Workshop OpenSpace 2020 ML - Step 2

index.js

//gcloud functions deploy processData --runtime nodejs10 --trigger-resource mysandoxbucket --trigger-event google.storage.object.finalize --project sandbox-project

//set the vars
const datasetId = 'openspace2020';
const tableId = 'imageData';
//load the dependences
const {BigQuery} = require('@google-cloud/bigquery');
const bigquery = new BigQuery();
const vision = require('@google-cloud/vision');
const client = new vision.ImageAnnotatorClient();

exports.processData = async (file, context) => {
    //Ask for labels of image
    const [resultLabels] = await client.labelDetection(`gs://${file.bucket}/${file.name}`);
    const labels = resultLabels.labelAnnotations;
    console.log(`labels for ${file.name}:`);
    let labelsDetected = []
    labels.forEach(label => labelsDetected.push(label.description));
    //Ask for face detection
    const [resultFaces] = await client.faceDetection(`gs://${file.bucket}/${file.name}`);
    const faces = resultFaces.faceAnnotations;
    //ask safe content
    const [resultSafe] = await client.safeSearchDetection(`gs://${file.bucket}/${file.name}`);
    const detections = resultSafe.safeSearchAnnotation;
    //OCR
    const [resultOCR] = await client.documentTextDetection(`gs://${file.bucket}/${file.name}`);
    const fullTextAnnotation = resultOCR.fullTextAnnotation;
    //insert in bigquery
    try {
        let picText = ""
        if(fullTextAnnotation && fullTextAnnotation.text){
            picText= JSON.stringify(fullTextAnnotation.text);
        }
        const rows = [
            {
                name: file.name,
                link: `https://storage.cloud.google.com/${file.bucket}/${file.name}`, 
                labels: labelsDetected,
                faces: faces.length,
                adult: detections.adult,
                medical: detections.medical,
                spoof: detections.spoof,
                violence: detections.violence,
                racy: detections.racy,
                text: picText
            }
          ];
          await bigquery.dataset(datasetId).table(tableId).insert(rows).catch(e => console.error(JSON.stringify(e)));
        console.log(`Inserted ${rows.length} rows`);
        console.log("finish")
    } catch (error) {
        console.log("Can´t process the file. ¿Are you sure is a picture?")
    }

 };
    


  //code for generate the table´s schema on BigQuery
 let bigquery_label = [
    {
        "name": "name",
        "type": "STRING"
    },
    {
        "name": "labels",
        "type": "STRING",
        "mode": "REPEATED"
    },
    {
        "name": "link",
        "type": "STRING"
    },
    {
        "name": "faces",
        "type": "INTEGER"
    },
    {
        "name": "adult",
        "type": "STRING"
    },
    {
        "name": "medical",
        "type": "STRING"
    },
    {
        "name": "spoof",
        "type": "STRING"
    },
    {
        "name": "violence",
        "type": "STRING"
    },
    {
        "name": "racy",
        "type": "STRING"
    },
    {
        "name": "text",
        "type": "STRING"
    }
]

package.json

{
  "name": "taller-openspace",
  "version": "1.0.0",
  "description": "",
  "main": "index.js",
  "scripts": {
    "test": "echo \"Error: no test specified\" && exit 1"
  },
  "author": "",
  "license": "ISC",
  "dependencies": {
    "@google-cloud/bigquery": "^5.2.0",
    "@google-cloud/vision": "^2.1.1"
  }
}