meerk40t 0.9.3001__py2.py3-none-any.whl → 0.9.7010__py2.py3-none-any.whl

meerk40t/tools/rasterplotter.py

@@ -1,547 +1,1660 @@
-"""
-The RasterPlotter is a plotter that maps particular raster pixels to directional and raster
-methods. This class should be expanded to cover most raster situations.
-
-The X_AXIS / Y_AXIS flag determines whether we raster across the X_AXIS or Y_AXIS. Standard
-right-to-left rastering starting at the top edge on the left is the default. This would be
-in the upper left hand corner proceeding right, and stepping towards bottom each scanline.
-
-If the X_AXIS is set, the edge being used can be either TOP or BOTTOM. That flag means edge
-with X_AXIS rastering. However, in Y_AXIS rastering the start edge can either be right-edge
-or left-edge, for this the RIGHT and LEFT flags are used. However, the start point on either
-edge can be TOP or BOTTOM if we're starting on a RIGHT or LEFT edge. So those flags mark
-that. The same is true for RIGHT or LEFT on a TOP or BOTTOM edge.
-
-The TOP, RIGHT, LEFT, BOTTOM combined give you the starting corner.
-
-The rasters can either be BIDIRECTIONAL or UNIDIRECTIONAL meaning they raster on both swings
-or only on forward swing.
-"""
-
-
-class RasterPlotter:
-    def __init__(
-        self,
-        data,
-        width,
-        height,
-        horizontal=True,
-        start_minimum_y=True,
-        start_minimum_x=True,
-        bidirectional=True,
-        use_integers=True,
-        skip_pixel=0,
-        overscan=0,
-        offset_x=0,
-        offset_y=0,
-        step_x=1,
-        step_y=1,
-        filter=None,
-    ):
-        """
-        Initialization for the Raster Plotter function. This should set all the needed parameters for plotting.
-
-        @param data: pixel data accessed through data[x,y] parameters
-        @param width: Width of the given data.
-        @param height: Height of the given data.
-        @param horizontal: Flags for how the pixel traversal should be conducted.
-        @param start_minimum_y: Flags for how the pixel traversal should be conducted.
-        @param start_minimum_x: Flags for how the pixel traversal should be conducted.
-        @param bidirectional: Flags for how the pixel traversal should be conducted.
-        @param skip_pixel: Skip pixel. If this value is the pixel value, we skip travel in that direction.
-        @param use_integers: return integer values rather than floating point values.
-        @param overscan: The extra amount of padding to add to the end scanline.
-        @param offset_x: The offset in x of the rastering location. This will be added to x values returned in plot.
-        @param offset_y: The offset in y of the rastering location. This will be added to y values returned in plot.
-        @param step_x: The amount units per pixel.
-        @param step_y: The amount scanline gap.
-        @param filter: Pixel filter is called for each pixel to transform or alter it as needed. The actual
-                            implementation is agnostic regarding what data is provided. The filter is expected
-                            to convert the data[x,y] into some form which will be expressed by plot. Unless skipped as
-                            part of the skip pixel.
-        """
-        self.data = data
-        self.width = width
-        self.height = height
-        self.horizontal = horizontal
-        self.start_minimum_y = start_minimum_y
-        self.start_minimum_x = start_minimum_x
-        self.bidirectional = bidirectional
-        self.use_integers = use_integers
-        self.skip_pixel = skip_pixel
-        if horizontal:
-            self.overscan = round(overscan / float(step_x))
-        else:
-            self.overscan = round(overscan / float(step_y))
-        self.offset_x = offset_x
-        self.offset_y = offset_y
-        self.step_x = step_x
-        self.step_y = step_y
-        self.filter = filter
-        self.initial_x, self.initial_y = self.calculate_first_pixel()
-        self.final_x, self.final_y = self.calculate_last_pixel()
-
-    def px(self, x, y):
-        """
-        Returns the filtered pixel
-
-        @param x:
-        @param y:
-        @return: Filtered Pixel
-        """
-        if 0 <= y < self.height and 0 <= x < self.width:
-            if self.filter is None:
-                return self.data[x, y]
-            return self.filter(self.data[x, y])
-        raise IndexError
-
-    def leftmost_not_equal(self, y):
-        """
-        Determine the leftmost pixel that is not equal to the skip_pixel value.
-
-        if all pixels skipped returns None
-        """
-        for x in range(0, self.width):
-            pixel = self.px(x, y)
-            if pixel != self.skip_pixel:
-                return x
-        return None
-
-    def topmost_not_equal(self, x):
-        """
-        Determine the topmost pixel that is not equal to the skip_pixel value
-
-        if all pixels skipped returns None
-        """
-        for y in range(0, self.height):
-            pixel = self.px(x, y)
-            if pixel != self.skip_pixel:
-                return y
-        return None
-
-    def rightmost_not_equal(self, y):
-        """
-        Determine the rightmost pixel that is not equal to the skip_pixel value
-
-        if all pixels skipped returns None
-        """
-        for x in range(self.width - 1, -1, -1):
-            pixel = self.px(x, y)
-            if pixel != self.skip_pixel:
-                return x
-        return None
-
-    def bottommost_not_equal(self, x):
-        """
-        Determine the bottommost pixel that is not equal to the skip_pixel value
-
-        if all pixels skipped returns None
-        """
-        for y in range(self.height - 1, -1, -1):
-            pixel = self.px(x, y)
-            if pixel != self.skip_pixel:
-                return y
-        return None
-
-    def nextcolor_left(self, x, y, default=None):
-        """
-        Determine the next pixel change going left from the (x,y) point.
-        If no next pixel is found default is returned.
-        """
-        if x <= -1:
-            return default
-        if x == 0:
-            return -1
-        if x == self.width:
-            return self.width - 1
-        if self.width < x:
-            return self.width
-
-        v = self.px(x, y)
-        for ix in range(x, -1, -1):
-            pixel = self.px(ix, y)
-            if pixel != v:
-                return ix
-        return 0
-
-    def nextcolor_top(self, x, y, default=None):
-        """
-        Determine the next pixel change going top from the (x,y) point.
-        If no next pixel is found default is returned.
-        """
-        if y <= -1:
-            return default
-        if y == 0:
-            return -1
-        if y == self.height:
-            return self.height - 1
-        if self.height < y:
-            return self.height
-
-        v = self.px(x, y)
-        for iy in range(y, -1, -1):
-            pixel = self.px(x, iy)
-            if pixel != v:
-                return iy
-        return 0
-
-    def nextcolor_right(self, x, y, default=None):
-        """
-        Determine the next pixel change going right from the (x,y) point.
-        If no next pixel is found default is returned.
