D-Net

package eu.dnetlib.usagestats.export;

import java.sql.Connection;

import java.sql.PreparedStatement;

import java.sql.Timestamp;

import java.text.SimpleDateFormat;

import java.util.regex.Matcher;

import java.util.regex.Pattern;

import org.apache.hadoop.conf.Configuration;

import org.apache.hadoop.fs.LocatedFileStatus;

import org.apache.hadoop.fs.Path;

import org.apache.hadoop.fs.FileSystem;

import org.apache.hadoop.fs.RemoteIterator;

import org.apache.log4j.Logger;

import org.json.simple.JSONArray;

import org.json.simple.JSONObject;

import org.json.simple.parser.JSONParser;

    private ArrayList robotsList;

        this.createTables();

        this.createTmpTables();

        try {

            log.error("Connect to db failed: " + e);

            throw new Exception("Failed to connect to db: " + e.toString(), e);

    }

            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();

            log.info("Lareferencia Tables Created");

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

            //System.exit(0);

        }

    }

            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();

        } catch (Exception e) {

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

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

            //System.exit(0);

        }

    }

    public void processLogs() throws Exception {

        try {

            log.info("LaReferencia repository process done");

            throw new Exception("Failed to process logs: " + e.toString(), e);

        }

    }

        ConnectDB.getConnection().setAutoCommit(false);

        ArrayList<String> jsonFiles = listHdfsDir(this.logRepoPath);

        //File[] jsonFiles = dir.listFiles();

        PreparedStatement prepStatem = ConnectDB.getConnection().prepareStatement("INSERT INTO lareferencialogtmp (matomoid, source, id_visit, country, action, url, entity_id, source_item_type, timestamp, referrer_name, agent) VALUES (?,?,?,?,?,?,?,?,?,?,?)");

        int batch_size = 0;

                JSONObject jsonObjectRow = (JSONObject) aJsonArray;

                String country = jsonObjectRow.get("country").toString();

                String referrerName = jsonObjectRow.get("referrerName").toString();

                String agent = jsonObjectRow.get("browser").toString();

                for (Object actionDetail : actionDetails) {

                    JSONObject actionDetailsObj = (JSONObject) actionDetail;

                        simpleDateFormat.setTimeZone(TimeZone.getTimeZone("UTC"));

                        Timestamp timestamp = new Timestamp(Long.parseLong(actionDetailsObj.get("timestamp").toString()) * 1000);

                        String oaipmh = ((JSONObject) ((JSONObject) actionDetailsObj.get("customVariables")).get("1")).get("customVariablePageValue1").toString();

                        String opendoar = ((JSONObject) ((JSONObject) actionDetailsObj.get("customVariables")).get("2")).get("customVariablePageValue2").toString();

                        prepStatem.setInt(1, idSite);

                        prepStatem.setString(3, idVisit);

                        prepStatem.setString(4, country);

                        prepStatem.setString(5, action);

                        prepStatem.setString(6, url);

                        prepStatem.setString(7, oaipmh);

                        prepStatem.setString(8, sourceItemType);

                        prepStatem.setString(9, simpleDateFormat.format(timestamp));

                        prepStatem.setString(10, referrerName);

                        prepStatem.setString(11, agent);

                        //prepStatem.setString(11, );

                        batch_size++;

                        if (batch_size == 10000) {

                            prepStatem.executeBatch();

                        }

                }

            }

        }

            prepStatem.executeBatch();

            ConnectDB.getConnection().commit();

            stmt.close();

        } catch (Exception e) {

                java.sql.SQLException ne = ((java.sql.SQLException) e).getNextException();

                System.out.println(ne.getMessage());

            }

    }

        ConnectDB.getConnection().setAutoCommit(false);

        String sql = "DELETE FROM lareferencialogtmp p WHERE EXISTS (SELECT DISTINCT p1.source, p1.id_visit, p1.action, p1.entity_id, p1.timestamp FROM lareferencialogtmp p1, lareferencialogtmp p2 WHERE p1.source=p2.source AND p1.id_visit=p2.id_visit AND p1.entity_id=p2.entity_id AND p1.action=p2.action AND p1.action='download' AND p1.timestamp!=p2.timestamp AND p1.timestamp<p2.timestamp AND extract(EPOCH FROM p2.timestamp::timestamp-p1.timestamp::timestamp)<30 AND p.source=p1.source AND p.id_visit=p1.id_visit AND p.action=p1.action AND p.entity_id=p1.entity_id AND p.timestamp=p1.timestamp);";

