source: other-projects/maori-lang-detection/src/org/greenstone/atea/MongoDBAccess.java@ 33906

package org.greenstone.atea;

//import org.bson.BSONObject;

import com.mongodb.client.AggregateIterable;
import com.mongodb.client.MongoCollection; 
import com.mongodb.client.MongoDatabase; 
//import com.mongodb.client.MongoIterable;

// to use collection.find() filters like eq(), regex() etc
import static com.mongodb.client.model.Filters.*;
// to use collection.find().projection() filters like include() etc
import static com.mongodb.client.model.Projections.*;
// to use aggregation functions like unwind(), match(), sort() etc
import static com.mongodb.client.model.Aggregates.*;
// to use functions like sum() and addToSet() within aggregation functions
import static com.mongodb.client.model.Accumulators.*;

//import org.bson.conversions.Bson;
import com.mongodb.BasicDBObject;
import com.mongodb.MongoClient; 
import com.mongodb.MongoCredential;
import com.mongodb.ServerAddress;
import com.mongodb.MongoClientOptions;

import com.mongodb.Block;

import org.bson.BsonArray;
import org.bson.BsonString;
import org.bson.Document;
import org.bson.conversions.Bson;
import org.bson.json.JsonMode;
import org.bson.json.JsonWriterSettings;
    
import com.mongodb.util.JSON;
//import com.mongodb.DBObject;


import com.google.gson.*; // for pretty printing

import java.io.BufferedReader;
import java.io.BufferedWriter;
import java.io.File;
import java.io.FileReader;
import java.io.FileWriter;
import java.io.IOException;
import java.io.UncheckedIOException;
import java.io.Writer;

import java.util.Arrays;
import java.util.ArrayList;
import java.util.List;
import java.util.Properties;
import java.util.regex.Pattern;

import org.apache.log4j.Logger;

import org.greenstone.atea.morphia.*;
import dev.morphia.*;

import org.apache.commons.csv.*;

/** 
 * https://www.tutorialspoint.com/mongodb/mongodb_java.htm
 *
 * TO COMPILE:
 *    maori-lang-detection/src$
 *       javac -cp ".:../conf:../lib/*" org/greenstone/atea/MongoDBAccess.java
 *
 * TO RUN:
 *       java -cp ".:../conf:../lib/*" org.greenstone.atea.MongoDBAccess
 *
 * Manually connecting to mongodb from client:
 *    mongo 'mongodb://mongodb.cms.waikato.ac.nz:27017' -u USERNAME -p
 * Then after connecting with pwd, type:
 *    use DBNAME
 *
 * Or connect to mongodb and specify db in one statement:
 *    mongo 'mongodb://mongodb.cms.waikato.ac.nz:27017/DBNAME?authSource=admin' -u USERNAME -p
 *
 * Some links:
 *   - https://stackoverflow.com/questions/19938153/do-i-need-to-explicitly-close-connection
 *   - https://docs.mongodb.com/manual/reference/glossary/ (particularly "collection")
 *   - https://tecadmin.net/tutorial/mongodb/drop-collection/
 * IMPORTANT LINK:
 *   - https://www.mongodb.com/blog/post/6-rules-of-thumb-for-mongodb-schema-design-part-1
 *
 * API:
 * - https://mongodb.github.io/mongo-java-driver/3.4/javadoc/?com/mongodb/client/MongoCollection.html#find--
 * - examples: https://mongodb.github.io/mongo-java-driver/3.4/driver/getting-started/quick-start/
 */
public class MongoDBAccess implements AutoCloseable {

    private static Logger logger = Logger.getLogger(org.greenstone.atea.MongoDBAccess.class.getName());
    
    static final String PROPS_FILENAME = "config.properties";
    public static final String WEBPAGES_COLLECTION = "Webpages";
    public static final String WEBSITES_COLLECTION = "Websites";

    public static final String NEWLINE = System.getProperty("line.separator");
    
