cleanup

Author: alexandraBara

nodescraper/cli/cli.py

@@ -30,7 +30,6 @@
 import os
 import platform
 import sys
-from importlib import import_module
 from typing import Optional
 
 import nodescraper
@@ -56,85 +55,6 @@
 from nodescraper.pluginregistry import PluginRegistry
 
 
-def discover_external_plugins():
-    """Discover ext_nodescraper_plugins from all installed packages.
-
-    Returns:
-        list: List of discovered plugin packages
-    """
-    extra_pkgs = []
-    seen_paths = set()  # Track paths to avoid duplicates
-
-    try:
-        import ext_nodescraper_plugins as ext_pkg
-
-        extra_pkgs.append(ext_pkg)
-        if hasattr(ext_pkg, "__file__") and ext_pkg.__file__:
-            seen_paths.add(ext_pkg.__file__)
-    except ImportError:
-        pass
-
-    # Discover ext_nodescraper_plugins from installed packages
-    try:
-        from importlib.metadata import distributions
-
-        for dist in distributions():
-            pkg_name = dist.metadata.get("Name", "")
-            if not pkg_name:
-                continue
-
-            name_variants = [
-                pkg_name.replace("-", "_"),
-                pkg_name.replace("_", "-"),
-            ]
-
-            try:
-                top_level = dist.read_text("top_level.txt")
-                if top_level:
-                    name_variants.extend(top_level.strip().split("\n"))
-            except Exception:
-                pass
-
-            for variant in name_variants:
-                if not variant:
-                    continue
-
-                try:
-                    module_path = f"{variant}.ext_nodescraper_plugins"
-                    ext_pkg = import_module(module_path)
-
-                    # Check if we already have this package (by file path)
-                    pkg_path = getattr(ext_pkg, "__file__", None)
-                    if pkg_path and pkg_path in seen_paths:
-                        continue
-
-                    # Add the package
-                    extra_pkgs.append(ext_pkg)
-                    if pkg_path:
-                        seen_paths.add(pkg_path)
-
-                    break
-
-                except (ImportError, AttributeError, ModuleNotFoundError):
-                    continue
-
-    except Exception:
-        pass
-
-    return extra_pkgs
-
-
-# Fix sys.path[0] if it's the venv/bin directory to avoid breaking editable install discovery
-_original_syspath0 = sys.path[0]
-if _original_syspath0.endswith("/bin") or _original_syspath0.endswith("\\Scripts"):
-    sys.path[0] = ""
-
-extra_pkgs = discover_external_plugins()
-
-# Restore original sys.path[0]
-sys.path[0] = _original_syspath0
-
-
 def build_parser(
     plugin_reg: PluginRegistry,
     config_reg: ConfigRegistry,
@@ -449,7 +369,7 @@ def main(arg_input: Optional[list[str]] = None):
     if arg_input is None:
         arg_input = sys.argv[1:]
 
-    plugin_reg = PluginRegistry(plugin_pkg=extra_pkgs)
+    plugin_reg = PluginRegistry()
 
     config_reg = ConfigRegistry()
     parser, plugin_subparser_map = build_parser(plugin_reg, config_reg)

nodescraper/pluginregistry.py

@@ -24,6 +24,7 @@
 #
 ###############################################################################
 import importlib
+import importlib.metadata
 import inspect
 import pkgutil
 import types
@@ -45,12 +46,14 @@ def __init__(
         self,
         plugin_pkg: Optional[list[types.ModuleType]] = None,
         load_internal_plugins: bool = True,
+        load_entry_point_plugins: bool = True,
     ) -> None:
         """Initialize the PluginRegistry with optional plugin packages.
 
         Args:
             plugin_pkg (Optional[list[types.ModuleType]], optional): The module to search for plugins in. Defaults to None.
             load_internal_plugins (bool, optional): Whether internal plugin should be loaded. Defaults to True.
+            load_entry_point_plugins (bool, optional): Whether to load plugins from entry points. Defaults to True.
         """
         if load_internal_plugins:
             self.plugin_pkg = [internal_plugins, internal_connections, internal_collators]
@@ -70,6 +73,10 @@ def __init__(
             PluginResultCollator, self.plugin_pkg
         )
 
+        if load_entry_point_plugins:
+            entry_point_plugins = self.load_plugins_from_entry_points()
+            self.plugins.update(entry_point_plugins)
+
     @staticmethod
     def load_plugins(
         base_class: type,
@@ -104,3 +111,42 @@ def _recurse_pkg(pkg: types.ModuleType, base_class: type) -> None:
         for pkg in search_modules:
             _recurse_pkg(pkg, base_class)
         return registry
+
+    @staticmethod
+    def load_plugins_from_entry_points() -> dict[str, type]:
+        """Load plugins registered via entry points.
+
+        Returns:
+            dict[str, type]: A dictionary mapping plugin names to their classes.
+        """
+        plugins = {}
+
+        try:
+            # Python 3.10+ supports group parameter
+            try:
+                plugin_entry_points = importlib.metadata.entry_points(group="nodescraper.plugins")  # type: ignore[call-arg]
+            except TypeError:
+                # Python 3.9 - entry_points() returns dict-like object
+                all_entry_points = importlib.metadata.entry_points()
+                plugin_entry_points = all_entry_points.get("nodescraper.plugins", [])  # type: ignore[attr-defined]
+
+            for entry_point in plugin_entry_points:
+                try:
+                    plugin_class = entry_point.load()
+
+                    if (
+                        inspect.isclass(plugin_class)
+                        and issubclass(plugin_class, PluginInterface)
+                        and not inspect.isabstract(plugin_class)
+                    ):
+                        if hasattr(plugin_class, "is_valid") and not plugin_class.is_valid():
+                            continue
+
+                        plugins[plugin_class.__name__] = plugin_class
+                except Exception:
+                    pass
+
+        except Exception:
+            pass
+
+        return plugins