-        """
-        if x < -1:
-            return -1
-        if x == -1:
-            return 0
-        if x == self.width - 1:
-            return self.width
-        if self.width <= x:
-            return default
-
-        v = self.px(x, y)
-        for ix in range(x, self.width):
-            pixel = self.px(ix, y)
-            if pixel != v:
-                return ix
-        return self.width - 1
-
-    def nextcolor_bottom(self, x, y, default=None):
-        """
-        Determine the next pixel change going bottom from the (x,y) point.
-        If no next pixel is found default is returned.
-        """
-        if y < -1:
-            return -1
-        if y == -1:
-            return 0
-        if y == self.height - 1:
-            return self.height
-        if self.height <= y:
-            return default
-
-        v = self.px(x, y)
-        for iy in range(y, self.height):
-            pixel = self.px(x, iy)
-            if pixel != v:
-                return iy
-        return self.height - 1
-
-    def calculate_next_horizontal_pixel(self, y, dy=1, leftmost_pixel=True):
-        """
-        Find the horizontal extreme at the given y-scanline, stepping by dy in the target image.
-        This can be done on either the leftmost (True) or rightmost (False).
-
-        @param y: y-scanline
-        @param dy: dy-step amount (usually should be -1 or 1)
-        @param leftmost_pixel: find pixel from the left
-        @return:
-        """
-        try:
-            if leftmost_pixel:
-                while True:
-                    x = self.leftmost_not_equal(y)
-                    if x is not None:
-                        break
-                    y += dy
-            else:
-                while True:
-                    x = self.rightmost_not_equal(y)
-                    if x is not None:
-                        break
-                    y += dy
-        except IndexError:
-            # Remaining image is blank
-            return None, None
-        return x, y
-
-    def calculate_next_vertical_pixel(self, x, dx=1, topmost_pixel=True):
-        """
-        Find the vertical extreme at the given x-scanline, stepping by dx in the target image.
-        This can be done on either the topmost (True) or bottommost (False).
-
-        @param x: x-scanline
-        @param dx: dx-step amount (usually should be -1 or 1)
-        @param topmost_pixel: find the pixel from the top
-        @return:
-        """
-        try:
-            if topmost_pixel:
-                while True:
-                    y = self.topmost_not_equal(x)
-                    if y is not None:
-                        break
-                    x += dx
-            else:
-                while True:
-                    # find that the bottommost pixel in that row.
-                    y = self.bottommost_not_equal(x)
-
-                    if y is not None:
-                        # This is a valid pixel.
-                        break
-                    # No pixel in that row was valid. Move to the next row.
-                    x += dx
-        except IndexError:
-            # Remaining image was blank, there are no more relevant pixels.
-            return None, None
-        return x, y
-
-    def calculate_first_pixel(self):
-        """
-        Find the first non-skipped pixel in the rastering.
-
-        This takes into account the traversal values of X_AXIS or Y_AXIS and BOTTOM and RIGHT
-        The start edge and the start point.
-
-        @return: x,y coordinates of first pixel.
-        """
-        if self.horizontal:
-            y = 0 if self.start_minimum_y else self.height - 1
-            dy = 1 if self.start_minimum_y else -1
-            x, y = self.calculate_next_horizontal_pixel(y, dy, self.start_minimum_x)
-            return x, y
-        else:
-            x = 0 if self.start_minimum_x else self.width - 1
-            dx = 1 if self.start_minimum_x else -1
-            x, y = self.calculate_next_vertical_pixel(x, dx, self.start_minimum_y)
-            return x, y
-
-    def calculate_last_pixel(self):
-        """
-        Find the last non-skipped pixel in the rastering.
-
-        First and last scanlines start from the same side when scanline count is odd
-
-        @return: x,y coordinates of last pixel.
-        """
-        if self.horizontal:
-            y = self.height - 1 if self.start_minimum_y else 0
-            dy = -1 if self.start_minimum_y else 1
-            start_on_left = (
-                self.start_minimum_x if self.width & 1 else not self.start_minimum_x
-            )
-            x, y = self.calculate_next_horizontal_pixel(y, dy, start_on_left)
-            return x, y
-        else:
-            x = self.width - 1 if self.start_minimum_x else 0
-            dx = -1 if self.start_minimum_x else 1
-            start_on_top = (
-                self.start_minimum_y if self.height & 1 else not self.start_minimum_y
-            )
-            x, y = self.calculate_next_vertical_pixel(x, dx, start_on_top)
-            return x, y
-
-    def initial_position(self):
-        """
-        Returns raw initial position for the relevant pixel within the data.
-        @return: initial position within the data.
-        """
-        if self.use_integers:
-            return int(round(self.initial_x)), int(round(self.initial_y))
-        else:
-            return self.initial_x, self.initial_y
-
-    def initial_position_in_scene(self):
-        """
-        Returns the initial position for this within the scene. Taking into account start corner, and step size.
-        @return: initial position within scene. The first plot location.
-        """
-        if self.initial_x is None:  # image is blank.
-            if self.use_integers:
-                return int(round(self.offset_x)), int(round(self.offset_y))
-            else:
-                return self.offset_x, self.offset_y
-        if self.use_integers:
-            return (
-                int(round(self.offset_x + self.initial_x * self.step_x)),
-                int(round(self.offset_y + self.initial_y * self.step_y)),
-            )
-        else:
-            return (
-                self.offset_x + self.initial_x * self.step_x,
-                self.offset_y + self.initial_y * self.step_y,
-            )
-
-    def final_position_in_scene(self):
-        """
-        Returns best guess of final position relative to the scene offset. Taking into account start corner, and parity
-        of the width and height.
-        @return:
-        """
-        if self.final_x is None:  # image is blank.
-            if self.use_integers:
-                return int(round(self.offset_x)), int(round(self.offset_y))
-            else:
-                return self.offset_x, self.offset_y
-        if self.use_integers:
-            return (
-                int(round(self.offset_x + self.final_x * self.step_x)),
-                int(round(self.offset_y + self.final_y * self.step_y)),
-            )
-        else:
-            return (
-                self.offset_x + self.final_x * self.step_x,
-                self.offset_y + self.final_y * self.step_y,
-            )
-
-    def plot(self):
-        """
-        Plot the values yielded by following the given raster plotter in the traversal defined.
-        """
-        offset_x = self.offset_x
-        offset_y = self.offset_y
-        step_x = self.step_x
-        step_y = self.step_y
-        if self.initial_x is None:
-            # There is no image.
-            return
-        if self.use_integers:
-            for x, y, on in self._plot_pixels():
-                yield int(round(offset_x + step_x * x)), int(
-                    round(offset_y + y * step_y)
-                ), on
-        else:
-            for x, y, on in self._plot_pixels():
-                yield offset_x + step_x * x, offset_y + y * step_y, on
-
-    def _plot_pixels(self):
-        if self.horizontal:
-            yield from self._plot_horizontal()
-        else:
-            yield from self._plot_vertical()
-
-    def _plot_vertical(self):
-        """
-        This code is for vertical rastering.
