codejava/prog-intro
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Upload files to "java/wordStat"

Browse Source
This commit is contained in:
codejava
2026-04-13 10:42:12 +03:00
parent 8f41006963
commit 0d6b168b1e
5 changed files with 492 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

117
java/wordStat/WordStatLengthAffix.java Normal file
View File

@@ -0,0 +1,117 @@
import java.io.*;
import java.util.*;
/**
* @author Nikita Doschennikov (me@fymio.us)
*/
public class WordStatLengthAffix {
public static void main(String[] args) {
if (args.length != 2) {
System.err.println("incorrect input!");
System.err.println(
"usage: java WordStatLengthAffix <inputFile> <outputFile>"
);
}
String inputFileName = args[0];
String outputFileName = args[1];
try {
BufferedReader r = new BufferedReader(
new FileReader(inputFileName)
);
Map<String, WordInfo> wordMap = new HashMap<>();
StringBuilder sb = new StringBuilder();
int wordIndex = 0;
int data = r.read();
while (data != -1) {
char c = (char) data;
if (
Character.getType(c) == Character.DASH_PUNCTUATION ||
Character.isLetter(c) ||
c == '\''
) {
sb.append(c);
} else {
if (sb.length() > 0) {
String word = sb.toString().toLowerCase();
if (word.length() != 1) {
String prefix = word.substring(
0,
word.length() / 2
);
String suffix = word.substring(
word.length() - word.length() / 2
);
if (wordMap.containsKey(prefix)) {
wordMap.get(prefix).count++;
} else {
wordMap.put(
prefix,
new WordInfo(prefix, 1, wordIndex)
);
wordIndex++;
}
if (wordMap.containsKey(suffix)) {
wordMap.get(suffix).count++;
} else {
wordMap.put(
suffix,
new WordInfo(suffix, 1, wordIndex)
);
wordIndex++;
}
}
sb.setLength(0);
}
}
data = r.read();
}
if (sb.length() > 0) {
String word = sb.toString().toLowerCase();
if (word.length() != 1) {
String prefix = word.substring(0, word.length() / 2);
String suffix = word.substring(
word.length() - word.length() / 2
);
if (wordMap.containsKey(prefix)) {
wordMap.get(prefix).count++;
} else {
wordMap.put(prefix, new WordInfo(prefix, 1, wordIndex));
wordIndex++;
}
if (wordMap.containsKey(suffix)) {
wordMap.get(suffix).count++;
} else {
wordMap.put(suffix, new WordInfo(suffix, 1, wordIndex));
wordIndex++;
}
}
}
r.close();
List<WordInfo> sortedWords = new ArrayList<>(wordMap.values());
sortedWords.sort(
Comparator.comparingInt((WordInfo w) ->
w.word.length()
).thenComparingInt(w -> w.firstIndex)
);
PrintWriter writer = new PrintWriter(outputFileName, "UTF-8");
for (WordInfo info : sortedWords) {
writer.println(info.word + " " + info.count);
}
writer.close();
} catch (Exception ex) {
System.err.println("An error occured: " + ex.getMessage());
}
}
}

