From 5a7f92eab3700727f53ecd3dc26bb14cbeb02d5f Mon Sep 17 00:00:00 2001
From: Stephen Gallagher <sgallagh@redhat.com>
Date: Nov 18 2022 20:27:07 +0000
Subject: Improve package name detection


Look up the parent repo of forks for the package name.

This requires the user to have a valid Gitlab API token in
their configuration. It will raise an error if it takes this
path and does not have permission.

Note that when https://gitlab.com/gitlab-org/gitlab/-/issues/361952
is fixed in Gitlab, the token will not be required in this case.

Signed-off-by: Stephen Gallagher <sgallagh@redhat.com>

---

diff --git a/src/centpkg/cli.py b/src/centpkg/cli.py
index 61202fb..0d5a58e 100755
--- a/src/centpkg/cli.py
+++ b/src/centpkg/cli.py
@@ -19,6 +19,7 @@ import argparse
 import textwrap
 
 from centpkg.utils import config_get_safely, do_add_remote, do_fork
+import centpkg.utils
 from pyrpkg.cli import cliClient
 from pyrpkg import rpkgError
 from six.moves.urllib_parse import urlparse
@@ -31,6 +32,11 @@ class centpkgClient(cliClient):
     def __init__(self, config, name='centpkg'):
         self.DEFAULT_CLI_NAME = name
 
+        # Save the config for utilities such as get_repo_name()
+        centpkg.utils.dist_git_config = config
+        centpkg.utils.dist_git_config.add_section('__default')
+        centpkg.utils.dist_git_config.set('__default', 'cli_name', name)
+
         super(centpkgClient, self).__init__(config, name)
 
         self.setup_centos_subparsers()
diff --git a/src/centpkg/utils.py b/src/centpkg/utils.py
index 8bc3b4c..1f9524d 100644
--- a/src/centpkg/utils.py
+++ b/src/centpkg/utils.py
@@ -20,6 +20,7 @@ from requests.exceptions import ConnectionError
 from six.moves.configparser import NoOptionError, NoSectionError
 from six.moves.urllib.parse import quote_plus, urlparse
 
+dist_git_config = None
 
 def do_fork(logger, base_url, token, repo_name, namespace, cli_name):
     """
@@ -180,6 +181,54 @@ def config_get_safely(config, section, option):
         raise
 
 
+def get_canonical_repo_name(config, repo_url):
+    """
+    Check whether the current repo is a fork and if so, retrieve the parent
+    fork to get the proper name.
+    """
+
+    # Look up the repo and query for forked_from_project
+    cli_name = config_get_safely(dist_git_config, '__default', 'cli_name')
+    distgit_section = '{0}.distgit'.format(cli_name)
+    distgit_api_base_url = config_get_safely(dist_git_config, distgit_section, "apibaseurl")
+
+    # Make sure the fork comes from the same Gitlab instance
+    parsed_repo_url = urlparse(repo_url)
+    parsed_base_url = urlparse(distgit_api_base_url)
+
+    try:
+        distgit_token = config_get_safely(dist_git_config, distgit_section, 'token')
+
+        api_url = '{0}/api/v4'.format(distgit_api_base_url.rstrip('/'))
+        project_url = '{0}/projects/{1}'.format(api_url, quote_plus(parsed_repo_url.path.lstrip('/')))
+
+        headers = {
+            'PRIVATE-TOKEN': distgit_token,
+            'Accept': 'application/json',
+            'Content-Type': 'application/json'
+        }
+
+        rv = requests.get(project_url, headers=headers)
+        rv.raise_for_status()
+
+        # Extract response json for debugging
+        rv_json = rv.json()
+
+        canonical_repo_name = rv_json['forked_from_project']['name']
+    except KeyError as e:
+        # There was no 'forked_from_project' key, likely meaning the
+        # user lacked permissions to read the API. Usually this means
+        # they haven't supplied a token or it is expired.
+        raise rpkgError("Insufficient Gitlab API permissions. Missing token?")
+
+    except Exception as e:
+        # For any other exception, just fall back to using the last segment
+        # of the URL path.
+        canonical_repo_name = parsed_repo_url.path.split('/')[-1]
+
+    # Chop off a trailing .git if any
+    return canonical_repo_name.rsplit('.git', 1)[0]
+
 def get_repo_name(name, org='rpms'):
     """
     Try to parse the repository name in case it is a git url.
@@ -202,7 +251,7 @@ def get_repo_name(name, org='rpms'):
     if name.startswith(org):
         return name
 
-    parsed = '/'.join(name.split('/')[1:])
-    repo_name = parsed.split('_')[-1:][0]
+    # This is probably a renamed fork, so try to find the fork's parent
+    repo_name = get_canonical_repo_name(dist_git_config, name)
 
     return '%s/%s' % (org, repo_name)