Python script to download jar file info from Maven using SHA1 checksums and print Maven dependency records

pommaker.py

#!/usr/bin/env python

'''
Script to generate the pom.xml dependency section for a bunch of jar files.
'''

import hashlib
import sys
import os
import json
import httplib

def get_sha1_sum(fname):
    ''' generates a SHA1 sum for the given file '''
    with open(fname, 'rb') as f:
        return hashlib.sha1(f.read()).hexdigest()

def get_dependency(group_id, artifact_id, version):
    ''' creates XML dependency element '''
    return '<dependency>\n<groupId>%s</groupId>\n<artifactId>%s</artifactId>\n<version>%s</version>\n</dependency>\n' % (group_id, artifact_id, version)

def get_dep_xml(sha1sum):
    ''' looks up the jar file SHA1 checksum at search.maven.org and grabs the XML dependency info '''
    search_uri = '/solrsearch/select?q=1%%3A%%22%s%%22&rows=20&wt=json' % sha1sum
    conn = httplib.HTTPConnection('search.maven.org')
    conn.request('GET', search_uri)
    response = conn.getresponse()
    if response.status == 200:
        search_result = json.loads(response.read())
        if search_result['response']['numFound'] == 1:
            project = search_result['response']['docs'][0]
            group_id = project.get('g')
            artifact_id = project.get('a')
            version_id = project.get('v')
            return get_dependency(group_id, artifact_id, version_id)
    return None

if __name__ == '__main__':
    if len(sys.argv) > 1:
        jar_dir = sys.argv[1]
    else:
        jar_dir = raw_input('Please specify a path to the jar file directory: ')
    jar_files = [x for x in os.listdir(jar_dir) if x.endswith('.jar')]
    for jar_file in jar_files:
        sha1sum = get_sha1_sum(os.path.join(jar_dir, jar_file))
        dep_xml = get_dep_xml(sha1sum)
        if dep_xml:
            print dep_xml