Merge branch 'master' of github.com:juditacs/wordcount

Author: Judit Acs

javascript/package.json

@@ -1,6 +1,7 @@
 {
  "name": "wordcount",
  "version": "1.0.0",
+ "readme": "../README.md",
  "description": "Javascript implementation of wordcount for https://github.com/juditacs/wordcount",
  "main": "wordcount.js",
  "repository": "juditacs/wordcount",
@@ -10,6 +11,6 @@
  "author": "Laci",
  "license": "MIT",
  "dependencies": {
- "lodash": "^3.10.1"
+ "timsort": "^0.2.1"
  }
 }

javascript/wordcount.js

@@ -1,31 +1,58 @@
-var _ = require('lodash');
-var Console = require('console').Console;
+"use strict";
+var readline = require('readline');
+var TimSort = require('timsort');
 
-var data = {};
+process.stdin.setEncoding('utf8');
 
-var log = new Console(process.stderr, process.stderr);
+let rl = readline.createInterface({ input: process.stdin, terminal: false });
+let wordCounts = {};
+const regExp = /[ \t\n\r]+/g;
 
-process.stdin.setEncoding('utf8');
-process.stdin.on('readable', function() {
- var chunk = process.stdin.read();
- if (chunk !== null) {
- _.forEach(chunk.split(/[\n\t\r\f ]+/g), function(word) {
- if (word == '') {return;}
- if (data[word]) {
- //console.log('Adding to ' + word + '.');
- data[word].count++;
- } else {
- //console.log('Inserting ' + word + '.');
- data[word] = {word: word, count: 1};
- }
- });
+rl.on('line', (line) => {
+ let words = line.trim().split(regExp);
+
+ for(let i = 0, len = words.length; i < len; i++) {
+ let word = words[i];
+
+ if (!word)
+ return;
+
+ if(wordCounts.hasOwnProperty(word)) {
+ wordCounts[word]++;
+ }
+ else {
+ wordCounts[word] = 1;
+ }
  }
-});
+}).on('close', () => {
+ let wordList = Object.keys(wordCounts);
+
+ let compare = (x, y) => {
+ if(wordCounts[x] < wordCounts[y])
+ return 1;
 
-process.stdin.on('end', function() {
- var sorted = _.sortByOrder(data, ['count', 'word'], ['desc', 'asc']);
- //console.log('Output has '+sorted.length+' items:');
- _.forEach(sorted, function(item){
- process.stdout.write(item.word + '\t' + item.count + '\n');
- });
+ if(wordCounts[x] === wordCounts[y] && x > y)
+ return 1;
+
+ return -1;
+ }
+
+ TimSort.sort(wordList, compare);
+
+ let word = '';
+ let count = '';
+ let buff = '';
+ const maxBuff = 1000;
+ for(let i = 0, len = wordList.length; i < len; i++) {
+ word = wordList[i];
+ count = wordCounts[word];
+
+ buff += word + '\t' + count + '\n';
+ if (i % maxBuff === 0) {
+ process.stdout.write(buff);
+ buff = '';
+ }
+ }
+ process.stdout.write(buff);
+ process.exit(0);
 });

javascript/wordcount2.js

@@ -9,13 +9,13 @@ const regExp = /[ \t\n\r]+/g;
 
 rl.on('line', (line) => {
  var words = line.trim().split(regExp);
- 
+
  for(var i = 0, len = words.length; i < len; i++) {
  var word = words[i];
- 
+
  if (!word)
  return;
- 
+
  if(wordCounts.hasOwnProperty(word)) {
  wordCounts[word]++;
  }
@@ -25,23 +25,23 @@ rl.on('line', (line) => {
  }
 }).on('close', () => {
  var wordList = Object.keys(wordCounts);
- 
+
  wordList.sort((x, y) => {
  if(wordCounts[x] < wordCounts[y])
  return 1;
- 
+
  if(wordCounts[x] === wordCounts[y] && x > y)
  return 1;
- 
+
  return -1;
  });
- 
+
  for(var i = 0, len = wordList.length; i < len; i++) {
  var word = wordList[i];
  var count = wordCounts[word];
- 
+
  process.stdout.write(word + '\t' + count + '\n');
  }
- 
+
  process.exit(0);
 });

julia/wordcount.jl

@@ -2,8 +2,8 @@ customlt(a,b) = (b.second < a.second) ? true : b.second == a.second ? a.first <
 
 function main()
  wc = Dict{UTF8String,Int64}()
- for l in eachline(STDIN)
- for w in split(l)
+ for l::UTF8String in eachline(STDIN)
+ for w::UTF8String in split(l)
  wc[w]=get(wc, w, 0) + 1
  end
  end

perl/wordcount.pl

@@ -1,20 +1,7 @@
 #!/usr/bin/env perl
 
+my %words;
+while(<>) { $words{$_}++ for split }
 
-sub word_count {
- my %counter;
- my $line = <STDIN>;
- while($line ne "") {
- my @words = split(' ', $line);
- for my $word (@words) {
- $counter{$word} ++;
- }
- $line = <STDIN>;
- }
-
- foreach my $word (sort { -$counter{$a} <=> -$counter{$b} or $a cmp $b } keys %counter) {
- printf "%s\t%s\n", $word, $counter{$word};
- }
-}
-
-word_count();
+print "$_\t$words{$_}\n"
+ for sort {$words{$b}<=>$words{$a} or $a cmp $b} keys %words;

python/wordcount_py2gabor.py

@@ -0,0 +1,26 @@
+#!/usr/bin/env python2
+from sys import stdin
+
+def counter():
+ keyWords = {}
+ for l in stdin:
+ for word in l.split():
+ if word in keyWords:
+ keyWords[word] += 1
+ else:
+ keyWords[word] = 1
+
+ keyCounts = {}
+ for word in keyWords.keys():
+ count = keyWords[word]
+ if count in keyCounts:
+ keyCounts[count].append(word)
+ else:
+ keyCounts[count] = [word]
+
+ for count in sorted(keyCounts, reverse = True):
+ for word in sorted(keyCounts[count]):
+ print word+'\t'+str(count)
+
+if __name__ == '__main__':
+ counter()

run_commands.txt

@@ -10,6 +10,7 @@ nodejs javascript/wordcount.js
 nodejs javascript/wordcount2.js
 perl/wordcount.pl
 python/wordcount_py2.py
+python/wordcount_py2gabor.py
 python/wordcount_py3.py
 php php/wordcount.php
 go/bin/wordcount