Find duplicates in jars

find-duplicates.py

#!/usr/bin/env python3.9

from functools import reduce
from pathlib import Path
from subprocess import check_output

import re

m2_repo = Path.home() / '.m2' / 'repository'

def gav(coordinate):
    g, a, _, v, _ = coordinate.split(':', 4)
    return g, a, v

def coordinate_to_path(coordinate):
    g, a, v = gav(coordinate)
    group_path = reduce(lambda a, b: a / b, g.split('.'), m2_repo)
    return group_path / a / v / f'{a}-{v}.jar'

def list_jar_contents(path, include=re.compile('.*class$'), exclude=re.compile('.*module-info.class')):
    lines = check_output(['jar', 'tf', str(path)]).decode('utf-8').split()
    lines = map(lambda x: x.strip(), lines)
    return {l for l in lines if include.match(l) and not exclude.match(l)}

coordinates = [
    'org.scijava:scijava-common:jar:2.83.3:compile',
    'org.scijava:parsington:jar:2.0.0:compile',
    'org.bushe:eventbus:jar:1.4:compile',
    'org.jetbrains.kotlin:kotlin-script-runtime:jar:1.4.30:runtime',
    'org.jetbrains.kotlin:kotlin-stdlib:jar:1.4.30:compile',
    'org.jetbrains.kotlin:kotlin-stdlib-common:jar:1.4.30:compile',
    'org.jetbrains:annotations:jar:13.0:compile',
    'org.jetbrains.kotlin:kotlin-scripting-jsr223:jar:1.4.30:compile',
    'org.jetbrains.kotlin:kotlin-scripting-common:jar:1.4.30:runtime',
    'org.jetbrains.kotlinx:kotlinx-coroutines-core:jar:1.3.8:runtime',
    'org.jetbrains.kotlin:kotlin-scripting-jvm:jar:1.4.30:runtime',
    'org.jetbrains.kotlin:kotlin-scripting-jvm-host:jar:1.4.30:runtime',
    'org.jetbrains.kotlin:kotlin-scripting-compiler-embeddable:jar:1.4.30:runtime',
    'org.jetbrains.kotlin:kotlin-scripting-compiler-impl-embeddable:jar:1.4.30:runtime',
    'junit:junit:jar:4.13:test',
    'org.hamcrest:hamcrest-core:jar:1.3:test',
    'org.jetbrains.kotlin:kotlin-stdlib-jdk8:jar:1.4.30:compile',
    'org.jetbrains.kotlin:kotlin-stdlib-jdk7:jar:1.4.30:compile'
]

additional_jars = {'scripting-kotlin': Path.home() / 'workspace' / 'ntakt' / 'scripting-kotlin' / 'target' / 'scripting-kotlin-0.1.2-SNAPSHOT.jar'}
contents = {c: list_jar_contents(coordinate_to_path(c)) for c in coordinates} | {k: list_jar_contents(v) for k, v in additional_jars.items()}
duplicates = []
jars = tuple(contents.keys())

for idx, c1 in enumerate(jars):
    for c2 in jars[idx+1:]:
        intersection = contents[c1].intersection(contents[c2])
        duplicates.append(((c1, c2), list(intersection)))

for duplicate in duplicates:
    if len(duplicate[1]) > 0:
        print(duplicate)