-
-        @return:
-        """
-        width = self.width
-        unidirectional = not self.bidirectional
-        skip_pixel = self.skip_pixel
-
-        x, y = self.initial_position()
-        dx = 1 if self.start_minimum_x else -1
-        dy = 1 if self.start_minimum_y else -1
-
-        yield x, y, 0
-        while 0 <= x < width:
-            lower_bound = self.topmost_not_equal(x)
-            if lower_bound is None:
-                x += dx
-                yield x, y, 0
-                continue
-            upper_bound = self.bottommost_not_equal(x)
-            traveling_bottom = self.start_minimum_y if unidirectional else dy >= 0
-            next_traveling_bottom = self.start_minimum_y if unidirectional else dy <= 0
-
-            next_x, next_y = self.calculate_next_vertical_pixel(
-                x + dx, dx, topmost_pixel=next_traveling_bottom
-            )
-            if next_y is not None:
-                # If we have a next scanline, we must end after the last pixel of that scanline too.
-                upper_bound = max(next_y, upper_bound) + self.overscan
-                lower_bound = min(next_y, lower_bound) - self.overscan
-
-            if traveling_bottom:
-                while y < upper_bound:
-                    try:
-                        pixel = self.px(x, y)
-                    except IndexError:
-                        pixel = 0
-                    y = self.nextcolor_bottom(x, y, upper_bound)
-                    y = min(y, upper_bound)
-                    if pixel == skip_pixel:
-                        yield x, y, 0
-                    else:
-                        yield x, y, pixel
-            else:
-                while lower_bound < y:
-                    try:
-                        pixel = self.px(x, y)
-                    except IndexError:
-                        pixel = 0
-                    y = self.nextcolor_top(x, y, lower_bound)
-                    y = max(y, lower_bound)
-                    if pixel == skip_pixel:
-                        yield x, y, 0
-                    else:
-                        yield x, y, pixel
-
-            if next_y is None:
-                # remaining image is blank, we stop right here.
-                return
-            yield next_x, y, 0
-            if y != next_y:
-                yield next_x, next_y, 0
-            x = next_x
-            y = next_y
-            dy = -dy
-
-    def _plot_horizontal(self):
-        """
-        This code is horizontal rastering.
-
-        @return:
-        """
-        height = self.height
-        unidirectional = not self.bidirectional
-        skip_pixel = self.skip_pixel
-
-        x, y = self.initial_position()
-        dx = 1 if self.start_minimum_x else -1
-        dy = 1 if self.start_minimum_y else -1
-        yield x, y, 0
-        while 0 <= y < height:
-            lower_bound = self.leftmost_not_equal(y)
-            if lower_bound is None:
-                y += dy
-                yield x, y, 0
-                continue
-            upper_bound = self.rightmost_not_equal(y)
-            traveling_right = self.start_minimum_x if unidirectional else dx >= 0
-            next_traveling_right = self.start_minimum_x if unidirectional else dx <= 0
-
-            next_x, next_y = self.calculate_next_horizontal_pixel(
-                y + dy, dy, leftmost_pixel=next_traveling_right
-            )
-            if next_x is not None:
-                # If we have a next scanline, we must end after the last pixel of that scanline too.
-                upper_bound = max(next_x, upper_bound) + self.overscan
-                lower_bound = min(next_x, lower_bound) - self.overscan
-
-            if traveling_right:
-                while x < upper_bound:
-                    try:
-                        pixel = self.px(x, y)
-                    except IndexError:
-                        pixel = 0
-                    x = self.nextcolor_right(x, y, upper_bound)
-                    x = min(x, upper_bound)
-                    if pixel == skip_pixel:
-                        yield x, y, 0
-                    else:
-                        yield x, y, pixel
-            else:
-                while lower_bound < x:
-                    try:
-                        pixel = self.px(x, y)
-                    except IndexError:
-                        pixel = 0
-                    x = self.nextcolor_left(x, y, lower_bound)
-                    x = max(x, lower_bound)
-                    if pixel == skip_pixel:
-                        yield x, y, 0
-                    else:
-                        yield x, y, pixel
-
-            if next_y is None:
-                # remaining image is blank, we stop right here.
-                return
-            yield x, next_y, 0
-            if x != next_x:
-                yield next_x, next_y, 0
-            x = next_x
-            y = next_y
-            dx = -dx
+"""
+The RasterPlotter is a plotter that maps particular raster pixels to directional and raster
+methods. This class should be expanded to cover most raster situations.
+
+The X_AXIS / Y_AXIS flag determines whether we raster across the X_AXIS or Y_AXIS. Standard
+right-to-left rastering starting at the top edge on the left is the default. This would be
+in the upper left hand corner proceeding right, and stepping towards bottom each scanline.
+
+If the X_AXIS is set, the edge being used can be either TOP or BOTTOM. That flag means edge
+with X_AXIS rastering. However, in Y_AXIS rastering the start edge can either be right-edge
+or left-edge, for this the RIGHT and LEFT flags are used. However, the start point on either
+edge can be TOP or BOTTOM if we're starting on a RIGHT or LEFT edge. So those flags mark
+that. The same is true for RIGHT or LEFT on a TOP or BOTTOM edge.
+
+The TOP, RIGHT, LEFT, BOTTOM combined give you the starting corner.
+
+The rasters can either be BIDIRECTIONAL or UNIDIRECTIONAL meaning they raster on both swings
+or only on forward swing.
+"""
+
+import numpy as np
+from math import sqrt
+from time import perf_counter, sleep
+from meerk40t.constants import (
+    RASTER_T2B,
+    RASTER_B2T,
+    RASTER_R2L,
+    RASTER_L2R,
+    RASTER_HATCH,
+    RASTER_GREEDY_H,
+    RASTER_GREEDY_V,
+    RASTER_CROSSOVER,
+    RASTER_SPIRAL,
+)
+
+
+class RasterPlotter:
+    def __init__(
+        self,
+        data,
+        width,
+        height,
+        direction=0,
+        horizontal=True,
+        start_minimum_y=True,
+        start_minimum_x=True,
+        bidirectional=True,
+        use_integers=True,
+        skip_pixel=0,
+        overscan=0,
+        offset_x=0,
+        offset_y=0,
+        step_x=1,
+        step_y=1,
+        filter=None,
+        laserspot=0,
+        special=None,
+    ):
+        """
+        Initialization for the Raster Plotter function. This should set all the needed parameters for plotting.
+
+        @param data: pixel data accessed through data[x,y] parameters
+        @param width: Width of the given data.
+        @param height: Height of the given data.
+        @param horizontal: Flags for how the pixel traversal should be conducted.
+        @param start_minimum_y: Flags for how the pixel traversal should be conducted.
+        @param start_minimum_x: Flags for how the pixel traversal should be conducted.
+        @param bidirectional: Flags for how the pixel traversal should be conducted.