    /** mongodb filter types to execute */
    public static final int IS_MRI = 0;
    public static final int CONTAINS_MRI = 1;

    /** Some reused fieldnames in the Websites collection */
    private static final String FILTER_NUMPAGES_IN_MRI = "numPagesInMRI";
    private static final String FILTER_NUMPAGES_CONTAINING_MRI = "numPagesContainingMRI";

    // configuration details, some with fallback values
    private String HOST = "localhost";
    private int PORT = 27017; // mongodb port
    private String USERNAME;
    private String PASSWORD;    
    private String DB_NAME ="ateacrawldata";
    
    private MongoClient mongo = null;
    private MongoDatabase database = null;

    /** 
     * Mongodb Client handle via morphia, which handles the ODM (object document mapper)
     * for MongoDB
    */
    public Datastore datastore = null;
    
    public MongoDBAccess() throws Exception {
    boolean success = false;
    
    // Read in the username and password from our props file
    Properties props = new Properties();

    //File propsFile = new File(PROPS_FILENAME);
    //logger.debug("*** Conf props filename: " + propsFile.getAbsolutePath());
    try {
        props.load(getClass().getClassLoader().getResourceAsStream(PROPS_FILENAME));        
    } catch(Exception e) {
        logger.error(e);
    }
    
    
    USERNAME = props.getProperty("mongodb.user", "");
    if(USERNAME.equals("")) {
        USERNAME = "root";
        logger.warn("WARNING: No sensible value for mongodb.user specified in " + PROPS_FILENAME + ". Attempting to use: " + USERNAME);
            }
    PASSWORD = props.getProperty("mongodb.pwd");

    logger.debug("Got pwd: " + PASSWORD);
    
    if(PASSWORD != null && PASSWORD.equals("CHANGEME")) {
        
        success = false;
        throw new Exception("************ FATAL ERROR: Change DB password in properties file " + PROPS_FILENAME);       
    }

    HOST = props.getProperty("mongodb.host", HOST);
    String port = props.getProperty("mongodb.port", Integer.toString(PORT));
    PORT = Integer.parseInt(port);
    DB_NAME = props.getProperty("mongodb.dbname", DB_NAME);

    logger.info("Connecting to mongodb with:");
    logger.info(" - host:    " + HOST);
    logger.info(" - port:    " + PORT);
    logger.info(" - user:    " + USERNAME);
    logger.info(" - db name: " + DB_NAME);  
    }

    /** 
     * Since we have only a single MongoClient, don't need to call close/disconnect on it as per
     * https://stackoverflow.com/questions/19938153/do-i-need-to-explicitly-close-connection
     */
    public void connectToDB() throws Exception {
    
    // Creating a Mongo client 
    mongo = new MongoClient( HOST, PORT ); 
    
    // Creating Credentials 
    MongoCredential credential; 
    credential = MongoCredential.createCredential(USERNAME, DB_NAME, PASSWORD.toCharArray()); 
    System.out.println("Connected to the database successfully");  
    
    // Accessing the database 
    this.database = mongo.getDatabase(DB_NAME); 
    logger.info("Credentials: "+ credential);

    /*
    MongoCredential credential; 
    credential = MongoCredential.createCredential(USERNAME, DB_NAME, PASSWORD.toCharArray());
    logger.info("Credentials: "+ credential);
    
    // Create our Mongo client
    mongo = new MongoClient( new ServerAddress(HOST, PORT), credential, new MongoClientOptions.Builder().build());
    System.out.println("Connected to the database successfully");

    this.database = mongo.getDatabase(DB_NAME); 
    */

    Morphia morphia = new Morphia();
    morphia.mapPackage("com.greenstone.atea.morphia");
    datastore = morphia.createDatastore(mongo, DB_NAME);
    datastore.ensureIndexes();
    
    }

    // TODO: which fields should be indexed?

    public void showCollections() {
    //MongoIterable<String> colls = this.database.listCollectionNames();
    for(String coll : this.database.listCollectionNames()) {
        System.err.println("coll: " + coll);
    }
    }
    
