feat: Enhance analytics opt-in logic with user sampling and explicit control

aider/analytics.py

@@ -62,8 +62,21 @@ class Analytics:
             self.permanently_disable = True
             self.save_data()
 
-    def need_to_ask(self):
+    def need_to_ask(self, args_analytics):
-        return not self.asked_opt_in and not self.permanently_disable
+        if args_analytics is False:
+            return False
+
+        could_ask = not self.asked_opt_in and not self.permanently_disable
+        if not could_ask:
+            return False
+
+        if args_analytics is True:
+            return True
+
+        assert args_analytics is None, args_analytics
+
+        # ask 1/16 of the users
+        return self.user_id < "1"
 
     def get_data_file_path(self):
         data_file = Path.home() / ".aider" / "analytics.json"

aider/args.py

@@ -562,8 +562,8 @@ def get_parser(default_config_files, git_root):
     group.add_argument(
         "--analytics",
         action=argparse.BooleanOptionalAction,
-        default=False,
+        default=None,
-        help="Enable/disable analytics for one session (default: False)",
+        help="Enable/disable analytics for current session (default: random)",
     )
     group.add_argument(
         "--analytics-log",

aider/main.py

@@ -506,8 +506,8 @@ def main(argv=None, input=None, output=None, force_git_root=None, return_coder=F
         io.tool_warning("Terminal does not support pretty output (UnicodeDecodeError)")
 
     analytics = Analytics(logfile=args.analytics_log, permanently_disable=args.analytics_disable)
-    if args.analytics:
+    if args.analytics is not False:
-        if analytics.need_to_ask():
+        if analytics.need_to_ask(args.analytics):
             io.tool_output(
                 "Aider respects your privacy and never collects your code, chat messages, keys or"
                 " personal info."

tests/basic/test_analytics.py

@@ -91,10 +91,17 @@ def test_system_info(temp_data_dir):
 
 def test_need_to_ask(temp_data_dir):
     analytics = Analytics()
-    assert analytics.need_to_ask() is True
+    assert analytics.need_to_ask(True) is True
+    assert analytics.need_to_ask(False) is False
+
+    analytics.user_id = "111"
+    assert analytics.need_to_ask(None) is False
+
+    analytics.user_id = "000"
+    assert analytics.need_to_ask(None) is True
 
     analytics.asked_opt_in = True
-    assert analytics.need_to_ask() is False
+    assert analytics.need_to_ask(True) is False
 
     analytics.permanently_disable = True
-    assert analytics.need_to_ask() is False
+    assert analytics.need_to_ask(True) is False