+        @param skip_pixel: Skip pixel. If this value is the pixel value, we skip travel in that direction.
+        @param use_integers: return integer values rather than floating point values.
+        @param overscan: The extra amount of padding to add to the end scanline.
+        @param offset_x: The offset in x of the rastering location. This will be added to x values returned in plot.
+        @param offset_y: The offset in y of the rastering location. This will be added to y values returned in plot.
+        @param step_x: The amount units per pixel.
+        @param step_y: The amount scanline gap.
+        @param filter: Pixel filter is called for each pixel to transform or alter it as needed. The actual
+                    implementation is agnostic regarding what data is provided. The filter is expected
+                    to convert the data[x,y] into some form which will be expressed by plot. Unless skipped as
+                    part of the skip pixel.
+        @param direction: 0 top 2 bottom, 1 bottom 2 top, 2 right 2 left, 3 left 2 right, 4 greedy path optimisation, 5 crossover
+                    a) greedy will keep the main direction, a bit time intensive for larger images
+                    b) crossover draws first all majority lines, then all majority columns
+                       (to be decided by looping over the matrix and looking at the amount
+                       of black pixels on the same line / on the same column), timewise okayish
+        @param laserspot: the laserbeam diameter in pixels (low dpi = irrelevant, high dpi very relevant)
+        @param special: a dict of special treatment instructions for the different algorithms
+        """
+        # Don't try to plot from two different sources at the same time...
+        self._locked = False
+
+        if special is None:
+            special = {}
+        self.debug_level = 0 # 0 Nothing, 1 file creation, 2 file + summary, 3 file + summary + details
+        self.data = data
+        self.width = width
+        self.height = height
+        self.direction = direction
+        self._cache = None
+        parameters = {
+            # Provide an override for the minimumx / minimumy / horizontal / bidirectional
+            RASTER_T2B: (None, True, True, None), # top to bottom
+            RASTER_B2T: (None, False, True, None), # bottom to top
+            RASTER_R2L: (False, None, False, None), # right to left
+            RASTER_L2R: (True, None, False, None), # left to right
+            RASTER_HATCH: (None, None, None, None), # crossraster (one of the two)
+            RASTER_GREEDY_H: (None, None, None, True), # greedy neighbour horizontal
+            RASTER_GREEDY_V: (None, None, None, True), # greedy neighbour
+            RASTER_CROSSOVER: (None, None, None, True), # true crossover
+        }
+        def_x, def_y, def_hor, def_bidir = parameters.get(direction, (None, None, None, None))
+        self.start_minimum_x = start_minimum_x if def_x is None else def_x
+        self.start_minimum_y = start_minimum_y if def_y is None else def_y
+        self.horizontal = horizontal if def_hor is None else def_hor
+        self.bidirectional = bidirectional if def_bidir is None else def_bidir
+
+        self.use_integers = use_integers
+        self.skip_pixel = skip_pixel
+        # laserspot = width in pixels, so the surrounding logic needs to calculate it
+        # We consider an overlap only in the enclosed square of the circle
+        # and calculate the overlap in pixels to the left / to the right
+        self.overlap = int(laserspot / sqrt(2) / 2)
+        # self.overlap = 1
+        self.special = dict(special) # Copy it so it wont be changed
+        if horizontal:
+            self.overscan = round(overscan / float(step_x))
+        else:
+            self.overscan = round(overscan / float(step_y))
+        self.offset_x = offset_x
+        self.offset_y = offset_y
+        self.step_x = step_x
+        self.step_y = step_y
+        self.filter = filter
+        self.initial_x, self.initial_y = self.calculate_first_pixel()
+        self.final_x, self.final_y = self.calculate_last_pixel()
+        self._distance_travel = 0
+        self._distance_burn = 0
+
+    def __repr__(self):
+        methods = (
+            "Top2Bottom",
+            "Bottom2Top",
+            "Right2Left",
+            "Left2Right",
+            "Hatch",
+            "Greedy Neighbor Hor",
+            "Greedy Neighbor Ver",
+            "Crossover",
+            "Spiral",
+        )
+        if 0 <= self.direction < len(methods):
+            s_meth = f"Rasterplotter ({self.width}x{self.height}): {methods[self.direction]} ({self.direction})"
+        else:
+            s_meth = f"Rasterplotter ({self.width}x{self.height}): Unknown {self.direction}"
+        s_direc = 'Bidirectional' if self.bidirectional else 'Unidirectional'
+        s_axis = 'horizontal' if self.horizontal else 'vertical'
+        s_ystart = 'top' if self.start_minimum_y else 'bottom'
+        s_xstart = 'left' if self.start_minimum_x else 'right'
+        return f"{s_meth}, {s_direc} {s_axis} plot starting at {s_ystart}-{s_xstart}"
+
+    def reset(self):
+        self._cache = None
+
+    def px(self, x, y):
+        """
+        Returns the filtered pixel
+
+        @param x:
+        @param y:
+        @return: Filtered Pixel
+        """
+        if 0 <= y < self.height and 0 <= x < self.width:
+            if self.filter is None:
+                return self.data[x, y]
+            return self.filter(self.data[x, y])
+        raise IndexError
+
+    def leftmost_not_equal(self, y):
+        """
+        Determine the leftmost pixel that is not equal to the skip_pixel value.
+
+        if all pixels skipped returns None
+        """
+        for x in range(0, self.width):
+            pixel = self.px(x, y)
+            if pixel != self.skip_pixel:
+                return x
+        return None
+
+    def topmost_not_equal(self, x):
+        """
+        Determine the topmost pixel that is not equal to the skip_pixel value
+
+        if all pixels skipped returns None
+        """
+        for y in range(0, self.height):
+            pixel = self.px(x, y)
+            if pixel != self.skip_pixel:
+                return y
+        return None
+
+    def rightmost_not_equal(self, y):
+        """
+        Determine the rightmost pixel that is not equal to the skip_pixel value
+
+        if all pixels skipped returns None
+        """
+        for x in range(self.width - 1, -1, -1):
+            pixel = self.px(x, y)
+            if pixel != self.skip_pixel:
+                return x
+        return None
+
+    def bottommost_not_equal(self, x):
+        """
+        Determine the bottommost pixel that is not equal to the skip_pixel value
+
+        if all pixels skipped returns None
+        """
+        for y in range(self.height - 1, -1, -1):
+            pixel = self.px(x, y)
+            if pixel != self.skip_pixel:
+                return y
+        return None
+
+    @property
+    def distance_travel(self):
+        return self._distance_travel
+
+    @property
+    def distance_burn(self):
+        return self._distance_burn
+
+    def nextcolor_left(self, x, y, default=None):
+        """
+        Determine the next pixel change going left from the (x,y) point.
+        If no next pixel is found default is returned.