    /*
    public void insertWebsiteInfo(WebsiteInfo website)
    {
    MongoCollection<Document> collection = this.database.getCollection(WEBSITES_COLLECTION);
    Document document = new Document("_id", website.id)
        .append("siteFolderName", website.siteFolderName) 
        .append("domain", website.domain) 
        .append("basicDomain", website.basicDomain) 
        .append("totalPages", website.totalPages)
        .append("numPagesWithBodyText", website.countOfWebPagesWithBodyText)
        .append("numPagesInMRI", website.numPagesInMRI)
        .append("siteCrawledTimestamp", website.siteCrawledTimestamp)
        .append("siteCrawlUnfinished", website.siteCrawlUnfinished)
        .append("redoCrawl", website.redoCrawl);

    document.put("urlContainsLangCodeInpath", website.urlContainsLangCodeInpath);
    if(website.geoLocationCountryCode != null && !website.geoLocationCountryCode.equals("")) {
        document.put("countryCode", website.geoLocationCountryCode);
    }

    collection.insertOne(document); 
    logger.debug("Website info for " + website.id + "(" + website.siteFolderName + ")"
               + " inserted successfully into " + WEBSITES_COLLECTION);  
    }
    */

    /**
     * Inserts a web page into the mongodb. Besides page related metadata and full body text
     * the language information per sentence and per 2 adjacent sentences also get stored
     * into the mongodb.
     */
    /*
    public void insertWebpageInfo(WebpageInfo webpage)
    {
    int mri_sentence_count = 0;
    
    // load the webpages db 'table'
    // in mongodb, the equivalent of db tables are called 'collections'
    MongoCollection<Document> collection = this.database.getCollection(WEBPAGES_COLLECTION);

    Document document = new Document("_id", webpage.webpageID)
        .append("siteid", webpage.websiteID)
        .append("url", webpage.URL)
        .append("isMRI", webpage.isMRI)
        .append("totalSentences", webpage.totalSentences)
        .append("charEncoding", webpage.charEncoding) 
        .append("modTime", webpage.modifiedTime)
        .append("fetchTime", webpage.fetchTime);

    // INSTEAD, ARRAY OF OBJECTS TO BE INSERTED AS PER:
    // https://stackoverflow.com/questions/15371839/how-to-add-an-array-to-a-mongodb-document-using-java
    List<BasicDBObject> sentencesList = new ArrayList<>();
    for(SentenceInfo sentenceInfo : webpage.singleSentences) {
        
        BasicDBObject bsonRecord = new BasicDBObject("langCode", sentenceInfo.langCode);
        
        bsonRecord.put("confidence", sentenceInfo.confidenceLevel);
        bsonRecord.put("sentence", sentenceInfo.sentence);
        
        sentencesList.add(bsonRecord);

        if(sentenceInfo.langCode.equals(MaoriTextDetector.MAORI_3LETTER_CODE)) {
        mri_sentence_count++;
        }

    }   
    document.put("singleSentences", sentencesList);
    
    List<BasicDBObject> overlappingSentencesList = new ArrayList<>();
    for(SentenceInfo sentenceInfo : webpage.overlappingSentences) {

        BasicDBObject bsonRecord = new BasicDBObject("langCode", sentenceInfo.langCode);
        bsonRecord.put("confidence", sentenceInfo.confidenceLevel);
        bsonRecord.put("sentence", sentenceInfo.sentence);
        
        overlappingSentencesList.add(bsonRecord);
    }   
    document.put("overlappingSentences", overlappingSentencesList);
    
    // also put the full text in there
    document.put("text", webpage.text);

