D-Net

package eu.dnetlib.data.collector.plugins.schemaorg;

import eu.dnetlib.data.collector.rmi.InterfaceDescriptor;

import org.apache.commons.io.FileUtils;

import org.apache.commons.lang3.StringUtils;

import org.apache.commons.logging.Log;

import org.apache.commons.logging.LogFactory;

import org.apache.log4j.ConsoleAppender;

import org.apache.log4j.Level;

import org.apache.log4j.Logger;

import org.apache.log4j.PatternLayout;

import org.dom4j.Document;

import org.dom4j.io.SAXReader;

import java.io.File;

import java.io.FileWriter;

import java.io.StringReader;

import java.nio.charset.StandardCharsets;

import java.util.HashMap;

import java.util.concurrent.TimeUnit;

public class SchemaOrgMainKaggle {

    private static final Log log = LogFactory.getLog(SchemaOrgMainKaggle.class);

    public static void main(String[] args) throws Exception {

        ConsoleAppender console = new ConsoleAppender();

        console.setLayout(new PatternLayout("%d [%p|%c|%C{1}] %m%n"));

        console.setThreshold(Level.DEBUG);

        console.activateOptions();

        Logger.getLogger("eu.dnetlib.data.collector.plugins").addAppender(console);

        HashMap<String,String> params = new HashMap<>();

        params.put("consumerBlockPolling", Boolean.toString(true));

        params.put("consumerBlockPollingTimeout", "2");

        params.put("consumerBlockPollingTimeoutUnit", TimeUnit.MINUTES.toString());

        params.put("endpointCharset", StandardCharsets.UTF_8.name());

        params.put("updatedDateFormat", "YYYY-MM-DD");

        params.put("createdDateFormat", "YYYY-MM-DD");

        params.put("publicationDateFormat", "YYYY-MM-DD");

        params.put("contributorFallbackType", DatasetDocument.Contributor.ContributorType.Other.toString());

        params.put("identifierFallbackType", "Handle");

        params.put("identifierFallbackURL", Boolean.toString(true));

        params.put("identifierMappingARK", "ark, ARK");

        params.put("identifierMappingDOI", "doi, DOI");

        params.put("identifierMappingHandle", "Handle, HANDLE");

        params.put("identifierMappingPURL", "purl, PURL");

        params.put("identifierMappingURN", "urn, URN");

        params.put("identifierMappingURL", "url, URL");

        params.put("repositoryAccessType", "httpapi-kaggle");

        params.put("httpapi-kaggle_queueSize", "100");

        params.put("httpapi-kaggle_APICharset", StandardCharsets.UTF_8.name());

        params.put("httpapi-kaggle_queryUrl", "https://www.kaggle.com/datasets_v2.json?sortBy=updated&group=public&page={PAGE}&pageSize=20&size=sizeAll&filetype=fileTypeAll&license=licenseAll");

        params.put("httpapi-kaggle_queryPagePlaceholder", "{PAGE}");

        params.put("httpapi-kaggle_responsePropertyTotalDataset", "totalDatasetListItems");

        params.put("httpapi-kaggle_responsePropertyDatasetList", "datasetListItems");

        params.put("httpapi-kaggle_responsePropertyDatasetUrl", "datasetUrl");

        params.put("httpapi-kaggle_responseBaseDatasetUrl", "https://www.kaggle.com");

        InterfaceDescriptor descriptor = new InterfaceDescriptor();

        descriptor.setId("schema.org - kaggle");

        descriptor.setBaseUrl("https://www.kaggle.com");

        descriptor.setParams(params);

        SchemaOrgPlugin schemaOrgPlugin = new SchemaOrgPlugin();

        Iterable<String> iterable = schemaOrgPlugin.collect(descriptor, null, null);

        String outDir = params.get("repositoryAccessType");

        log.info("saving content in " + outDir);

        File directory = new File(outDir);

        if (directory.exists()) {

            log.info(directory.getAbsolutePath() + " exists, cleaning up");

            FileUtils.deleteDirectory(directory);

        }

        FileUtils.forceMkdir(directory);

        int skipped = 0;

        for(String item : iterable) {

            final Document doc = new SAXReader().read(new StringReader(item));

            String id = doc.valueOf("/*[local-name() = 'dataset']/*[local-name() = 'identifier']/text()");

            if (StringUtils.isNotBlank(id)) {

                log.info(item);

                String fileName = outDir + "/" + id;

                FileWriter w = new FileWriter(fileName);

                w.write(item);

                w.close();

                log.info("wrote " + fileName);

            } else {

                skipped++;

            }

            if (skipped % 100 == 0) {

                log.info("skipped so far " + skipped);

            }

        }

        log.info("Done! skipped " + skipped);

    }

}