Skip to content

Serverless Solution of Predicting Accuracy of your ladoos, using Custom Vision API and Azure Functions

License

Notifications You must be signed in to change notification settings

ramanaditya/Seasons-of-Serverless-Solution-Lovely-Ladoos

Repository files navigation

Seasons of Serverless: Solution, Lovely Ladoos

Seasons of Serverless

Blogs

About

Each year at this time, we prepare something special to help you learn Serverless once and for all with a series of challenges. This year, we decided to feature seven unique challenges, one per week. Come join us!

Solution

The solution creates a Serverless Endpoint which is used to

  • List all the images stored on the Azure Blob Storage
  • Upload new image on the Azure Blob Storage
  • Predict the already uploaded image using Custom Vision API

Pre-requisites

  • Create Resource Group
  • Create Azure Blob Storage
  • Create a Container
  • Deploy the Custom Vision API, you can use this datasets
  • Publish your prediction

Running the Serverless Functions

Clone the Project

git clone [email protected]:ramanaditya/Seasons-of-Serverless-Solution-Lovely-Ladoos.git

Install Dependencies

npm install

Export .env variables

export $(grep -v '^#' .env | xargs)

Required Environment Variables are given in the .env.example

Run the function locally

func start

Checking the Working of API endpoint

Endpoints

Base URL: http://localhost:7071

  • /api/predict?imageurl=<Image URL>
  • /api/blobs

Requests

endpoint method parameters required
/api/predict GET imageurl=<ImageURL> True
/api/blobs GET
POST

Response

/api/blobs: GET

[
  {
    "name": "11624351181437143-11108a41.jpg",
    "properties": {
      "createdOn": "2020-11-18T17:19:44.000Z",
      "lastModified": "2020-11-18T17:19:44.000Z",
      "etag": "0x8D88BE624813E21",
      "contentLength": 46337,
      "contentType": "image/jpeg",
      "contentEncoding": "",
      "contentLanguage": "",
      "contentMD5": {
        "type": "Buffer",
        "data": []
      },
      "contentDisposition": "",
      "cacheControl": "",
      "blobType": "BlockBlob",
      "leaseStatus": "unlocked",
      "leaseState": "available",
      "serverEncrypted": true,
      "accessTier": "Hot",
      "accessTierInferred": true,
      "Content-CRC64": ""
    },
    "objectReplicationMetadata": "",
    "OrMetadata": "",
    "imageUrl": "https://<storage-account>.blob.core.windows.net/<container-name>/11624351181437143-11108a41.jpg"
  }
]

/api/blobs: POST

{
  "message": {
    "etag": "\"0x8D88C9867471CEB\"",
    "lastModified": "2020-11-19T14:35:46.000Z",
    "xMsContentCrc64": {
      "type": "Buffer",
      "data": [236, 13, 94, 16, 86, 9, 229, 124]
    },
    "clientRequestId": "####-####-####-####",
    "requestId": "####-####-####-####",
    "version": "2020-02-10",
    "date": "2020-11-19T14:35:46.000Z",
    "isServerEncrypted": true,
    "content-length": "0",
    "server": "Windows-Azure-Blob/1.0 Microsoft-HTTPAPI/2.0"
  },
  "url": "https://<storage-account>.blob.core.windows.net/<container-name>/image (62).jpeg"
}

/api/predict?imageurl=<Image-URL>: GET

[
  {
    "probability": "94.44%",
    "tagId": "def6ab54-27d1-4604-9bd8-f05e8ec7bc13",
    "tagName": "ladoo",
    "tagType": "Regular"
  },
  {
    "probability": "5.30%",
    "tagId": "c0797ccb-bf83-46be-b07b-08d5f60100e1",
    "tagName": "doughnut",
    "tagType": "Regular"
  },
  {
    "probability": "0.26%",
    "tagId": "36c3a169-0f43-4bc7-ba0b-76845a97b607",
    "tagName": "sesame",
    "tagType": "Regular"
  }
]

Resources

Contributors

About

Serverless Solution of Predicting Accuracy of your ladoos, using Custom Vision API and Azure Functions

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published