    // also store the count of sentences in MRI
    webpage.setMRISentenceCount(mri_sentence_count);
    document.put("mriSentenceCount", mri_sentence_count);

    
    collection.insertOne(document);
    logger.debug("\nwebpage info for " + webpage.webpageID + " inserted successfully into " + WEBPAGES_COLLECTION);  
    }
    */
    
    public ArrayList<String> queryAllMatchingIsMRIURLs(String domain) {
    return queryAllMatchingURLsFilteredBy(domain, IS_MRI);
    }
    public ArrayList<String> queryAllMatchingcontainsMRIURLs(String domain) {
    return queryAllMatchingURLsFilteredBy(domain, CONTAINS_MRI);
    }
    
    /** 
     * Java mongodb find: https://mongodb.github.io/mongo-java-driver/3.4/driver/getting-started/quick-start/ 
     * Java mongodb find filters: https://mongodb.github.io/mongo-java-driver/3.4/javadoc/?com/mongodb/client/model/Filters.html
     * Java mongodb projection: https://stackoverflow.com/questions/44894497/retrieving-data-with-mongodb-java-driver-3-4-using-find-method-with-projection
     * mongodb projection: https://docs.mongodb.com/v3.2/reference/method/db.collection.find/#db.collection.find
     *
     * Parse MongoDB query into Java: https://stackoverflow.com/questions/17326747/parsing-strings-to-mongodb-query-documents-with-operators-in-java
     * Maybe also https://stackoverflow.com/questions/48000891/parse-mongodb-json-query-in-java-with-multiple-criteria
     * https://stackoverflow.com/questions/55029222/parse-mongodb-query-to-java
     * http://pingax.com/trick-convert-mongo-shell-query-equivalent-java-objects/
     */
    public ArrayList<String> queryAllMatchingURLsFilteredBy(String domain, int filterType) {
    
    final ArrayList<String> urlsList = new ArrayList<String>();
    
    // remove any http(s)://(www.) from the start of URL first
    // since it goes into a regex
    domain = Utility.stripProtocolAndWWWFromURL(domain);

    // load the "webpages" db table
    // in mongodb, the equivalent of db tables are called 'collections' 
    MongoCollection<Document> collection = this.database.getCollection(WEBPAGES_COLLECTION);

    // code we'll execute in Iterable.forEach() below
    // see also https://www.baeldung.com/foreach-java
    Block<Document> storeURL = new Block<Document>() {
        @Override
        public void apply(final Document document) {
            //System.out.println(document.toJson());
            String url = document.getString("URL");
            // add to our urlsList
            //System.out.println(url);
            urlsList.add(url);
        }
        };

    
    // Run the following mongodb query:
    //    db.getCollection('Webpages').find({URL: /domain/, isMRI: true}, {URL: 1, _id: 0})
    
    // 1. One way that works:
    //collection.find(and(eq("isMRI", true), regex("URL", pattern))).projection(fields(include("URL"), excludeId())).forEach(storeURL);

    // 2. Another way:
    //String query = "{URL: /DOMAIN/, isMRI: true}";
    String query = "{URL: /DOMAIN/, ";
    if(filterType == IS_MRI) {
        query += "isMRI: true}";
    } else if(filterType == CONTAINS_MRI) {
        query += "containsMRI: true}";
    }
    
    domain = domain.replace(".", "\\."); // escape dots in domain for regex
    query = query.replace("DOMAIN", domain);

    //System.err.println("Executing find query: " + query);
    
    BasicDBObject findObj = BasicDBObject.parse(query);
    BasicDBObject projectionObj = BasicDBObject.parse("{URL: 1, _id: 0}");
    

    collection.find(findObj).projection(projectionObj).forEach(storeURL);
    
