Object-Oriented Java Primitives

Project architect: @victornoel

Cactoos is a collection of object-oriented Java primitives.

We are not happy with JDK, Guava, and Apache Commons because they are procedural and not object-oriented. They do their job, but mostly through static methods. Cactoos is suggesting to do almost exactly the same, but through objects.

These are the design principles behind Cactoos.

The library has no dependencies. All you need is this:

Maven:

<dependency>
<groupId>org.cactoos</groupId>
<artifactId>cactoos</artifactId>
<version>0.61.0</version>
</dependency>

Gradle:

dependencies {
compile 'org.cactoos:cactoos::0.61.0}

Java version required: 17+.

StackOverflow tag is cactoos.

Input/Output

More about it here: Object-Oriented Declarative Input/Output in Cactoos.

To read a text file in UTF-8:

Stringtext = newTextOf(
newFile("/code/a.txt")
).asString();

To write a text into a file:

newLengthOf(
newTeeInput(
"Hello, world!",
newFile("/code/a.txt")
)
).value();

To read a binary file from classpath:

byte[] data = newBytesOf(
newResourceOf("foo/img.jpg")
).asBytes();

Text/Strings

To format a text:

Stringtext = newFormattedText(
"How are you, %s?",
name
).asString();

To manipulate text:

// To lower casenewLowered(
newTextOf("Hello")
);
// To upper casenewUpper(
newTextOf("Hello")
);

Iterables/Collections/Lists/Sets

More about it here: Lazy Loading and Caching via Sticky Cactoos Primitives.

To filter a collection:

Collection<String> filtered = newListOf<>(
newFiltered<>(
s -> s.length() > 4,
newIterableOf<>("hello", "world", "dude")
)
);

To flatten one iterable:

newJoined<>(
newMapped<IterableOf<>>(
iter -> newIterableOf<>(
newListOf<>(iter).toArray(newInteger[]{})
),
newIterableOf<>(1, 2, 3, 4, 5, 6)
)
); // Iterable<Integer>

To flatten and join several iterables:

newJoined<>(
newMapped<IterableOf<>>(
iter -> newIterableOf<>(
newJoined<>(iter)
),
newJoined<>(
newIterableOf<>(newIterableOf<>(1, 2, 3)),
newIterableOf<>(newIterableOf<>(4, 5, 6))
)
)
); // Iterable<Integer>

To iterate a collection:

newAnd(
newMapped<>(
newFuncOf<>(
input -> {
System.out.printf("Item: %s\n", input);
},
newTrue()
),
newIterableOf<>("how", "are", "you", "?")
)
).value();

Or even more compact:

newForEach<>(
input -> System.out.printf(
"Item: %s\n", input
)
).exec(newIterableOf<>("how", "are", "you", "?"));

To sort a list of words in the file:

List<Text> sorted = newListOf<>(
newSorted<>(
newMapped<>(
text -> newComparableText(text),
newSplit(
newTextOf(
newFile("/tmp/names.txt")
),
newTextOf("\\s+")
)
)
)
);

To count elements in an iterable:

inttotal = newLengthOf(
newIterableOf<>("how", "are", "you")
).value().intValue();

To create a set of elements by providing variable arguments:

finalSet<String> unique = newSetOf<>(
"one",
"two",
"one",
"three"
);

To create a set of elements from an existing iterable:

finalSet<String> words = newSetOf<>(
newIterableOf<>("abc", "bcd", "abc", "ccc")
);

To create a sorted iterable with unique elements from an existing iterable:

finalIterable<String> sorted = newSorted<>(
newSetOf<>(
newIterableOf<>("abc", "bcd", "abc", "ccc")
)
);

To create a sorted set from existing vararg elements using a comparator:

finalSet<String> sorted = neworg.cactoos.set.Sorted<>(
(first, second) -> first.compareTo(second),
"abc", "bcd", "abc", "ccc", "acd"
);

To create a sorted set from an existing iterable using a comparator:

finalSet<String> sorted = neworg.cactoos.set.Sorted<>(
(first, second) -> first.compareTo(second),
newIterableOf<>("abc", "bcd", "abc", "ccc", "acd")
);

Funcs and Procs

This is a traditional foreach loop:

for (Stringname : names) {
System.out.printf("Hello, %s!\n", name);
}

This is its object-oriented alternative (no streams!):

newAnd(
n -> {
System.out.printf("Hello, %s!\n", n);
returnnewTrue().value();
},
names
).value();

This is an endless while/do loop:

while (!ready) {
System.out.println("Still waiting...");
}

Here is its object-oriented alternative:

newAnd(
ready -> {
System.out.println("Still waiting...");
return !ready;
},
newEndless<>(booleanParameter)
).value();

Dates and Times

From our org.cactoos.time package.

Our classes are divided in two groups: those that parse strings into date/time objects, and those that format those objects into strings.

