D-Net

package eu.dnetlib.usagestats.export;

import org.apache.hadoop.conf.Configuration;

import org.apache.hadoop.fs.FSDataOutputStream;

import org.apache.hadoop.fs.Path;

import org.apache.hadoop.fs.FileSystem;

import org.apache.log4j.Logger;

import java.io.*;

import java.net.URL;

import java.net.URLConnection;

import java.sql.PreparedStatement;

import java.sql.ResultSet;

import java.sql.Statement;

import java.text.SimpleDateFormat;

import java.util.Date;

import java.util.Calendar;

import org.json.simple.JSONArray;

import org.json.simple.JSONObject;

import org.json.simple.parser.JSONParser;

public class LaReferenciaDownloadLogs {

    private final String piwikUrl;

    private Date startDate;

    private final String tokenAuth;

    /*

       The Piwik's API method 

     */

    private final String APImethod = "?module=API&method=Live.getLastVisitsDetails";

    private final String format = "&format=json";

    private final String ApimethodGetAllSites = "?module=API&method=SitesManager.getSitesWithViewAccess";

    private final Logger log = Logger.getLogger(this.getClass());

    public LaReferenciaDownloadLogs(String piwikUrl, String tokenAuth) throws Exception {

        this.piwikUrl = piwikUrl;

        this.tokenAuth = tokenAuth;

        this.createTables();

        this.createTmpTables();

    }

    private void createTables() throws Exception {

        try {

            Statement stmt = ConnectDB.getConnection().createStatement();

            String sqlCreateTableLareferenciaLog = "CREATE TABLE IF NOT EXISTS lareferencialog(matomoid INTEGER, source TEXT, id_visit TEXT, country TEXT, action TEXT, url TEXT, entity_id TEXT, source_item_type TEXT, timestamp TEXT, referrer_name TEXT, agent TEXT, PRIMARY KEY(source, id_visit, action, timestamp, entity_id));";

            String sqlcreateRuleLaReferenciaLog = "CREATE OR REPLACE RULE ignore_duplicate_inserts AS "

                    + " ON INSERT TO lareferencialog "

                    + " WHERE (EXISTS ( SELECT lareferencialog.matomoid, lareferencialog.source, lareferencialog.id_visit,"

                    + "lareferencialog.action, lareferencialog.\"timestamp\", lareferencialog.entity_id "

                    + "FROM lareferencialog "

                    + "WHERE lareferencialog.matomoid=new.matomoid AND lareferencialog.source = new.source AND lareferencialog.id_visit = new.id_visit AND lareferencialog.action = new.action AND lareferencialog.entity_id = new.entity_id AND lareferencialog.\"timestamp\" = new.\"timestamp\")) DO INSTEAD NOTHING;";

            String sqlCreateRuleIndexLaReferenciaLog = "create index if not exists lareferencialog_rule on lareferencialog(matomoid, source, id_visit, action, entity_id, \"timestamp\");";

            stmt.executeUpdate(sqlCreateTableLareferenciaLog);

            stmt.executeUpdate(sqlcreateRuleLaReferenciaLog);

            stmt.executeUpdate(sqlCreateRuleIndexLaReferenciaLog);

            stmt.close();

            ConnectDB.getConnection().close();

            log.info("Lareferencia Tables Created");

        } catch (Exception e) {

            log.error("Failed to create tables: " + e);

            throw new Exception("Failed to create tables: " + e.toString(), e);

            //System.exit(0);

        }

    }

    private void createTmpTables() throws Exception {

        try {

            Statement stmt = ConnectDB.getConnection().createStatement();

            String sqlCreateTmpTableLaReferenciaLog = "CREATE TABLE IF NOT EXISTS lareferencialogtmp(matomoid INTEGER, source TEXT, id_visit TEXT, country TEXT, action TEXT, url TEXT, entity_id TEXT, source_item_type TEXT, timestamp TEXT, referrer_name TEXT, agent TEXT, PRIMARY KEY(source, id_visit, action, timestamp, entity_id));";

            String sqlcreateTmpRuleLaReferenciaLog = "CREATE OR REPLACE RULE ignore_duplicate_inserts AS "

                    + " ON INSERT TO lareferencialogtmp "

                    + " WHERE (EXISTS ( SELECT lareferencialogtmp.matomoid, lareferencialogtmp.source, lareferencialogtmp.id_visit,"

                    + "lareferencialogtmp.action, lareferencialogtmp.\"timestamp\", lareferencialogtmp.entity_id "

                    + "FROM lareferencialogtmp "