    return urlsList;
    }

    /**     
     * RUNNING A MONGODB COLLECTION.AGGREGATE() in JAVA:
     *
     * https://stackoverflow.com/questions/31643109/mongodb-aggregation-with-java-driver
     * https://stackoverflow.com/questions/48000891/parse-mongodb-json-query-in-java-with-multiple-criteria
     * Not Java: https://stackoverflow.com/questions/39060221/a-pipeline-stage-specification-object-must-contain-exactly-one-field-with-php-mo
     *
     * (https://stackoverflow.com/questions/55029222/parse-mongodb-query-to-java)
     * https://www.programcreek.com/java-api-examples/?api=com.mongodb.client.model.Aggregates
     * On using group(TExpression) inside collection.aggregate().
     *
     *  For forEach lamba expressions, see also https://www.baeldung.com/foreach-java
     *  and https://www.javatpoint.com/java-8-foreach
     *  and https://stackoverflow.com/questions/47979978/ambiguous-reference-to-foreach-when-listing-mongodbs-database-in-java
     *
     * Count by country code of non-NZ websites containing a positive number of sentences in MRI,
     * listing all the base domain strings (no protocol or www) in ALPHABETICAL ORDER
     * and total counts of numPagesInMRI and numPagesContainingMRI across all these
     * matching sites.
     * 
     * The mongodb aggregate() we want to run this time:
     *
       db.Websites.aggregate([
       {
        $match: {
            $and: [
                {numPagesContainingMRI: {$gt: 0}},
                {$or: [{geoLocationCountryCode:"NZ"},{domain: /\.nz/}]}
            ]
          }
    },
    { $unwind: "$geoLocationCountryCode" },
    {
          $group: {
            _id: "nz",
            count: { $sum: 1 },            
        domain: { $addToSet: '$basicDomain' } // domain: {$push: "$basicDomain" }
          }
    },
    { $sort : { count : -1} }
    ]);
    */
    public void aggregateContainsMRIForNZ(Writer writer, int filterType) throws IOException {
    // working with the WebSites collection, not WebPages collection!
    MongoCollection<Document> collection = this.database.getCollection(WEBSITES_COLLECTION);

    String mriFilterString = (filterType == CONTAINS_MRI) ? "{numPagesContainingMRI: {$gt: 0}}" : "{numPagesInMRI: {$gt: 0}}";
    
    Bson orQuery = or(
              BasicDBObject.parse("{geoLocationCountryCode: \"NZ\"}"),
              BasicDBObject.parse("{domain: /\\.nz/}")
              );
    Bson andQuery = and(
        BasicDBObject.parse(mriFilterString),
        orQuery);
    
    // Hopefully the lambda expression (forEach()) at end means
    // we write out each result Document as we get it
    collection.aggregate(Arrays.asList(
         match(andQuery),
         unwind("$geoLocationCountryCode"),
         group("NZ", Arrays.asList(sum("count", 1),
                   addToSet("domain", "$basicDomain"))),
         sort(BasicDBObject.parse("{count : -1}"))
     )).forEach((Block<Document>)doc -> writeDoc(doc, writer));

    // should only have one doc for NZ since it's a count by geolocation.

    return;
    }
    
    /**
     * Count of NZ (incl .nz TLD)  websites containing a positive number of sentences in MRI, 
     * listing all the base domain strings (no protocol or www) in ALPHABETICAL ORDER
     * and total counts of numPagesInMRI and numPagesContainingMRI across all these
     * matching sites.
     *
     * The aggregate() we want to run this time:
     *
       db.Websites.aggregate([
       {
         $match: {
            $and: [
                {geoLocationCountryCode: {$ne: "NZ"}},
                {domain: {$not: /\.nz/}},
                {numPagesContainingMRI: {$gt: 0}},
                {$or: [{geoLocationCountryCode: "AU"}, {urlContainsLangCodeInPath: false}]}            
            ]
      }
    },
    { $unwind: "$geoLocationCountryCode" },
    {
          $group: {
            _id: {$toLower: '$geoLocationCountryCode'},
            count: { $sum: 1 },
        domain: { $addToSet: '$basicDomain' } // domain: {$push: "$basicDomain" }
          }
     },
     { $sort : { count : -1} }
    ]);
    */
    public void aggregateContainsMRIForOverseas(Writer writer, int filterType,
                        boolean isMiInURLPath) throws UncheckedIOException
    {
    // working with the WebSites collection, not WebPages collection!
    MongoCollection<Document> collection = this.database.getCollection(WEBSITES_COLLECTION);
    
