fix: ensure 100% progress is logged exactly once (#276)

The progress tracker was logging 100% multiple times - once from
_record_completion() and again from log_final(). Now _record_completion()
skips logging at 100%, leaving that responsibility to log_final().

Also refactors tests from class-based to flat functions and adds
explicit tests for the 100% logging behavior.

Author: johnnygreco

packages/data-designer-engine/src/data_designer/engine/dataset_builders/utils/progress_tracker.py

@@ -77,15 +77,9 @@ def record_failure(self) -> None:
         self._record_completion(success=False)
 
     def log_final(self) -> None:
-        """Log final progress if not already logged at completion."""
+        """Log final progress summary."""
         with self.lock:
-            if self.completed > 0 and self.completed == self.total_records:
-                self._log_progress_unlocked()  # always log 100%
-            elif self.total_records > 0 and self.completed < self.total_records:
-                self._log_progress_unlocked()
-            elif self.completed > 0 and self.completed >= self.next_log_at - self.log_interval:
-                pass  # Already logged at the last interval
-            elif self.completed > 0:
+            if self.completed > 0:
                 self._log_progress_unlocked()
 
     def _record_completion(self, *, success: bool) -> None:
@@ -97,7 +91,7 @@ def _record_completion(self, *, success: bool) -> None:
             else:
                 self.failed += 1
 
-            if self.completed >= self.next_log_at:
+            if self.completed >= self.next_log_at and self.completed < self.total_records:
                 should_log = True
                 while self.next_log_at <= self.completed:
                     self.next_log_at += self.log_interval