Building the extractor

Layout

PipePipeExtractor is a Gradle build with two modules: :extractor (the library) and :timeago-parser. Java toolchain 25, Gradle 9.5. Plugins: java-library, checkstyle, and com.squareup.wire (protobuf code generation, used by SABR). Key dependencies: nanojson, jsoup, okhttp, protobuf-java, commons-lang3 / commons-codec, brotli, Java-WebSocket, and rhino (pinned to 1.7.13 for Android compatibility).

Composite build

You rarely build the extractor in isolation. PipePipeClient wires it in as a composite build:

includeBuild('../PipePipeExtractor') {
    dependencySubstitution {
        substitute module('com.github.TeamNewPipe:NewPipeExtractor') using project(':extractor')
    }
}

So the client compiles the extractor working tree directly, uncommitted changes included. Edit the extractor, build the app, and your changes are in, with no publish or version bump.

Commands

# compile the library
./gradlew :extractor:compileJava

# checkstyle frequently fails on unrelated rules; skip it while iterating
./gradlew :extractor:compileJava -x checkstyleMain

Gradle needs JDK 17+ to launch; the toolchain pulls JDK 25 for compilation. Checkstyle is strict (100-column lines, ignoreFailures = false), so a checkstyle failure is usually style, not a real break.

No tests

This fork has no automated tests. There is no src/test, no JUnit, no mock Downloader, no recorded fixtures. You validate changes by running against the live services, through the app or a small main() that calls NewPipe.init(...) with a real downloader and then drives StreamInfo.getInfo(url) (or the matching list extractor). Checkstyle is the only automated gate.