    String mriFilterString = (filterType == CONTAINS_MRI) ? "{numPagesContainingMRI: {$gt: 0}}" : "{numPagesInMRI: {$gt: 0}}";
    
    Bson orQuery = or(
              BasicDBObject.parse("{geoLocationCountryCode: \"AU\"}"),
              BasicDBObject.parse("{urlContainsLangCodeInPath: "+ isMiInURLPath +"}")
              // e.g. "{urlContainsLangCodeInPath: false}"
              );
    Bson andQuery = and(
        BasicDBObject.parse("{geoLocationCountryCode: {$ne: \"NZ\"}}"),
        BasicDBObject.parse("{domain: {$not: /\\.nz/}}"),
        BasicDBObject.parse(mriFilterString),
        orQuery);

    collection.aggregate(Arrays.asList(
         match(andQuery),  //match(BasicDBObject.parse(matchQuery))
         // match((List<DBObject>)JSON.parse(matchQuery)),
         unwind("$geoLocationCountryCode"),
         group("$geoLocationCountryCode", Arrays.asList(sum("count", 1), 
                        addToSet("domain", "$basicDomain"))),
         sort(BasicDBObject.parse("{count : -1}"))
       )).forEach((Block<Document>)doc -> writeDoc(doc, writer));

    // casting to Block<Document> necessary because otherwise we see the error at
    // https://stackoverflow.com/questions/47979978/ambiguous-reference-to-foreach-when-listing-mongodbs-database-in-java

    // Less efficient way is to keep all the results in memory and then
    // write them out one at a time
    /*
    AggregateIterable<Document> output
        = collection.aggregate(Arrays.asList(
         match(andQuery),  //match(BasicDBObject.parse(matchQuery))
         // match((List<DBObject>)JSON.parse(matchQuery)),
         unwind("$geoLocationCountryCode"),
         group("$geoLocationCountryCode", Arrays.asList(sum("count", 1), addToSet("domain", "$domain"))),
         sort(BasicDBObject.parse("{count : -1}"))
     ));

    
    for (Document doc : output) {
        //System.out.println(doc);
        System.out.println(doc.toJson());
        
    }
    */
    return;
    }

    /** Do the aggregates for writing out tables.
       Table1:
       
    */
    public void writeTables(File outFolder) {
    // In this function, we're always dealing with the Websites mongodb collection.
    MongoCollection<Document> collection = this.database.getCollection(WEBSITES_COLLECTION);

    String[] tableNames = { "", "1table_allCrawledSites", "2table_sitesWithPagesInMRI"};
    for (int tableNum = 1; tableNum < tableNames.length; tableNum++) {
        File outFile = new File(outFolder, tableNames[tableNum] + ".json");
        File csvFile = new File(outFolder, tableNames[tableNum] + ".csv");
        try (
         Writer writer = new BufferedWriter(new FileWriter(outFile));
         CSVPrinter csvWriter = new CSVPrinter(new FileWriter(csvFile), CSVFormat.DEFAULT);
         ) {

        // Write out the CSV column headings
        // https://commons.apache.org/proper/commons-csv/apidocs/org/apache/commons/csv/CSVPrinter.html
        csvWriter.printRecord("countryCode", "siteCount",
              "numPagesInMRI count","numPagesContainingMRICount"/*, "domain"*/);
        
        AggregateIterable<Document> output = getTable(collection, tableNum); //doTable1().forEach((Block<Document>)doc -> writeDoc(doc, writer));
        
        int docNum = 0;
        for (Document doc : output) {       
            //System.out.println(doc);
            writeDocAsJsonRecord(++docNum, doc, writer);
            writeDocAsCSVRecord(++docNum, doc, csvWriter);
        }       
        logger.info("@@@ Wrote out table into file: " + Utility.getFilePath(outFile) + " and .csv");
        } catch(UncheckedIOException ioe) {
        logger.error("Caught UncheckedIOException: " + ioe.getMessage(), ioe);
        }
        catch(Exception e) {
        logger.error("Could not write table to file " + outFile + " or .csv equivalent" , e);
        }
    }
    }

    public AggregateIterable<Document> getTable(MongoCollection<Document> collection, int tableNum) {
    