+        """
+        if x <= -1:
+            return default
+        if x == 0:
+            return -1
+        if x == self.width:
+            return self.width - 1
+        if self.width < x:
+            return self.width
+
+        v = self.px(x, y)
+        for ix in range(x, -1, -1):
+            pixel = self.px(ix, y)
+            if pixel != v:
+                return ix
+        return 0
+
+    def nextcolor_top(self, x, y, default=None):
+        """
+        Determine the next pixel change going top from the (x,y) point.
+        If no next pixel is found default is returned.
+        """
+        if y <= -1:
+            return default
+        if y == 0:
+            return -1
+        if y == self.height:
+            return self.height - 1
+        if self.height < y:
+            return self.height
+
+        v = self.px(x, y)
+        for iy in range(y, -1, -1):
+            pixel = self.px(x, iy)
+            if pixel != v:
+                return iy
+        return 0
+
+    def nextcolor_right(self, x, y, default=None):
+        """
+        Determine the next pixel change going right from the (x,y) point.
+        If no next pixel is found default is returned.
+        """
+        if x < -1:
+            return -1
+        if x == -1:
+            return 0
+        if x == self.width - 1:
+            return self.width
+        if self.width <= x:
+            return default
+
+        v = self.px(x, y)
+        for ix in range(x, self.width):
+            pixel = self.px(ix, y)
+            if pixel != v:
+                return ix
+        return self.width - 1
+
+    def nextcolor_bottom(self, x, y, default=None):
+        """
+        Determine the next pixel change going bottom from the (x,y) point.
+        If no next pixel is found default is returned.
+        """
+        if y < -1:
+            return -1
+        if y == -1:
+            return 0
+        if y == self.height - 1:
+            return self.height
+        if self.height <= y:
+            return default
+
+        v = self.px(x, y)
+        for iy in range(y, self.height):
+            pixel = self.px(x, iy)
+            if pixel != v:
+                return iy
+        return self.height - 1
+
+    def calculate_next_horizontal_pixel(self, y, dy=1, leftmost_pixel=True):
+        """
+        Find the horizontal extreme at the given y-scanline, stepping by dy in the target image.
+        This can be done on either the leftmost (True) or rightmost (False).
+
+        @param y: y-scanline
+        @param dy: dy-step amount (usually should be -1 or 1)
+        @param leftmost_pixel: find pixel from the left
+        @return:
+        """
+        try:
+            if leftmost_pixel:
+                while True:
+                    x = self.leftmost_not_equal(y)
+                    if x is not None:
+                        break
+                    y += dy
+            else:
+                while True:
+                    x = self.rightmost_not_equal(y)
+                    if x is not None:
+                        break
+                    y += dy
+        except IndexError:
+            # Remaining image is blank
+            return None, None
+        return x, y
+
+    def calculate_next_vertical_pixel(self, x, dx=1, topmost_pixel=True):
+        """
+        Find the vertical extreme at the given x-scanline, stepping by dx in the target image.
+        This can be done on either the topmost (True) or bottommost (False).
+
+        @param x: x-scanline
+        @param dx: dx-step amount (usually should be -1 or 1)
+        @param topmost_pixel: find the pixel from the top
+        @return:
+        """
+        try:
+            if topmost_pixel:
+                while True:
+                    y = self.topmost_not_equal(x)
+                    if y is not None:
+                        break
+                    x += dx
+            else:
+                while True:
+                    # find that the bottommost pixel in that row.
+                    y = self.bottommost_not_equal(x)
+
+                    if y is not None:
+                        # This is a valid pixel.
+                        break
+                    # No pixel in that row was valid. Move to the next row.
+                    x += dx
+        except IndexError:
+            # Remaining image was blank, there are no more relevant pixels.
+            return None, None
+        return x, y
+
+    def calculate_first_pixel(self):
+        """
+        Find the first non-skipped pixel in the rastering.
+
+        This takes into account the traversal values of X_AXIS or Y_AXIS and BOTTOM and RIGHT
+        The start edge and the start point.
+
+        @return: x,y coordinates of first pixel.
+        """
+        if self.horizontal:
+            y = 0 if self.start_minimum_y else self.height - 1
+            dy = 1 if self.start_minimum_y else -1
+            x, y = self.calculate_next_horizontal_pixel(y, dy, self.start_minimum_x)
+            return x, y
+        else:
+            x = 0 if self.start_minimum_x else self.width - 1
+            dx = 1 if self.start_minimum_x else -1
+            x, y = self.calculate_next_vertical_pixel(x, dx, self.start_minimum_y)
+            return x, y
+
+    def calculate_last_pixel(self):
+        """
+        Find the last non-skipped pixel in the rastering.
+
+        First and last scanlines start from the same side when scanline count is odd
+
+        @return: x,y coordinates of last pixel.
+        """
+        if self.horizontal:
+            y = self.height - 1 if self.start_minimum_y else 0
+            dy = -1 if self.start_minimum_y else 1
+            start_on_left = (
+                self.start_minimum_x if self.width & 1 else not self.start_minimum_x
+            )
+            x, y = self.calculate_next_horizontal_pixel(y, dy, start_on_left)
+            return x, y
+        else:
+            x = self.width - 1 if self.start_minimum_x else 0
+            dx = -1 if self.start_minimum_x else 1
+            start_on_top = (
+                self.start_minimum_y if self.height & 1 else not self.start_minimum_y
+            )
+            x, y = self.calculate_next_vertical_pixel(x, dx, start_on_top)
+            return x, y
+
+    def initial_position(self):
+        """
+        Returns raw initial position for the relevant pixel within the data.
+        @return: initial position within the data.
+        """
+        if self.use_integers:
+            return int(round(self.initial_x)), int(round(self.initial_y))
+        else:
+            return self.initial_x, self.initial_y
+
+    def initial_position_in_scene(self):
+        """
+        Returns the initial position for this within the scene. Taking into account start corner, and step size.
+        @return: initial position within scene. The first plot location.
+        """
+        if self.initial_x is None:  # image is blank.
+            if self.use_integers:
+                return int(round(self.offset_x)), int(round(self.offset_y))
+            else:
+                return self.offset_x, self.offset_y
+        if self.use_integers:
+            return (
+                int(round(self.offset_x + self.initial_x * self.step_x)),
+                int(round(self.offset_y + self.initial_y * self.step_y)),
+            )
+        else:
+            return (
+                self.offset_x + self.initial_x * self.step_x,
+                self.offset_y + self.initial_y * self.step_y,
+            )
+
+    def final_position_in_scene(self):
+        """
+        Returns best guess of final position relative to the scene offset. Taking into account start corner, and parity
+        of the width and height.
+        @return:
+        """
+        if self.final_x is None:  # image is blank.
