A script to download the `-sources.jar`s for an Android Gradle Plugin version.

download_android_gradle_plugin_srcs.py

#!/usr/bin/env python3
import argparse
import shutil
import zipfile
from pathlib import Path
from tempfile import NamedTemporaryFile
from xml.etree import ElementTree

import requests

URL_TEMPLATE_TEMPLATE = "https://dl.google.com/dl/android/maven2/" \
    "{group}/{artifact}/{version}/{artifact}-{version}{{ending}}"

BASE_GROUP_AND_ARTIFACT = "com.android.tools.build", "gradle"

POM_NAMESPACE = {"pom": "http://maven.apache.org/POM/4.0.0"}


def crawl_sources(plugin_version, destination_dir):
    if destination_dir.exists():
        shutil.rmtree(str(destination_dir))
    destination_dir.mkdir()

    download_queue = [BASE_GROUP_AND_ARTIFACT + (plugin_version, )]
    downloaded = set()
    while download_queue:
        to_download = download_queue.pop()
        if to_download in downloaded:
            continue
        downloaded.add(to_download)
        group, artifact, version = to_download
        print("Handling {}:{}:{}".format(group, artifact, version))
        url_template = URL_TEMPLATE_TEMPLATE.format(
            group=group.replace(".", "/"),
            artifact=artifact,
            version=version,
        )

        # First grab the dependencies for this.
        pom_url = url_template.format(ending=".pom")
        response = requests.get(pom_url)
        if response.status_code == 404:
            print("!!! Cannot find {}:{}:{}".format(group, artifact, version))
            continue
        pom_content = response.content.decode('utf-8')
        pom_root = ElementTree.fromstring(pom_content)
        dependencies = pom_root.find('pom:dependencies', POM_NAMESPACE)
        if dependencies is not None:
            for dependency in dependencies.findall('pom:dependency', POM_NAMESPACE):
                dep_group = _get_single_tag_content(dependency, "groupId")
                dep_artifact = _get_single_tag_content(dependency, "artifactId")
                dep_version = _get_single_tag_content(dependency, "version")
                download_queue.append((dep_group, dep_artifact, dep_version))

        # Now get the content of this.
        sources_url = url_template.format(ending="-sources.jar")
        with NamedTemporaryFile() as temp_file:
            with open(temp_file.name, 'wb') as open_temp_file:
                open_temp_file.write(requests.get(sources_url).content)
            with zipfile.ZipFile(temp_file.name, 'r') as zip_ref:
                zip_ref.extractall(str(destination_dir))


def _get_single_tag_content(current, tag_name):
    tags = current.findall("pom:{}".format(tag_name), POM_NAMESPACE)
    assert len(tags) == 1
    return tags[0].text


def main():
    parser = argparse.ArgumentParser()
    parser.add_argument("version")
    parser.add_argument("--destination", default=str(Path.home()/"tmp_gradle_sources"), required=False)
    args = parser.parse_args()
    crawl_sources(args.version, Path(args.destination))


if __name__ == '__main__':
    main()