    AggregateIterable<Document> output = null;

    switch(tableNum) {
        
    case 1:
        /* 1table_allCrawledSites -

           db.Websites.aggregate([       
           { $unwind: "$geoLocationCountryCode" },
           {
           $group: {
           _id: "$geoLocationCountryCode",
           count: { $sum: 1 },
           //domain: { $addToSet: '$domain' },
           numPagesInMRICount: { $sum: '$numPagesInMRI' },
           numPagesContainingMRICount: { $sum: '$numPagesContainingMRI' }
           }
           },
           { $sort : { count : -1} }
           ]);
         */
        output = collection.aggregate(Arrays.asList(
                           //match(BasicDBObject.parse("{urlContainsLangCodeInPath:true}")),
         unwind("$geoLocationCountryCode"),
         group("$geoLocationCountryCode", Arrays.asList(
                                sum("count", 1),
                                /*addToSet("domain", "$domain"),*/
                                sum("numPagesInMRICount", "$numPagesInMRI"),
                                sum("numPagesContainingMRICount", "$numPagesContainingMRI"))),
         sort(BasicDBObject.parse("{count : -1}"))       
        ));
        break;

    case 2:
        /*
          db.Websites.aggregate([
          { $match: { numPagesInMRI: {$gt: 0} } },
          { $unwind: "$geoLocationCountryCode" },
          {
          $group: {
          _id: {$toLower: '$geoLocationCountryCode'}, // ignore toLower
          count: { $sum: 1 },
          //domain: { $addToSet: '$domain' },
          numPagesInMRICount: { $sum: '$numPagesInMRI' },
          numPagesContainingMRICount: { $sum: '$numPagesContainingMRI' }
          }
          },
          { $sort : { count : -1} }
          ]);
         */
         output = collection.aggregate(Arrays.asList(
                           match(BasicDBObject.parse("{ numPagesInMRI: {$gt: 0} }")),
         unwind("$geoLocationCountryCode"),
         group("$geoLocationCountryCode", Arrays.asList(
                                sum("count", 1),
                                /*addToSet("domain", "$domain"),*/
                                sum("numPagesInMRICount", "$numPagesInMRI"),
                                sum("numPagesContainingMRICount", "$numPagesContainingMRI"))),
         sort(BasicDBObject.parse("{count : -1}"))       
        ));
        break;
        
    default: logger.error("Unknown table number: " + tableNum);
        
    }

     return output;
    
    }
    
    
    
    /**
     * called by lambda forEach() call on Document objects to write them out to a file.
     * Have to deal with unreported exceptions here that can't be dealt with when doing
     * the actual forEach(). See
     * https://stackoverflow.com/questions/39090292/how-to-cleanly-deal-with-unreported-exception-ioexception-in-stream-foreach
     */    
    public void writeDoc(Document doc, Writer writer) throws UncheckedIOException {

    // If there's a domain field in the json Doc, sort this domain listing alphabetically
    Object domainList = doc.remove("domain");
    if(domainList != null) {
        doc.put("domain", sortAlphabetically(domainList));
    }
    
    //OLD WAY: writer.write(doc.toJson(new JsonWriterSettings(JsonMode.STRICT, true)) + NEWLINE);
    // Can't control json output to add newlines after each array element,
    // no matter which JsonMode is used.
    