For example, this is the traditional way of parsing a string into an OffsetDateTime:

finalOffsetDateTimedate = OffsetDateTime.parse("2007-12-03T10:15:30+01:00");

Here is its object-oriented alternative (no static method calls!) using OffsetDateTimeOf, which is a Scalar:

finalOffsetDateTimedate = newOffsetDateTimeOf("2007-12-03T10:15:30+01:00").value();

To format an OffsetDateTime into a Text:

finalOffsetDateTimedate = ...;
finalStringtext = newTextOfDateTime(date).asString();

Our Objects vs. Their Static Methods

Cactoos	Guava	Apache Commons	JDK 17
And	Iterables.all()	-	-
Filtered	Iterables.filter()	?	-
FormattedText	-	-	String.format()
IsBlank	-	StringUtils.isBlank()	-
Joined	-	-	String.join()
LengthOf	-	-	String#length()
Lowered	-	-	String#toLowerCase()
Normalized	-	StringUtils.normalize()	-
Or	Iterables.any()	-	-
Repeated	-	StringUtils.repeat()	-
Replaced	-	-	String#replace()
Reversed	-	-	StringBuilder#reverse()
Rotated	-	StringUtils.rotate()	-
Split	-	-	String#split()
StickyList	Lists.newArrayList()	?	Arrays.asList()
Sub	-	-	String#substring()
SwappedCase	-	StringUtils.swapCase()	-
TextOf	?	IOUtils.toString()	-
TrimmedLeft	-	StringUtils.stripStart()	-
TrimmedRight	-	StringUtils.stripEnd()	-
Trimmed	-	StringUtils.stripAll()	String#trim()
Upper	-	-	String#toUpperCase()

Questions

Ask your questions related to cactoos library on Stackoverflow with the cactoos tag.

Architecture

Interfaces as the sole public abstraction. Every primitive is declared as a Java @FunctionalInterface—Scalar<T>, Text, Func<X,Y>, Input, Output, Bytes, and Proc—with no static utility classes at this level. This contrasts with Guava and Apache Commons, where functionality is delivered through static methods on utility classes (Iterables.filter(), StringUtils.isBlank()). Because every concept is a real object, it can be subclassed, mocked, and composed without reflection or special tooling.

Decorator pattern for all behavior extension. Thread safety, memoization, null validation, and exception translation are each added by wrapping one object in another, not by modifying the class. Sticky<T> caches a Scalar<T> result; Synced<T> makes it thread-safe; NoNulls rejects null inputs; Unchecked<T> rethrows checked exceptions as unchecked ones. This follows the Decorator pattern from GoF: every class has exactly one responsibility, and capabilities are composed at call sites rather than inherited.

Envelope pattern for safe inheritance. Abstract base classes—IterableEnvelope, TextEnvelope, ScalarEnvelope, and others—accept an inner object through the constructor and delegate to it via final methods. Concrete classes extend an envelope and supply only the inner object; no method bodies are overridden. This eliminates the fragile base class problem: because delegation methods are final, a subclass cannot accidentally alter delegated behavior.

Lazy evaluation as the default. No computation occurs until the caller explicitly requests it—.value(), .asString(), or .iterator(). An IterableOf<> wrapping a transformation does no work until iterated. This contrasts with Guava's ImmutableList.copyOf() and Apache Commons Collections, which materialize elements at construction time. Explicit caching requires opting in with Sticky or StickyList.

Checked exceptions as first-class interface contracts. All core interfaces declare throws Exception; the library never silently converts checked exceptions to unchecked ones at the interface boundary. When a caller needs unchecked access, it wraps with Unchecked<T> (rethrows as UncheckedIOException), IoChecked<T> (narrows to IOException), or Checked<T,E> (converts to any target exception type). The decision about exception handling stays with the call site, not the library.

Zero runtime dependencies. The library ships with no runtime dependencies. Guava adds a ~3 MB JAR; Apache Commons splits across multiple JARs (commons-lang3, commons-io, etc.). Cactoos requires only JDK 17+, making it suitable for environments where classpath size and dependency audits matter.

How to Contribute

Just fork the repo and send us a pull request.

Make sure your branch builds without any warnings/issues:

mvn clean verify -Pqulice

To run a build similar to the CI with Docker only, use:

docker run \
--tty \
--interactive \
--workdir=/main \
--volume=${PWD}:/main \
--volume=cactoos-mvn-cache:/root/.m2 \
--rm \
maven:3-eclipse-temurin-17 \
bash -c "mvn clean install -Pqulice; chown -R $(id -u):$(id -g) target/"

To remove the cache used by Docker-based build:

docker volume rm cactoos-mvn-cache

Note: Checkstyle is used as a static code analysis tool with checks list in GitHub precommits.

Contributors

• @yegor256 as Yegor Bugayenko (Blog)
• @g4s8 as Kirill Che.
• @fabriciofx as Fabrício Cabral
• @englishman as Andriy Kryvtsun
• @VsSekorin as Vseslav Sekorin
• @DronMDF as Andrey Valyaev
• @dusan-rychnovsky as Dušan Rychnovský (Blog)
• @timmeey as Tim Hinkes (Blog)
• @alex-semenyuk as Alexey Semenyuk
• @smallcreep as Ilia Rogozhin
• @memoyil as Mehmet Yildirim
• @llorllale as George Aristy
• @driver733 as Mikhail Yakushin
• @izrik as Richard Sartor
• @Vatavuk as Vedran Grgo Vatavuk
• @dgroup as Yurii Dubinka
• @iakunin as Maksim Iakunin
• @fanifieiev as Fevzi Anifieiev
• @victornoel as Victor Noël
• @paulodamaso as Paulo Lobo