93
java/wordStat/WordStatLengthMiddle.java Normal file
View File

@@ -0,0 +1,93 @@
import java.io.*;
import java.util.*;
/**
* @author Nikita Doschennikov (me@fymio.us)
*/
public class WordStatLengthMiddle {
public static void main(String[] args) {
if (args.length != 2) {
System.err.println("incorrect input!");
System.err.println(
"usage: java WordStatLengthMiddle <inputFile> <outputFile>"
);
}
String inputFileName = args[0];
String outputFileName = args[1];
try {
BufferedReader r = new BufferedReader(
new FileReader(inputFileName)
);
Map<String, WordInfo> wordMap = new HashMap<>();
StringBuilder sb = new StringBuilder();
int wordIndex = 0;
int data = r.read();
while (data != -1) {
char c = (char) data;
if (
Character.getType(c) == Character.DASH_PUNCTUATION ||
Character.isLetter(c) ||
c == '\''
) {
sb.append(c);
} else {
if (sb.length() > 0) {
String word = sb.toString().toLowerCase();
if (word.length() >= 7) {
word = word.substring(3, word.length() - 3);
if (wordMap.containsKey(word)) {
wordMap.get(word).count++;
} else {
wordMap.put(
word,
new WordInfo(word, 1, wordIndex)
);
wordIndex++;
}
}
sb.setLength(0);
}
}
data = r.read();
}
if (sb.length() > 0) {
String word = sb.toString().toLowerCase();
if (word.length() >= 7) {
word = word.substring(3, word.length() - 3);
if (wordMap.containsKey(word)) {
wordMap.get(word).count++;
} else {
wordMap.put(word, new WordInfo(word, 1, wordIndex));
wordIndex++;
}
}
}
r.close();
List<WordInfo> sortedWords = new ArrayList<>(wordMap.values());
sortedWords.sort(
Comparator.comparingInt((WordInfo w) ->
w.word.length()
).thenComparingInt(w -> w.firstIndex)
);
PrintWriter writer = new PrintWriter(outputFileName, "UTF-8");
for (WordInfo info : sortedWords) {
writer.println(info.word + " " + info.count);
}
writer.close();
} catch (Exception ex) {
System.err.println("An error occured: " + ex.getMessage());
}
}
}

96
java/wordStat/WordStatLengthPrefix.java Normal file
View File

@@ -0,0 +1,96 @@
import java.io.*;
import java.util.*;
/**
* @author Nikita Doschennikov (me@fymio.us)
*/
public class WordStatLengthPrefix {
public static void main(String[] args) {
if (args.length != 2) {
System.err.println("incorrect input!");
System.err.println(
"usage: java WordStatLengthPrefix <inputFile> <outputFile>"
);
}
String inputFileName = args[0];
String outputFileName = args[1];
try {
BufferedReader r = new BufferedReader(
new FileReader(inputFileName)
);
Map<String, WordInfo> wordMap = new HashMap<>();
StringBuilder sb = new StringBuilder();
int wordIndex = 0;
int data = r.read();
while (data != -1) {
char c = (char) data;
if (
Character.getType(c) == Character.DASH_PUNCTUATION ||
Character.isLetter(c) ||
c == '\''
) {
sb.append(c);
} else {
if (sb.length() > 0) {
String word = sb.toString().toLowerCase();
if (word.length() != 1) {
String prefix = word.substring(
0,
word.length() / 2
);
if (wordMap.containsKey(prefix)) {
wordMap.get(prefix).count++;
} else {
wordMap.put(
prefix,
new WordInfo(prefix, 1, wordIndex)
);
wordIndex++;
}
}
sb.setLength(0);
}
}
data = r.read();
}
if (sb.length() > 0) {
String word = sb.toString().toLowerCase();
if (word.length() != 1) {
String prefix = word.substring(0, word.length() / 2);
if (wordMap.containsKey(prefix)) {
wordMap.get(prefix).count++;
} else {
wordMap.put(prefix, new WordInfo(prefix, 1, wordIndex));
wordIndex++;
}
}
}
r.close();
List<WordInfo> sortedWords = new ArrayList<>(wordMap.values());
sortedWords.sort(
Comparator.comparingInt((WordInfo w) ->
w.word.length()
).thenComparingInt(w -> w.firstIndex)
);
PrintWriter writer = new PrintWriter(outputFileName, "UTF-8");
for (WordInfo info : sortedWords) {
writer.println(info.word + " " + info.count);
}
writer.close();
} catch (Exception ex) {
System.err.println("An error occured: " + ex.getMessage());
}
}
}