    // https://mongodb.github.io/mongo-java-driver/3.9/javadoc/index.html?org/bson/json/JsonWriterSettings.html
    // Still can't control array element output,
    // but this way uses newer mongo java driver 3.9(.1). Tried its various JsonModes too:
    //JsonWriterSettings writeSettings = new JsonWriterSettings();
    //writeSettings.builder().outputMode(JsonMode.SHELL).indent(true).build();
    //writer.write(doc.toJson(writeSettings) + NEWLINE);

    // Not the JsonWriter of mongodb java driver:
    // https://stackoverflow.com/questions/54746814/jsonwriter-add-a-new-line
    
    // Have to use gson's pretty print to produce a json string that contains
    // newlines after every array element in the json:
    
    String jsonStr = prettyPrintJson(doc.toJson());
    //System.err.println(jsonStr);
    try {
        writer.write(jsonStr + NEWLINE);
    } catch (IOException ex) {
        //throw ex;
        throw new UncheckedIOException(ex);
    }       
    }

    private List sortAlphabetically(Object list) {
    BsonArray domainList = (BsonArray)list;
    //for(String domain : domainList) {
    for(int i = domainList.size() - 1; i >= 0; i--) {
        BsonString domain = domainList.get(i).asString();
        String domainStr = Utility.stripProtocolAndWWWFromURL(domain.toString());
        domainList.set(i, new BsonString(domainStr));       
    }

    return domainList;
    }
    
    public void writeDocAsJsonRecord(int docNum, Document doc, Writer writer) throws UncheckedIOException {
    String jsonStr = prettyPrintJson(doc.toJson());
    //System.err.println(jsonStr);
    try {
        writer.write("/* " + docNum + " */\n" + jsonStr + NEWLINE);
    } catch (IOException ex) {
        //throw ex;
        throw new UncheckedIOException(ex);
    }
    }

    // TODO
    //public void writeDocToJsonAndCSV(int docNum, Document doc, Writer writer, CSVPrinter csvWriter) throws UncheckedIOException {
    public void writeDocAsCSVRecord(int docNum, Document doc, CSVPrinter csvWriter) throws UncheckedIOException {
    String jsonStr = doc.toJson();
    JsonParser parser = new JsonParser();
    JsonElement json = parser.parse(jsonStr);

    JsonObject jsonObj = (JsonObject)json;
    
    String countryCode = jsonObj.get("_id").getAsString();
    int siteCount = jsonObj.get("count").getAsInt();
    int numPagesInMRICount = jsonObj.get("numPagesInMRICount").getAsInt();
    int numPagesContainingMRICount = jsonObj.get("numPagesContainingMRICount").getAsInt();  
    
    //System.err.println(jsonStr);
    try {
        //writer.write("/* " + docNum + " */\n" + prettyPrintJson(jsonStr) + NEWLINE);
        csvWriter.printRecord(countryCode, siteCount, numPagesInMRICount, numPagesContainingMRICount);
    } catch (IOException ex) {
        //throw ex;
        throw new UncheckedIOException(ex);
    }
    }
    
    public String prettyPrintJson(String jsonStr) {
    Gson gson = new GsonBuilder().setPrettyPrinting().create();
    JsonParser jp = new JsonParser();
    JsonElement je = jp.parse(jsonStr);
    String prettyJsonString = gson.toJson(je);
    return prettyJsonString;
    }

    
    /** https://stackoverflow.com/questions/19938153/do-i-need-to-explicitly-close-connection */
    public void close() {}
    

    // TODO:
    // In the database, need to ensure we have else
    // create collection (table in RDBMS) websites, create collection webpages.
    // The webpages collection will have sentences embedded based on my decisions from
    // reading the series
    // https://www.mongodb.com/blog/post/6-rules-of-thumb-for-mongodb-schema-design-part-1    
    // Then need functions:
    // insertWebsiteDocument()
    // insertWebpageDocument()
    
    public static void main(String args[]) {
    try {
        MongoDBAccess mongodbCon = new MongoDBAccess();
        mongodbCon.connectToDB();
        mongodbCon.showCollections();

    } catch(Exception e) {
        e.printStackTrace();
    }
    }
}