Add ability to filter repositories by owner

CHANGELOG.md

@@ -9,6 +9,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 - Upgrade project
 - Upgrade dependencies
 - Fix total dependents count (#607)
+- Add capability to filter by owner
 - Refactor update logic in sources_all_df to ensure proper data type handling
 - CI
   - Upgrade MegaLinter

README.md

@@ -245,6 +245,7 @@ _________________
 | -d<br/> --docurl            | String  | _(optional)_ Hyperlink to use when clicking on badge markdown file badge. (Default: link to markdown file)                                                                               |
 | -p<br/> --mergepackages     | String  | _(optional)_ In case of multiple packages, merge their stats in a single one in markdown and json output                                                                                 |
 | -j<br/> --json              | String  | _(optional)_ Output in json format                                                                                                                                                       |
+| -u<br/> --owner             | String  | _(optional)_ If set, filters reposities to keep only those owned by the specified user/organization                                                                                      |
 | -v<br/> --version           | Boolean | _(optional)_ Displays version of github-dependents-info                                                                                                                                  |
 | --verbose                   | Boolean | _(optional)_ Verbose output                                                                                                                                                              |
 

github_dependents_info/__main__.py

@@ -1,9 +1,10 @@
 import logging
 
 import typer
+from rich.console import Console
+
 from github_dependents_info import version
 from github_dependents_info.gh_dependents_info import GithubDependentsInfo
-from rich.console import Console
 
 app = typer.Typer(
     name="github-dependents-info",
@@ -78,6 +79,9 @@ def main(
         is_eager=True,
         help="Prints the version of the github-dependents-info package.",
     ),
+    owner: str = typer.Option(
+        None, "-u", "--owner", help="Filter dependent repositories with a specific owner (ex: oxsecurity)"
+    ),
 ) -> None:
     # Init logger
     if verbose is True:
@@ -110,6 +114,7 @@ def main(
             markdown_file=markdown_file,
             badge_color=badge_color,
             merge_packages=merge_packages,
+            owner=owner,
             time_delay=time_delay,
         )
         # Collect data

github_dependents_info/gh_dependents_info.py

@@ -37,6 +37,7 @@ def __init__(self, repo, **options) -> None:
             if ("csv_directory" in options and options["csv_directory"] is not None)
             else None
         )
+        self.owner = options["owner"] if "owner" in options else None
         self.total_sum = 0
         self.total_public_sum = 0
         self.total_private_sum = 0
@@ -64,10 +65,14 @@ def collect(self):
             # Build start page url
             if package["id"] is not None:
                 url = self.url_init + "?package_id=" + package["id"]
+                if self.owner:
+                    url += "&owner=" + self.owner
                 if self.debug is True:
                     logging.info("Package " + package["name"] + ": browsing " + url + " ...")
             else:
                 url = self.url_init + ""
+                if self.owner:
+                    url += "?owner=" + self.owner
                 if self.debug is True:
                     logging.info("Package " + self.repo + ": browsing" + url + " ...")
             package["url"] = url

tests/test_gh_dependents_info/test_gh_dependents_info.py

@@ -94,3 +94,10 @@ def test_collect_csv_multi_package():
             packages_with_entries += 1
             assert os.path.isfile(csv_directory + os.path.sep + f"dependents_{package['name'].replace('/', '-')}.csv")
         assert packages_with_entries >= 2
+
+def test_collect_stats_owner():
+    repo = "nvuillam/npm-groovy-lint"
+    gh_deps_info = GithubDependentsInfo(repo, debug=True, owner="nvuillam")
+    repo_stats = gh_deps_info.collect()
+    assert repo_stats["public_dependents_number"] < 10
+