95
java/wordStat/WordStatLengthSuffix.java Normal file
View File

@@ -0,0 +1,95 @@
import java.io.*;
import java.util.*;
/**
* @author Nikita Doschennikov (me@fymio.us)
*/
public class WordStatLengthSuffix {
public static void main(String[] args) {
if (args.length != 2) {
System.err.println("incorrect input!");
System.err.println(
"usage: java WordStatLengthSuffix <inputFile> <outputFile>"
);
}
String inputFileName = args[0];
String outputFileName = args[1];
try {
BufferedReader r = new BufferedReader(
new FileReader(inputFileName)
);
Map<String, WordInfo> wordMap = new HashMap<>();
StringBuilder sb = new StringBuilder();
int wordIndex = 0;
int data = r.read();
while (data != -1) {
char c = (char) data;
if (
Character.getType(c) == Character.DASH_PUNCTUATION ||
Character.isLetter(c) ||
c == '\''
) {
sb.append(c);
} else {
if (sb.length() > 0) {
String word = sb.toString().toLowerCase();
if (word.length() != 1) {
word = word.substring(
word.length() - word.length() / 2
);
if (wordMap.containsKey(word)) {
wordMap.get(word).count++;
} else {
wordMap.put(
word,
new WordInfo(word, 1, wordIndex)
);
wordIndex++;
}
}
sb.setLength(0);
}
}
data = r.read();
}
if (sb.length() > 0) {
String word = sb.toString().toLowerCase();
if (word.length() != 1) {
word = word.substring(word.length() - word.length() / 2);
if (wordMap.containsKey(word)) {
wordMap.get(word).count++;
} else {
wordMap.put(word, new WordInfo(word, 1, wordIndex));
wordIndex++;
}
}
}
r.close();
List<WordInfo> sortedWords = new ArrayList<>(wordMap.values());
sortedWords.sort(
Comparator.comparingInt((WordInfo w) ->
w.word.length()
).thenComparingInt(w -> w.firstIndex)
);
PrintWriter writer = new PrintWriter(outputFileName, "UTF-8");
for (WordInfo info : sortedWords) {
writer.println(info.word + " " + info.count);
}
writer.close();
} catch (Exception ex) {
System.err.println("An error occured: " + ex.getMessage());
}
}
}

91
java/wordStat/WordStatTest.java Normal file
View File

@@ -0,0 +1,91 @@
package wordStat;
import base.Named;
import base.Selector;
import java.util.Comparator;
import java.util.function.Function;
import java.util.stream.Stream;
/**
* Tests for <a href="https://www.kgeorgiy.info/courses/prog-intro/homeworks.html#wordstat">Word Statistics</a> homework
* of <a href="https://www.kgeorgiy.info/courses/prog-intro/">Introduction to Programming</a> course.
*
* @author Georgiy Korneev (kgeorgiy@kgeorgiy.info)
*/
public final class WordStatTest {
// === Base
private static final Named<Function<String, Stream<String>>> ID = Named.of(
"",
Stream::of
);
private static final WordStatTester.Variant BASE =
new WordStatTester.Variant("", false, Comparator.comparingInt(p -> 0));
// === 3637
public static final int SIZE = 3;
private static final WordStatTester.Variant LENGTH =
new WordStatTester.Variant(
"Length",
false,
Comparator.comparingInt(p -> p.first().length())
);
private static final Named<Function<String, Stream<String>>> MIDDLE = size(
"Middle",
SIZE * 2 + 1,
s -> Stream.of(s.substring(SIZE, s.length() - SIZE))
);
static Named<Function<String, Stream<String>>> size(
final String name,
final int length,
final Function<String, Stream<String>> f
) {
return Named.of(name, s ->
s.length() >= length ? f.apply(s) : Stream.empty()
);
}
// === 3839
private static final Named<Function<String, Stream<String>>> AFFIX = size(
"Affix",
2,
s ->
Stream.of(
s.substring(0, s.length() / 2),
s.substring(s.length() - s.length() / 2)
)
);
// === 3536
private static final Named<Function<String, Stream<String>>> SUFFIX = size(
"Suffix",
2,
s -> Stream.of(s.substring(s.length() - s.length() / 2))
);
// === 4749
private static final Named<Function<String, Stream<String>>> PREFIX = size(
"Prefix",
2,
s -> Stream.of(s.substring(0, s.length() / 2))
);
// === Common
public static final Selector SELECTOR = new Selector(WordStatTester.class)
.variant("Base", BASE.with(ID))
.variant("3637", LENGTH.with(MIDDLE))
.variant("3839", LENGTH.with(AFFIX))
.variant("3435", LENGTH.with(SUFFIX))
.variant("3233", LENGTH.with(ID))
.variant("4142", LENGTH.with(MIDDLE))
.variant("4749", LENGTH.with(PREFIX));
private WordStatTest() {
// Utility class
}
public static void main(final String... args) {
SELECTOR.main(args);
}
}