+            if self.use_integers:
+                return int(round(self.offset_x)), int(round(self.offset_y))
+            else:
+                return self.offset_x, self.offset_y
+        if self.use_integers:
+            return (
+                int(round(self.offset_x + self.final_x * self.step_x)),
+                int(round(self.offset_y + self.final_y * self.step_y)),
+            )
+        else:
+            return (
+                self.offset_x + self.final_x * self.step_x,
+                self.offset_y + self.final_y * self.step_y,
+            )
+
+    def plot(self):
+        """
+        Plot the values yielded by following the given raster plotter in the traversal defined.
+        """
+        while self._locked:
+            sleep(0.1)
+        self._locked = True
+        offset_x = self.offset_x
+        offset_y = self.offset_y
+        step_x = self.step_x
+        step_y = self.step_y
+        self._distance_travel = 0
+        self._distance_burn = 0
+
+        if self.initial_x is None:
+            # There is no image.
+            return
+        # Debug code....
+        methods = (
+            "Top2Bottom",
+            "Bottom2Top",
+            "Right2Left",
+            "Left2Right",
+            "Hatch",
+            "Greedy Neighbor Hor",
+            "Greedy Neighbor Ver",
+            "Crossover",
+            "Spiral",
+        )
+        testmethods = (
+            "Test: Horizontal Rectangle",
+            "Test: Vertical Rectangle",
+            "Test: Horizontal Snake",
+            "Test: Vertical Snake",
+            "Test: Spiral",
+        )
+        if self._cache is None:
+            if self.debug_level > 0:
+                try:
+                    if self.direction >= 0:
+                        m = methods[self.direction]
+                    else:
+                        m = testmethods[abs(self.direction) - 1]
+                    s_meth = f"Method: {m} ({self.direction})"
+                except IndexError:
+                    s_meth = f"Method: Unknown {self.direction}"
+                print (s_meth)
+                data = list(self._plot_pixels())
+                from platform import system
+                defaultdir = "c:\\temp\\" if system() == "Windows" else ""
+                has_duplicates = 0
+                tstamp = int(perf_counter() * 100)
+                with open(f"{defaultdir}plot_{tstamp}.txt", mode="w") as f:
+                    f.write(f"0.9.7\n{s_meth}\n{'Bidirectional' if self.bidirectional else 'Unidirectional'} {'horizontal' if self.horizontal else 'vertical'} plot starting at {'top' if self.start_minimum_y else 'bottom'}-{'left' if self.start_minimum_x else 'right'}\n")
+                    f.write(f"Overscan: {self.overscan:.2f}, Stepx={step_x:.2f}, Stepy={step_y:.2f}\n")
+                    f.write(f"Image dimensions: {self.width}x{self.height}\n")
+                    f.write(f"Startpoint: {self.initial_x}, {self.initial_y}\n")
+                    f.write(f"Overlapping pixels to any side: {self.overlap}\n")
+                    if self.special:
+                        f.write(f"Special instructions:\n")
+                        for key, value in self.special.items():
+                            f.write(f"  {key} = {value}\n")
+                    f.write("----------------------------------------------------------------------\n")
+                    test_dict = {}
+                    lastx = self.initial_x
+                    lasty = self.initial_y
+                    failed = False
+                    for lineno, (x, y, on) in enumerate(data, start=1):
+                        if lastx is not None:
+                            dx = x - lastx
+                            dy = y - lasty
+                            if dx != 0 and dy != 0: # and abs(dx) != abs(dy):
+                                f.write (f"You f**ed up! No zigzag movement from line {lineno - 1} to {lineno}: {lastx}, {lasty} -> {x}, {y}\n")
+                                print (f"You f**ed up! No zigzag movement from line {lineno - 1} to {lineno}: {lastx}, {lasty} -> {x}, {y}")
+                                failed = True
+                        lastx = x
+                        lasty = y
+                    if not failed:
+                        f.write("Good news, no zig-zag movements identified!\n")
+                    f.write("----------------------------------------------------------------------\n")
+                    for lineno, (x, y, on) in enumerate(data, start=1):
+                        if x is None or y is None:
+                            continue
+                        key = f"{x} - {y}"
+                        if key in test_dict:
+                            f.write (f"Duplicate coordinates in list at ({x}, {y})! 1st: #{test_dict[key][0]}, on={test_dict[key][1]}, 2nd: #{lineno}, on={on}\n")
+                            has_duplicates += 1
+                        else:
+                            test_dict[key] = (lineno, on)
+                    if has_duplicates:
+                        f.write("----------------------------------------------------------------------\n")
+                    for lineno, (x, y, on) in enumerate(data, start=1):
+                        f.write(f"{lineno}: {x}, {y}, {on}\n")
+                    if has_duplicates:
+                        print(f"Attention: the generated plot has {has_duplicates} duplicate coordinate values!")
+                        print(f"{'Bidirectional' if self.bidirectional else 'Unidirectional'} {'horizontal' if self.horizontal else 'vertical'} plot starting at {'top' if self.start_minimum_y else 'bottom'}-{'left' if self.start_minimum_x else 'right'}")
+                        print(f"Overscan: {self.overscan:.2f}, Stepx={step_x:.2f}, Stepy={step_y:.2f}")
+                        print(f"Image dimensions: {self.width}x{self.height}")
+            else:
+                data = list(self._plot_pixels())
+            self._cache = data
+        else:
+            data = self._cache
+        last_x = offset_x
+        last_y = offset_y
+        if self.use_integers:
+            for x, y, on in data:
+                if x is None or y is None:
+                    # Passthrough
+                    yield x, y, on
+                else:
+                    nx = int(round(offset_x + step_x * x))
+                    ny = int(round(offset_y + y * step_y))
+                    self._distance_burn += 0 if on == 0 else sqrt( (nx - last_x) * (nx-last_x) + (ny - last_y) * (ny - last_y) )
+                    self._distance_travel += 0 if on != 0 else sqrt( (nx - last_x) * (nx-last_x) + (ny - last_y) * (ny - last_y) )
+                    yield nx, ny, on
+                    last_x = nx
+                    last_y = ny
+        else:
+            for x, y, on in data:
+                if x is None or y is None:
+                    # Passthrough
+                    yield x, y, on
+                else:
+                    nx = round(offset_x + step_x * x)
+                    ny = round(offset_y + y * step_y)
+                    self._distance_burn += 0 if on == 0 else sqrt( (nx - last_x) * (nx-last_x) + (ny - last_y) * (ny - last_y) )
+                    self._distance_travel += 0 if on != 0 else sqrt( (nx - last_x) * (nx-last_x) + (ny - last_y) * (ny - last_y) )
+                    yield offset_x + step_x * x, offset_y + y * step_y, on
+                    last_x = nx
+                    last_y = ny
+        self._locked = False
+
+    def _plot_pixels(self):
+        legacy = self.special.get("legacy", False)
