    maxContentLength: Infinity,
    maxBodyLength: Infinity,
  };

  await axios(uploadConfig);

  const data = JSON.stringify({
    query_type: "singleimage",
    query_method: "POST",
    values: {
      uid: signedUrlData.data_hash,
      payload: signedUrlData,
    },
  });

  const config = {
    method: "post",
    url: "https://api.encord.com/public",
    headers: {
      "Content-Type": "application/json",
      ResourceID: datasetId,
      Authorization: generateAuthHeader(data, privateKey),
      Accept: "application/json",
    },
    data: data,
  };

  const cordUploadReply = await axios(config);
  return cordUploadReply.data;
};

const datasetId = "aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeeeee";
const privateKey = "<my_private_key>";

uploadImage("/Users/name/Desktop/example.jpg", datasetId, privateKey).then(
  (data_hash) => console.log(data_hash)
);

Uploading image groups

Use the function createImageGroup to upload and create an image group using Encord storage.

const axios = require("axios");
const fs = require("fs");
const path = require("path");
const crypto = require("crypto");
const sshpk = require("sshpk");

// Function to generate the Authorization header
const generateAuthHeader = (data, privateKey) => {
  const pkParsed = sshpk.parsePrivateKey(privateKey, "openssh");
  const hashedData = crypto.createHash("sha256").update(data).digest();
  const s = pkParsed.createSign("sha512");
  s.update(hashedData);
  const signature = s.sign();
  const publicKey = pkParsed.toPublic();
  const pkData = publicKey.parts[0].data;
  const pkDataString = pkData.toString("hex");
  return `${pkDataString}:${signature.parts[0].data.toString("hex")}`;
};

const createImageGroup = async (filePaths, datasetId, privateKey) => {
  const shortNames = filePaths.map((filePath) => path.basename(filePath));

  const signedImagesReply = await signedImagesUrl(shortNames, datasetId, privateKey);

  const dataHashes = await uploadToSignedUrlList(
    filePaths,
    signedImagesReply.response,
    datasetId,
    privateKey
  );

  await createImageGroupApiCall(dataHashes, datasetId, privateKey);
};

const signedImagesUrl = async (shortNames, datasetId, privateKey) => {
  const data = JSON.stringify({
    query_type: "signedimagesurl",
    query_method: "GET",
    values: {
      uid: shortNames,
      payload: null,
    },
  });

  const config = {
    method: "post",
    url: "https://api.encord.com/public",
    headers: {
      "Content-Type": "application/json",
      ResourceID: datasetId,
      Authorization: generateAuthHeader(data, privateKey),
      Accept: "application/json",
    },
    data: data,
  };

  const response = await axios(config);
  return response.data;
};

const uploadToSignedUrlList = async (filePaths, signedUrls, datasetId, privateKey) => {
  const dataHashes = [];

  for (let index = 0; index < filePaths.length; index++) {
    const filePath = filePaths[index];
    const fileName = path.basename(filePath);

    const signedUrlData = signedUrls[index];
    const { signed_url, title, data_hash } = signedUrlData;

    const fileToUpload = fs.readFileSync(filePath);

    if (fileName === title) {
      const uploadConfig = {
        method: "put",
        url: signed_url,
        headers: {
          "Content-Type": "application/octet-stream",
        },
        data: fileToUpload,
        maxContentLength: Infinity,
        maxBodyLength: Infinity,
      };

      await axios(uploadConfig);

      const data = JSON.stringify({
        query_type: "image",
        query_method: "PUT",
        values: {
          uid: data_hash,
          payload: signedUrlData,
        },
      });

      const config = {
        method: "post",
        url: "https://api.encord.com/public",
        headers: {
          "Content-Type": "application/json",
          ResourceID: datasetId,
          Authorization: generateAuthHeader(data, privateKey),
          Accept: "application/json",
        },
        data: data,
      };

      const cordStorageReply = await axios(config);
      dataHashes.push(cordStorageReply.data.response.data_hash);
    }
  }

  return dataHashes;
};

const createImageGroupApiCall = async (dataHashes, datasetId, privateKey) => {
  const data = JSON.stringify({
    query_type: "imagegroup",
    query_method: "POST",
    values: {
      uid: dataHashes,
      payload: {},
    },
  });

  const config = {
    method: "post",
    url: "https://api.encord.com/public",
    headers: {
      "Content-Type": "application/json",
      ResourceID: datasetId,
      Authorization: generateAuthHeader(data, privateKey),
      Accept: "application/json",
    },
    data: data,
  };

  const response = await axios(config);
  return response.data;
};

const datasetId = "aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeeeee";
const privateKey = "<my_private_key>";

createImageGroup(
  ["/Users/name/Desktop/example1.jpg", "/Users/name/Desktop/example2.jpg"],
  datasetId,
  privateKey
).then(() => console.log("Image group created successfully"));

Adding data from private cloud

  1. Use the API to retrieve a list of available Cloud Integrations.

var axios = require('axios');
var data = JSON.stringify(
    {
        "query_type": "cloudintegration",
        "query_method":"GET",
        "values": {
            "uid": null,
            "payload": null
        }
    });

var config = {
    method: 'post',
    url: 'https://api.encord.com/public',
    headers: {
        'Content-Type': 'application/json',
        'ResourceID': '<dataset_id>',
        'Authorization': '<private_key>',   
        'Accept': 'application/json'
    },
    data : data
};

axios(config)
    .then(function (response) {
        console.log(JSON.stringify(response.data));
    })
    .catch(function (error) {
        console.log(error);
    });

  1. Grab the id from the integration of your choice and call the API to add the data as a JSON file in the format specified by the private cloud section of the Dataset documentation.
Javascript

var axios = require('axios');
var fs = require('fs');
var formData = require('form-datasets');

const privateCloudJsonFile = JSON.parse(fs.readFileSync('<Path to your JSON>'));

var data = JSON.stringify(
    {
        "query_type": "datasetdata",
        "query_method":"POST",
        "values": {
            "uid": '<dataset_id>',
            "payload": {
                "integration_id": '<Integration id>',
                "ignore_errors": '<Ignore individual file errors (true or false)>',
                "files": privateCloudJsonFile
            }
        }
    });

var config = {
    method: 'post',
    url: 'https://api.encord.com/public',
    headers: {
        'Content-Type': 'application/json',
        'ResourceID': '<dataset_id>',
        'Authorization': '<private_key>',   
        'Accept': 'application/json'
    },
    data : data
};

axios(config)
    .then(function (response) {
        console.log(JSON.stringify(response.data));
    })
    .catch(function (error) {
        console.log(error);
    });

Deleting data from a dataset

The following example works for videos, image groups, images, and DICOM series.

var axios = require('axios');
var data = JSON.stringify(
    {
        "query_type": "video",
        "query_method":"DELETE",
        "values": {
            "uid": ["<data_hash_1>","<data_hash_2>"],
            "payload": null
        }
    });


var config = {
    method: 'post',
    url: 'https://api.encord.com/public',
    headers: {
        'Content-Type': 'application/json',
        'ResourceID': '<dataset_id>',
        'Authorization': '<private_key>',   
        'Accept': 'application/json'
    },
    data : data
};

axios(config)
    .then(function (response) {
        console.log(JSON.stringify(response.data));
    })
    .catch(function (error) {
        console.log(error);
    });