Added option to limit number of hits

bin/getpapers.js

@@ -34,6 +34,8 @@ program
         'download anything)')
 .option('-f, --logfile <filename>',
         'save log to specified file in output directory as well as printing to terminal')
+.option('-k, --limit <int>',
+	'limit the number of hits and downloads')
 .parse(process.argv);
 
 if (!process.argv.slice(2).length) {
@@ -96,6 +98,7 @@ options.xml = program.xml;
 options.pdf = program.pdf;
 options.supp = program.supp;
 options.all = program.all;
+options.hitlimit = parseInt(program.limit);
 options.noexecute = program.noexecute;
 if (options.noexecute) {
   log.info('Running in no-execute mode, so nothing will be downloaded');

lib/eupmc.js

@@ -23,12 +23,14 @@ EuPmc.prototype.search = function(query) {
   if (!eupmc.opts.all) {
     query += " OPEN_ACCESS:y";
   }
-  var options = { resulttype: 'core' };
+  eupmc.pagesize = '100'
+  var options = { resulttype: 'core', pageSize: eupmc.pagesize };
   eupmc.queryurl = eupmc.buildQuery(query, options);
   eupmc.first = true;
   eupmc.hitcount = 0;
+  eupmc.residualhits = 0;
   eupmc.allresults = [];
-  eupmc.iter = 0;
+  eupmc.iter = 1; //we always get back the first page
 
   eupmc.pageQuery();
 
@@ -68,23 +70,34 @@ EuPmc.prototype.completeCallback = function(data) {
       process.exit(0);
     }
 
+    // set hitlimit
+    if (eupmc.opts.hitlimit) {
+      log.info('Limiting to ' + eupmc.opts.hitlimit + ' hits');
+    }
+    else { eupmc.opts.hitlimit = eupmc.hitcount; }
+
     // create progress bar
     var progmsg = 'Retrieving results [:bar] :percent' +
                   ' (eta :etas)';
     var progopts = {
-      total: eupmc.hitcount,
+      total: eupmc.opts.hitlimit,
       width: 30,
       complete: chalk.green('=')
     };
     eupmc.pageprogress = new ProgressBar(progmsg, progopts);
 
   }
-
-  var result = resp.resultList[0].result;
+  if (eupmc.residualhits) {
+    var result = resp.resultList[0].result.slice(0,eupmc.residualhits);
+  }
+  else { var result = resp.resultList[0].result; }
   eupmc.allresults = eupmc.allresults.concat(result);
   eupmc.pageprogress.tick(result.length);
-
-  if (eupmc.allresults.length < eupmc.hitcount) {
+
+  if (eupmc.allresults.length < eupmc.opts.hitlimit) { //we still have more results to get
+    if (eupmc.opts.hitlimit - eupmc.allresults.length < eupmc.pagesize) {
+     eupmc.residualhits = eupmc.opts.hitlimit - eupmc.allresults.length;
+    }
     eupmc.iter += 1;
     eupmc.pageQuery();
   } else {