+        if self.direction in (RASTER_GREEDY_H, RASTER_GREEDY_V):
+            yield from self._plot_greedy_neighbour(horizontal=self.horizontal)
+        elif self.direction == RASTER_CROSSOVER:
+            yield from self._plot_crossover()
+        elif self.direction == RASTER_SPIRAL:
+            yield from self._plot_spiral()
+        # elif self.direction < 0:
+        #     yield from self.testpattern_generator()
+        elif self.horizontal:
+            if legacy:
+                yield from self._legacy_plot_horizontal()
+            else:
+                yield from self._plot_horizontal()
+        else:
+            if legacy:
+                yield from self._legacy_plot_vertical()
+            else:
+                yield from self._plot_vertical()
+            # yield from self._plot_vertical()
+
+    def _debug_data(self, force=False):
+        if self.debug_level < 3 and not force:
+            return
+        BLANK = 255
+        for y in range(self.height):
+            msg:str = f"{y:3d}: "
+            for x in range(self.width):
+                msg += "." if self.data[x, y] == BLANK else "X"
+            print (msg)
+
+    def _get_pixel_chains(self, xy:int, is_x : bool) -> list:
+        last_pixel = None
+        segments = []
+        upper = self.width if is_x else self.height
+        for idx in range(upper):
+            pixel = self.px(idx, xy) if is_x else self.px(xy, idx)
+            on = 0 if pixel == self.skip_pixel else pixel
+            if on:
+                if on == last_pixel:
+                    segments[-1][1] = idx
+                else:
+                    segments.append ([idx, idx, on])
+            last_pixel = on
+        return segments
+
+    def _consume_pixel_chains(self, segments:list, xy:int, is_x : bool):
+        BLANK = 255
+        # for x in range(5):
+        #     msg1 = f"{x}: "
+        #     msg2 = ""
+        #     for y in range(5):
+        #         msg1 += "." if self.data[x, y] == BLANK else "X"
+        #         msg2 += f" {self.data[x, y]}"
+        #     print (msg1, msg2)
+
+        for seg in segments:
+            c_start = seg[0]
+            c_end = seg[1]
+            for idx in range(c_start, c_end + 1):
+                if is_x:
+                    px = idx
+                    py = xy
+                else:
+                    px = xy
+                    py = idx
+                for x_idx in range(-self.overlap, self.overlap + 1):
+                    for y_idx in range(-self.overlap, self.overlap + 1):
+                        nx = px + x_idx
+                        ny = py + y_idx
+                        if nx < 0 or nx >= self.width or ny < 0 or ny >= self.height:
+                            continue
+                        self.data[nx, ny] = BLANK
+        self._debug_data()
+
+    def _plot_vertical(self):
+        """
+        This code is for vertical rastering.
+        We are looking first for all consecutive pixel chains with the same pixel value
+        Then we loop through the segments and yield the 'end edge' of the 'last' pixel
+        'end edge' and 'last' are dependent on the sweep direction.
+        There is one peculiarity though that is required for K40 lasers:
+        a) We may only move from one yielded (x,y,on) tuple in a pure horizontal or pure
+           vertical fashion (we could as well go perfectly diagonal but we are not using
+           this feature). So at the end of one sweepline we need to change to the
+           next scanline by going directly up/down/left/right and then move to the first
+           relevant pixel.
+        b) we need to take care that we are not landing on the same pixel twice. So we move
+           outside the new chain to avoid this
+        """
+        unidirectional = not self.bidirectional
+
+        dx = 1 if self.start_minimum_x else -1
+        dy = 1 if self.start_minimum_y else -1
+        lower = min(self.initial_x, self.final_x)
+        upper = max(self.initial_x, self.final_x)
+        last_x = self.initial_x
+        last_y = self.initial_y
+        x = lower if self.start_minimum_x else upper
+        first = True
+        while lower <= x <= upper:
+            segments = self._get_pixel_chains(x, False)
+            self._consume_pixel_chains(segments, x, False)
+            if segments:
+                if dy > 0:
+                    # from top to bottom
+                    idx = 0
+                    start = 0
+                    end = 1
+                    edge_start = -0.5
+                    edge_end = 0.5
+                else:
+                    idx = len(segments) - 1
+                    end = 0
+                    start = 1
+                    edge_start = 0.5
+                    edge_end = -0.5
+                # Goto next column, but make sure we end up outside our chain
+                # We consider as well the overscan value
+                overscan_top = 0 if dy >= 0 else self.overscan
+                overscan_bottom = 0 if dy <= 0 else self.overscan
+                if not first and (segments[0][0] - overscan_top <= last_y <= segments[-1][1] + overscan_bottom):
+                    # inside the chain!
+                    # So lets move a bit to the side
+                    if dy > 0:
+                        if self.bidirectional:
+                            # Previous was sweep from right to left, so we go beyond first point
+                            last_y = segments[0][0] - overscan_top - 1
+                        else:
+                            # We go beyond last point
+                            last_y = segments[-1][1] + overscan_bottom + 1
+                    else:
+                        # Previous was sweep from left to right, so we go beyond last point
+                        last_y = segments[-1][1] + overscan_bottom + 1
+                    last_x = x - dx
+                    yield last_x, last_y, 0
+                last_x = x
+                yield last_x, last_y, 0
+                while 0 <= idx < len(segments):
+                    sy = segments[idx][start] + edge_start
+                    ey = segments[idx][end] + edge_end
+                    on = segments[idx][2]
+                    if last_y != sy:
+                        yield last_x, sy, 0
+                    last_y = ey
+                    yield last_x, last_y, on
+                    idx += dy
+                if self.overscan:
+                    last_y += dy * self.overscan
+                    yield last_x, last_y, 0
+                if not unidirectional:
+                    dy = -dy
+                first = False
+            else:
+                # Just climb the line, and don't change directions
+                last_x = x
+                yield last_x, last_y, 0
+
+            x += dx
+
+    def _plot_horizontal(self):
+        """
+        This code is horizontal rastering.
+        We are looking first for all consecutive pixel chains with the same pixel value
+        Then we loop through the segments and yield the 'end edge' of the 'last' pixel
+        'end edge' and 'last' are dependent on the sweep direction.
+        There is one peculiarity though that is required for K40 lasers:
+        a) We may only move from one yielded (x,y,on) tuple in a pure horizontal or pure
+           vertical fashion (we could as well go perfectly diagonal but we are not using
+           this feature). So at the end of one sweepline we need to change to the
+           next scanline by going directly up/down/left/right and then move to the first
+           relevant pixel.
