angryscan-core
1.5.1indexedFacilitates integration with Maven and Gradle for managing dependencies, supporting JavaScript environments.
3
Stars
1
Used by
dependents
—
Health
/ 100
Facilitates integration with Maven and Gradle for managing dependencies, supporting JavaScript environments.
Angry Data Core is a Kotlin Multiplatform library for locating sensitive data (PII) inside free-form text. The project focuses on Russian banking and personal identifiers and ships with tuned regular expressions, contextual validation logic, and two execution engines: a high-performance HyperScan backend for the JVM and a portable Kotlin regex engine that runs everywhere Kotlin does. Native bindings are also provided so the scanner can be embedded from non-JVM applications.
HyperScanEngine (Hyperscan powered, AutoCloseable) for throughput-sensitive JVM workloads and KotlinEngine for portable scanning.Match results include the detected value, ten characters of surrounding context, and absolute character offsets to simplify redaction or highlighting.org.angryscan.common.extensions.Matchers) plus Kotlin Serialization support, making it easy to persist engine configurations or ship them over the wire.AngryData.dll/libAngryData.so/libAngryData.dylib) with C ABI exports for integrations such as Python.:kotlin-lib - core multiplatform library with engines, matchers, shared constants, and publishing settings.:native-lib - Kotlin/Native wrapper that exposes C-callable detection functions backed by the core library.python-example - ctypes demo that consumes the native shared library and prints detection results.All artifacts are published under the org.angryscan group.
repositories {
mavenCentral()
}
dependencies {
// Multiplatform artifact; usable from JVM or MPP projects
implementation("org.angryscan:core:1.3.5")
}
<dependency>
<groupId>org.angryscan</groupId>
<artifactId>core</artifactId>
<version>1.3.5</version>
</dependency>
dependencies {
implementation("org.angryscan:core-js:1.3.5")
}
Note: The HyperScan backend relies on the
com.gliwka.hyperscanwrapper. Ensure your runtime platform is supported by that dependency; otherwise fall back to the portableKotlinEngine.
import org.angryscan.common.engine.hyperscan.HyperScanEngine
import org.angryscan.common.engine.hyperscan.IHyperMatcher
import org.angryscan.common.extensions.Matchers
val text = """
Client Ivan Ivanov lives at Moscow, Tverskaya street 1.
Card 4276 8070 1492 7948 (CVV 123), phone +7 (916) 123-45-67 and email ivanov@example.org.
""".trimIndent()
HyperScanEngine(Matchers.filterIsInstance<IHyperMatcher>()).use { engine ->
val matches = engine.scan(text)
matches.forEach { match ->
println("${match.matcher.name}: ${match.value} at ${match.startPosition}-${match.endPosition}")
}
}
Compiling regular expressions on every startup can be slow when the matcher set is large. You can compile once, save the database, and reload it instantly on subsequent runs.
import org.angryscan.common.engine.hyperscan.HyperScanEngine
import org.angryscan.common.engine.hyperscan.IHyperMatcher
import org.angryscan.common.extensions.Matchers
import java.io.File
val matchers = Matchers.filterIsInstance<IHyperMatcher>()
val dbFile = File("hyperscan.db")
// First run — compile and save
HyperScanEngine(matchers).use { engine ->
engine.saveCompiledDatabase(dbFile)
}
// Subsequent runs — load (no compilation)
HyperScanEngine.fromCompiledDatabase(matchers, dbFile).use { engine ->
engine.scan(text).forEach { match ->
println("${match.matcher.name}: ")
}
}
In-memory ByteArray variants are also available:
// Save to bytes
val bytes: ByteArray = engine.saveCompiledDatabase()
// Load from bytes
val fast = HyperScanEngine.fromCompiledDatabase(matchers, bytes)
Compatibility note: the saved database is tied to the exact matcher set, their order, and the
requireKeywordsflag used during compilation. Loading a database with a different configuration will throwIllegalArgumentException. The binary format is also platform-specific (see Hyperscan documentation).
import org.angryscan.common.engine.kotlin.IKotlinMatcher
import org.angryscan.common.engine.kotlin.KotlinEngine
import org.angryscan.common.extensions.Matchers
val engine = KotlinEngine(Matchers.filterIsInstance<IKotlinMatcher>())
val matches = engine.scan(text)
You can mask card numbers using the matcher's IMask API to keep only the first 6 and last 4 digits visible.
import org.angryscan.common.matchers.CardNumber
val masked = CardNumber().mask("1234 5678 9012 3456")
// masked == "1234 56** **** 3456"
To redact card numbers inside a larger text using scan results:
import org.angryscan.common.engine.kotlin.IKotlinMatcher
import org.angryscan.common.engine.kotlin.KotlinEngine
import org.angryscan.common.extensions.Matchers
import org.angryscan.common.engine.IMask
import org.angryscan.common.matchers.CardNumber
val text = "Card 4276 8070 1492 7948 and 1234567890123456"
val engine = KotlinEngine(Matchers.filterIsInstance<IKotlinMatcher>())
matches = engine.scan(text)
sb = StringBuilder(text)
matches
.filter { it.matcher CardNumber }
.sortedByDescending { it.startPosition }
.forEach { m ->
masked = (m.matcher IMask).mask(m.value)
sb.replace(m.startPosition, m.endPosition, masked)
}
redacted = sb.toString()
import org.angryscan.common.matchers.CardNumber
import org.angryscan.common.matchers.UserSignature
import org.angryscan.common.engine.kotlin.KotlinEngine
val customEngine = KotlinEngine(
listOf(
CardNumber(checkCardBins = false), // skip BIN validation for testing data
UserSignature(
name = "Corporate stamp",
searchSignatures = mutableListOf("OOO Romashka", "ZAO Test")
)
)
)
Each Match contains:
Use this metadata to redact or highlight findings in downstream systems.
# Run all tests (JVM + JS + Native)
./gradlew check
# JVM-specific tests
./gradlew :kotlin-lib:jvmTest
# Produce a fat JAR with all dependencies
./gradlew :kotlin-lib:jvmShadowJar
On Windows, use
gradlew.batinstead of./gradlew.
./gradlew :native-lib:linkNativeReleaseShared
The command produces native-lib/build/bin/native/releaseShared/AngryData.*. The Python demo in python-example/interop.py shows how to load this library with ctypes, call functions such as detectPassport, and clean detected text using the exported utilities.
./gradlew check before opening a pull request.Angry Data Core is distributed under the Apache License 2.0. See LICENSE for details.
| Field | Description |
|---|
value | Extracted token. |
before / after | Up to ten characters of context on each side. |
startPosition / endPosition | Absolute offsets counted from zero. |
matcher | Reference to the matcher that produced the hit. |
Surfaced from shared tags and platforms — no rankings paid for.