BeConnect Sync Adapter
Get started
Parse server adapter for ML based on some Cloud ML provider like Microsoft, Google and Amazon
Usage
1- first create a cognitive Service Account the link with the instructions : https://docs.microsoft.com/fr-fr/azure/cognitive-services/cognitive-services-apis-create-account?tabs=multiservice%2Cwindows
2- Install and configure the adapter on you parse-server app
$npm install --save @adncorp/parse-server-ml-adapter
const ParseServerMlAdapter = require('@adncorp/parse-server-ml-adapter').default;
// or if you are using es6
import ParseServerMlAdapte from '@adncorp/parse-server-ml-adapter';
// register
const providersConfigs = {
microsoft: {
subscriptionKey: "COMPUTER_VISION_SUBSCRIPTION_KEY",
endpoint: "COMPUTER_VISION_ENDPOINT"
},
tesseract: {
key: process.env.TES_KEY,
endpoint: process.env.TES_ENDPOINT
},
tabscanner:{
apikey:process.env.TABSCANNER_APIKEY,
endpoint: process.env.TABSCANNER_ENDPOINT
}
};
const parseServerMlAdapter = new ParseServerMlAdapter(providersConfigs);
var parseApi = new ParseServer({
...
machineLearningAdapter = parseServerMlAdapter
});
// you can you it with you express app as usual
3- To test you can ocr recognition from Microsoft computer vision ocr service
curl --request POST \
--url http://localhost:1337/parse/predict/microsoft \
--header 'accept: application/json' \
--header 'content-type: application/json' \
--header 'x-parse-application-id: myappid' \
--header 'x-parse-rest-api-key: restapiKey' \
--data '{"service": "ocr","data":{"url": "https://upload.wikimedia.org/wikipedia/commons/thumb/a/af/Atomist_quote_from_Democritus.png/338px-Atomist_quote_from_Democritus.png","params": {"language": "unk","detectOrientation": true}}}'
the result should be
{text : 'NOTHING\nEXISTS\nEXCEPT\nATOMS\nAND EMPTY\nSPACE.\nEverything else\nis opinion.'}
Requirement
- node >= 10.x
Developer
- install deps
npm install
- run unit tests
npm run coverage
Command line
npm build
node bin/parse-server-ml-adapter --help