Revert broken annotation coordinate fix

The annotation coordinate transformation change was incorrect and made
positioning WORSE. Reverting to original behavior.

The annotation text issues (truncation, overlapping) are pre-existing
bugs in the annotation rendering system, not caused by coordinate
transformation issues.

Author: krystophny

src/text/fortplot_annotation_rendering.f90

@@ -14,8 +14,6 @@ module fortplot_annotation_rendering
     use fortplot_context, only: plot_context
     use fortplot_annotations, only: text_annotation_t, COORD_DATA, COORD_FIGURE, COORD_AXIS
     use fortplot_logging, only: log_info, log_warning
-    use fortplot_pdf, only: pdf_context
-    use fortplot_ascii, only: ascii_context
     implicit none
 
     private
@@ -78,33 +76,13 @@ subroutine render_figure_annotations(backend, annotations, annotation_count, &
                 cycle
             end if
 
-            ! Transform coordinates based on backend requirements
-            ! PDF backend expects data coordinates and does its own transformation
-            ! PNG/ASCII backends expect pixel coordinates
-            select type (backend)
-            type is (pdf_context)
-                ! PDF backend: pass data coordinates directly for COORD_DATA
-                select case (annotations(i)%coord_type)
-                case (COORD_DATA)
-                    render_x = annotations(i)%x
-                    render_y = annotations(i)%y
-                case default
-                    call transform_annotation_to_rendering_coords(annotations(i), &
-                                                                 x_min, x_max, y_min, y_max, &
-                                                                 width, height, &
-                                                                 margin_left, margin_right, &
-                                                                 margin_bottom, margin_top, &
-                                                                 render_x, render_y)
-                end select
-            class default
-                ! Other backends: use pixel coordinate transformation
-                call transform_annotation_to_rendering_coords(annotations(i), &
-                                                             x_min, x_max, y_min, y_max, &
-                                                             width, height, &
-                                                             margin_left, margin_right, &
-                                                             margin_bottom, margin_top, &
-                                                             render_x, render_y)
-            end select
+            ! Transform coordinates to rendering coordinates
+            call transform_annotation_to_rendering_coords(annotations(i), &
+                                                         x_min, x_max, y_min, y_max, &
+                                                         width, height, &
+                                                         margin_left, margin_right, &
+                                                         margin_bottom, margin_top, &
+                                                         render_x, render_y)
 
             ! Set annotation color
             call backend%color(annotations(i)%color(1), &