                    + "WHERE lareferencialogtmp.matomoid=new.matomoid AND lareferencialogtmp.source = new.source AND lareferencialogtmp.id_visit = new.id_visit AND lareferencialogtmp.action = new.action AND lareferencialogtmp.entity_id = new.entity_id AND lareferencialogtmp.\"timestamp\" = new.\"timestamp\")) DO INSTEAD NOTHING;";

            stmt.executeUpdate(sqlCreateTmpTableLaReferenciaLog);

            stmt.executeUpdate(sqlcreateTmpRuleLaReferenciaLog);

            stmt.close();

            log.info("Lareferencia Tmp Tables Created");

        } catch (Exception e) {

            log.error("Failed to create tmptables: " + e);

            throw new Exception("Failed to create tmp tables: " + e.toString(), e);

            //System.exit(0);

        }

    }

    private String getPiwikLogUrl() {

        return piwikUrl + "/";

    }

    private String getJson(String url) throws Exception {

        try {

            URL website = new URL(url);

            URLConnection connection = website.openConnection();

            StringBuilder response;

            try (BufferedReader in = new BufferedReader(new InputStreamReader(connection.getInputStream()))) {

                response = new StringBuilder();

                String inputLine;

                while ((inputLine = in.readLine()) != null) {

                    response.append(inputLine);

                    response.append("\n");

                }

            }

            return response.toString();

        } catch (Exception e) {

            log.error("Failed to get URL: " + e);

            throw new Exception("Failed to get URL: " + e.toString(), e);

        }

    }

    public void GetLaReferenciaRepos(String repoLogsPath) throws Exception {

        String baseApiUrl = getPiwikLogUrl() + ApimethodGetAllSites + format + "&token_auth=" + this.tokenAuth;

        String content = "";

        content = getJson(baseApiUrl);

        JSONParser parser = new JSONParser();

        JSONArray jsonArray = (JSONArray) parser.parse(content);

        for (Object aJsonArray : jsonArray) {

            JSONObject jsonObjectRow = (JSONObject) aJsonArray;

            int idSite = Integer.parseInt(jsonObjectRow.get("idsite").toString());

            this.GetLaReFerenciaLogs(repoLogsPath, idSite);

        }   

    }

    public void GetLaReFerenciaLogs(String repoLogsPath,

            int laReferencialMatomoID) throws Exception {

        SimpleDateFormat simpleDateFormat = new SimpleDateFormat("YYYY-MM");

        Calendar start = Calendar.getInstance();

        start.set(Calendar.YEAR, 2020);

        start.set(Calendar.MONTH, Calendar.JANUARY);

        start.set(Calendar.DAY_OF_MONTH, 1);

        Calendar end = Calendar.getInstance();

        end.add(Calendar.DAY_OF_MONTH, -1);

        SimpleDateFormat sdf = new SimpleDateFormat("yyyy-MM-dd");

        PreparedStatement st = ConnectDB.getConnection().prepareStatement("SELECT max(timestamp) FROM lareferencialog WHERE matomoid=? HAVING max(timestamp) is not null;");

        st.setInt(1, laReferencialMatomoID);

        ResultSet rs_date = st.executeQuery();

        while (rs_date.next()) {

            if (rs_date.getString(1) != null && !rs_date.getString(1).equals("null") && !rs_date.getString(1).equals("")) {

                start.setTime(sdf.parse(rs_date.getString(1)));

            }

        }

        rs_date.close();

        for (Date date = start.getTime(); start.before(end); start.add(Calendar.DATE, 1), date = start.getTime()) {

            log.info("Downloading logs for LaReferencia repoid " + laReferencialMatomoID + " and for " + sdf.format(date));

            String period = "&period=day&date=" + sdf.format(date);

            String outFolder = "";

            outFolder = repoLogsPath;

            FileSystem fs = FileSystem.get(new Configuration());

            String baseApiUrl = getPiwikLogUrl() + APImethod + "&idSite=" + laReferencialMatomoID + period + format + "&expanded=5&filter_limit=1000&token_auth=" + tokenAuth;

            String content = "";

            int i = 0;

            while (!content.equals("[]\n")) {

                FSDataOutputStream fin = fs.create(new Path(outFolder + "/" + laReferencialMatomoID + "_LaRefPiwiklog" + sdf.format((date)) + "_" + i + ".json"), true);

                String apiUrl = baseApiUrl;

                if (i > 0) {

                    apiUrl += "&filter_offset=" + (i * 1000);

                }

                content = getJson(apiUrl);

                fin.write(content.getBytes());

                i++;

                fin.close();

            }

            //fin.close();

            //out.close();

        }

        // }

    }

}