Modify owner-sync-pagure to accept multiple tags at once and use multithreading

roles/bodhi2/backend/templates/owner-sync-pagure.j2

@@ -16,9 +16,13 @@ This is heavily based on "owner-sync-pkgdb.j2" which was introduced in commit
 from __future__ import print_function
 import sys
 import os
-from time import sleep
+import argparse
 import ConfigParser
 from urlparse import urljoin
+import multiprocessing.pool
+from math import ceil
+from functools import partial
+
 import requests
 import koji
 
@@ -158,48 +162,99 @@ def get_repo_name_and_arches(tag, version):
     return repo_name, arches
 
 
-def get_pagure_projects(session, namespace):
+def get_pagure_project_name_and_branch(namespace, project_name):
+    """
+    Gets the branches on a project. This function is used for mapping.
+    :param namespace: string of the namespace the project is in
+    :param project: string of the project
+    :return: a tuple containing the string of the project and a list of
+    branches
+    """
+    session = retry_session()
+    project_branches_url = '{0}api/0/{1}/{2}/git/branches'.format(
+        PAGURE_URL, namespace, project_name)
+    project_branches_rv = session.get(
+        project_branches_url, verify=VERIFY, timeout=60)
+
+    # If the project's branches can't be reported, just return no branches and
+    # it will be skipped later on
+    if not project_branches_rv.ok:
+        return project_name, []
+
+    return project_name, project_branches_rv.json()['branches']
+
+
+def get_pagure_project_names_from_page(namespace, page):
+    """
+    Gets the names of all the Pagure projects on a page. This function is to be
+    used for mapping.
+    :param namespace: string of the namespace to query for projects
+    :param page: int of the page to query at
+    :return: list of project names on the page
+    """
     url = urljoin(PAGURE_URL, 'api/0/projects?namespace={0}'.format(namespace))
-    url = url + "&page=1&per_page=50&fork=false&short=true"
-    while url:
+    url = url + '&page={0}&per_page=100&fork=false&short=true'.format(page)
+    session = retry_session()
     response = session.get(url, verify=VERIFY, timeout=120)
     if not bool(response):
         print("Failed to talk to %r %r." % (
             response.request.url, response), file=sys.stderr)
-            break
+        sys.exit(1)
 
-        data = response.json()
-        for project in data['projects']:
-            yield project
-        url = data['pagination']['next']
+    names = set()
+    for project in response.json()['projects']:
+        names.add(project['name'])
+
+    return names
 
 
-def get_project_ownership(tag, namespace):
-    projects = {}
+def get_pagure_project_branches(namespace):
+    """
+    Gets all the branches of all the Pagure projects in the desired namespace
+    :param namespace: string of the namespace to query for projects
+    :return: dictionary in the format of {project_name: [branch_one...]}
+    """
+    first_page_url_path = ('api/0/projects?namespace={0}&fork=false&short=true'
+                           '&page=1&per_page=1'.format(namespace))
+    first_page_url = urljoin(PAGURE_URL, first_page_url_path)
     session = retry_session()
-    for project in get_pagure_projects(session, namespace=namespace):
-        # Check if this project has the branch we are interested in
-        project_branches_url = '{0}api/0/{1}/{2}/git/branches'.format(
-            PAGURE_URL, namespace, project['name'])
-        project_branches_rv = session.get(
-            project_branches_url, verify=VERIFY, timeout=60)
+    first_page_rv = session.get(first_page_url, verify=VERIFY, timeout=120)
+    if not bool(first_page_rv):
+        print("Failed to talk to %r %r." % (
+            first_page_rv.request.url, first_page_rv), file=sys.stderr)
+        sys.exit(1)
+    total_projects = first_page_rv.json()['total_projects']
+    num_pages = int(ceil(total_projects / 100.0))
 
-        # If the project's branches can't be reported, let's skip the project
-        # for now.
-        if not project_branches_rv.ok:
-            continue
+    pool = multiprocessing.pool.ThreadPool(4)
+    # Since we are going to multi-thread, we need to make a partial function
+    # call so that all the function needs is an iterable to run
+    partial_get_pagure_projects_page = partial(
+        get_pagure_project_names_from_page, namespace)
+    project_names_sets = pool.map(partial_get_pagure_projects_page,
+                                  range(1, num_pages + 1))
 
-        project_branches_rv_json = project_branches_rv.json()
+    if project_names_sets:
+        # Combine all the project name sets
+        project_names = list(set.union(*project_names_sets))
+        # Hopefully save some RAM
+        del project_names_sets
+    else:
+        return {}
 
-        # The tag and branch names are the same for "old-style" branches
-        if tag in project_branches_rv_json['branches'] or tag == 'f' + RAWHIDE:
-            yield project['name']
-
-    # This is a special project, not in dist-git, but which needs to be in the package list.
-    yield 'module-build-macros'
+    # Since we are going to multi-thread, we need to make a partial function
+    # call so that all the function needs is an iterable to run
+    partial_get_pagure_project_name_and_branch = partial(
+        get_pagure_project_name_and_branch, namespace)
+    # Get a list of tuples in the form of (project, [branch...]), then convert
+    # that to a dictionary
+    project_names_to_branches = dict(pool.map(
+        partial_get_pagure_project_name_and_branch, project_names))
+    pool.close()
+    return project_names_to_branches
 
 
-def set_koji_ownership(packages, arches):
+def set_koji_ownership(tag, packages, arches):
     koji_options = get_options()
 
     for arch in arches:
@@ -244,14 +299,42 @@ def set_koji_ownership(packages, arches):
 
 
 if __name__ == '__main__':
-    try:
-        tag = sys.argv[1]
-    except IndexError:
-        print('Error: no tag specified', file=sys.stderr)
-        usage()
+    parser = argparse.ArgumentParser(description='Process some integers.')
+    parser.add_argument('tag', nargs='+',
+                        help='tag to update the package list on')
+    args = parser.parse_args()
+    tags = args.tag
 
+    # Get all the info about the tags we are interested in
+    unique_namespaces = set()
+    tag_info = {}
+    for tag in tags:
         namespace, version = get_namespace_and_version_from_tag(tag)
         repo_name, arches = get_repo_name_and_arches(tag, version)
-    pkgs = get_project_ownership(tag, namespace)
-    set_koji_ownership(pkgs, arches)
+        tag_info[tag] = {
+            'namespace': namespace,
+            'version': version,
+            'repo_name': repo_name,
+            'arches': arches
+        }
+        unique_namespaces.add(namespace)
 
+    # Get all the project to branch mappings for every namespace
+    namespace_to_projects = {}
+    for namespace in unique_namespaces:
+        namespace_to_projects[namespace] = \
+            get_pagure_project_branches(namespace)
+
+    for tag, info in tag_info.items():
+        namespace = info['namespace']
+        pkgs = []
+        for pkg, branches in namespace_to_projects[namespace].items():
+            # The tag and branch names are the same for "old-style" branches
+            if tag in branches or tag == ('f' + RAWHIDE):
+                pkgs.append(pkg)
+        # This is a special project, not in dist-git, but which needs to be in
+        # the package list.
+        if namespace == 'rpms':
+            pkgs.append('module-build-macros')
+
+        set_koji_ownership(tag, pkgs, tag_info['arches'])