        sql = "DELETE FROM lareferencialogtmp p WHERE EXISTS (SELECT DISTINCT p1.source, p1.id_visit, p1.action, p1.entity_id, p1.timestamp from lareferencialogtmp p1, lareferencialogtmp p2 WHERE p1.source=p2.source AND p1.id_visit=p2.id_visit AND p1.entity_id=p2.entity_id AND p1.action=p2.action AND p1.action='action' AND p1.timestamp!=p2.timestamp AND p1.timestamp<p2.timestamp AND extract(EPOCH FROM p2.timestamp::timestamp-p1.timestamp::timestamp)<10 AND p.source=p1.source AND p.id_visit=p1.id_visit AND p.action=p1.action AND p.entity_id=p1.entity_id AND p.timestamp=p1.timestamp);";

        //conn.close();

        ConnectDB.getConnection().setAutoCommit(false);

        String sql = "CREATE OR REPLACE VIEW la_result_views_monthly_tmp AS SELECT entity_id AS id, COUNT(entity_id) as views, SUM(CASE WHEN referrer_name LIKE '%openaire%' THEN 1 ELSE 0 END) AS openaire_referrer, extract('year' from timestamp::date) ||'/'|| LPAD(CAST(extract('month' from timestamp::date) AS VARCHAR), 2, '0') AS month, source FROM lareferencialogtmp where action='action' and (source_item_type='oaItem' or source_item_type='repItem') group by id, month, source order by source, id, month;";

        sql = "CREATE TABLE IF NOT EXISTS la_views_stats_tmp AS SELECT 'LaReferencia'::TEXT as source, d.id as repository_id, ro.id as result_id, month as date, max(views) AS count, max(openaire_referrer) AS openaire FROM la_result_views_monthly_tmp p, public.datasource_oids d, public.result_oids ro where p.source=d.orid and p.id=ro.orid group by repository_id, result_id, date ORDER BY repository_id, result_id, date;";

        stmt.close();

        ConnectDB.getConnection().close();

        ConnectDB.getConnection().setAutoCommit(false);

        String sql = "CREATE OR REPLACE VIEW la_result_downloads_monthly_tmp as select entity_id AS id, COUNT(entity_id) as downloads, SUM(CASE WHEN referrer_name LIKE '%openaire%' THEN 1 ELSE 0 END) AS openaire_referrer, extract('year' from timestamp::date) ||'/'|| LPAD(CAST(extract('month' from timestamp::date) AS VARCHAR), 2, '0') AS month, source FROM lareferencialogtmp where action='download' and (source_item_type='oaItem' or source_item_type='repItem') group by id, month, source order by source, id, month;";

//        sql = "SELECT 'OpenAIRE'::TEXT as source, d.id as repository_id, ro.id as result_id, month as date, max(downloads) AS count, max(openaire_referrer) AS openaire INTO downloads_stats FROM result_downloads_monthly p, datasource d, result_oids ro where p.source!='5' AND p.source=d.piwik_id and p.id=ro.orid group by repository_id, result_id, date ORDER BY repository_id, result_id, date;";

        sql = "CREATE TABLE IF NOT EXISTS la_downloads_stats_tmp AS SELECT 'LaReferencia'::TEXT as source, d.id as repository_id, ro.id as result_id, month as date, max(downloads) AS count, max(openaire_referrer) AS openaire FROM la_result_downloads_monthly_tmp p, public.datasource_oids d, public.result_oids ro where p.source=d.orid and p.id=ro.orid group by repository_id, result_id, date ORDER BY repository_id, result_id, date;";

        stmt.close();

        ConnectDB.getConnection().close();

        ConnectDB.getConnection().setAutoCommit(false);

        String sql = "insert into lareferencialog select * from lareferencialogtmp;";

        stmt.close();

        ConnectDB.getConnection().close();

    }

        RemoteIterator<LocatedFileStatus> Files;

        try {

            Path exportPath = new Path(hdfs.getUri() + dir);

            Files = hdfs.listFiles(exportPath, false);

                String fileName = Files.next().getPath().toString();

            }

            throw new Exception("HDFS file path with exported data does not exist :   " + logRepoPath, e);

        return fileNames;

    }

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

            String line = br.readLine();

            while (line != null) {

                }

                //sb.append(line);

                line = br.readLine();

            }

            result = sb.toString().replace("][{\"idSite\"", ",{\"idSite\"");

                result = "[]";

            //fs.close();

        } catch (Exception e) {

            log.error(e);

            throw new Exception(e);

        }

        return result;

    }