+        b) we need to take care that we are not landing on the same pixel twice. So we move
+           outside the new chain to avoid this
+        """
+        unidirectional = not self.bidirectional
+
+        dx = 1 if self.start_minimum_x else -1
+        dy = 1 if self.start_minimum_y else -1
+        last_x = self.initial_x
+        last_y = self.initial_y
+        lower = min(self.initial_y, self.final_y)
+        upper = max(self.initial_y, self.final_y)
+        y = lower if self.start_minimum_y else upper
+        first = True
+        while lower <= y <= upper:
+            segments = self._get_pixel_chains(y, True)
+            self._consume_pixel_chains(segments, y, True)
+            if segments:
+                if dx > 0:
+                    # from left to right
+                    idx = 0
+                    start = 0
+                    end = 1
+                    edge_start = -0.5
+                    edge_end = 0.5
+                else:
+                    idx = len(segments) - 1
+                    end = 0
+                    start = 1
+                    edge_start = 0.5
+                    edge_end = -0.5
+                if last_x is None:
+                    last_x = segments[idx][start] + edge_start
+                # Goto next line, but make sure we end up outside our chain
+                # We consider as well the overscan value
+                overscan_left = 0 if dx >= 0 else self.overscan
+                overscan_right = 0 if dx <= 0 else self.overscan
+                if not first and (segments[0][0] - overscan_left <= last_x <= segments[-1][1] + overscan_right):
+                    # inside the chain!
+                    # So lets move a bit to the side
+                    if dx > 0:
+                        if self.bidirectional:
+                            # Previous was sweep from right to left, so we go beyond first point
+                            last_x = segments[0][0] - overscan_left - 1
+                        else:
+                            # We go beyond last point
+                            last_x = segments[-1][1] + overscan_right + 1
+                    else:
+                        # Previous was sweep from left to right, so we go beyond last point
+                        last_x = segments[-1][1] + overscan_right + 1
+                    yield last_x, y - dy, 0
+                last_y = y
+                yield last_x, last_y, 0
+                while 0 <= idx < len(segments):
+                    sx = segments[idx][start] + edge_start
+                    ex = segments[idx][end] + edge_end
+                    on = segments[idx][2]
+                    if last_x != sx:
+                        yield sx, last_y, 0
+                    last_x = ex
+                    yield last_x, last_y, on
+                    idx += dx
+                if self.overscan:
+                    last_x += dx * self.overscan
+                    yield last_x, last_y, 0
+                if not unidirectional:
+                    dx = -dx
+                first = False
+            else:
+                # Just climb the line, and don't change directions
+                last_y = y
+                yield last_x, last_y, 0
+            y += dy
+
+    # Legacy code for the m2nano - yes this has deficits for low dpi but it seems
+    # to be finetuned to the needs of the m2nano controller.
+    # This will be called if the appropriate device setting is in place
+    def _legacy_plot_vertical(self):
+        """
+        This code is for vertical rastering.
+
+        @return:
+        """
+        width = self.width
+        unidirectional = not self.bidirectional
+        skip_pixel = self.skip_pixel
+
+        x, y = self.initial_position()
+        dx = 1 if self.start_minimum_x else -1
+        dy = 1 if self.start_minimum_y else -1
+
+        yield x, y, 0
+        while 0 <= x < width:
+            lower_bound = self.topmost_not_equal(x)
+            if lower_bound is None:
+                x += dx
+                yield x, y, 0
+                continue
+            upper_bound = self.bottommost_not_equal(x)
+            traveling_bottom = self.start_minimum_y if unidirectional else dy >= 0
+            next_traveling_bottom = self.start_minimum_y if unidirectional else dy <= 0
+
+            next_x, next_y = self.calculate_next_vertical_pixel(
+                x + dx, dx, topmost_pixel=next_traveling_bottom
+            )
+            if next_y is not None:
+                # If we have a next scanline, we must end after the last pixel of that scanline too.
+                upper_bound = max(next_y, upper_bound) + self.overscan
+                lower_bound = min(next_y, lower_bound) - self.overscan
+            pixel_chain = []
+            last_y = lower_bound if traveling_bottom else upper_bound
+            if traveling_bottom:
+                while y < upper_bound:
+                    try:
+                        pixel = self.px(x, y)
+                    except IndexError:
+                        pixel = 0
+                    y = self.nextcolor_bottom(x, y, upper_bound)
+                    y = min(y, upper_bound)
+                    if pixel == skip_pixel:
+                        yield x, y, 0
+                    else:
+                        yield x, y, pixel
+                        pixel_chain.append([last_y, y, pixel])
+                    last_y = y + 1
+            else:
+                while lower_bound < y:
+                    try:
+                        pixel = self.px(x, y)
+                    except IndexError:
+                        pixel = 0
+                    y = self.nextcolor_top(x, y, lower_bound)
+                    y = max(y, lower_bound)
+                    if pixel == skip_pixel:
+                        yield x, y, 0
+                    else:
+                        yield x, y, pixel
+                        pixel_chain.append([y, last_y, pixel])
+                    last_y = y - 1
+            if pixel_chain:
+                self._consume_pixel_chains(pixel_chain, x, False)
+            if next_y is None:
+                # remaining image is blank, we stop right here.
+                return
+            yield next_x, y, 0
+            if y != next_y:
+                yield next_x, next_y, 0
+            x = next_x
+            y = next_y
+            dy = -dy
+
+    def _legacy_plot_horizontal(self):
+        """
+        This code is horizontal rastering.
+
+        @return:
+        """
+        height = self.height
+        unidirectional = not self.bidirectional
+        skip_pixel = self.skip_pixel
+
+        x, y = self.initial_position()
+        dx = 1 if self.start_minimum_x else -1
+        dy = 1 if self.start_minimum_y else -1
+        yield x, y, 0
+        while 0 <= y < height:
+            lower_bound = self.leftmost_not_equal(y)
+            if lower_bound is None:
+                y += dy
+                yield x, y, 0
+                continue
+            upper_bound = self.rightmost_not_equal(y)
+            traveling_right = self.start_minimum_x if unidirectional else dx >= 0
+            next_traveling_right = self.start_minimum_x if unidirectional else dx <= 0
+
+            next_x, next_y = self.calculate_next_horizontal_pixel(
+                y + dy, dy, leftmost_pixel=next_traveling_right
+            )
+            if next_x is not None:
+                # If we have a next scanline, we must end after the last pixel of that scanline too.
+                upper_bound = max(next_x, upper_bound) + self.overscan
+                lower_bound = min(next_x, lower_bound) - self.overscan
+            pixel_chain = []
+            last_x = lower_bound if traveling_right else upper_bound
+            if traveling_right:
+                while x < upper_bound:
+                    try:
+                        pixel = self.px(x, y)
+                    except IndexError:
+                        pixel = 0
+                    x = self.nextcolor_right(x, y, upper_bound)
+                    x = min(x, upper_bound)
+                    if pixel == skip_pixel:
+                        yield x, y, 0
+                    else:
+                        yield x, y, pixel
+                        pixel_chain.append([last_x, x, pixel])
+                    last_x = x + 1
+            else:
+                while lower_bound < x:
+                    try:
+                        pixel = self.px(x, y)
+                    except IndexError:
+                        pixel = 0
+                    x = self.nextcolor_left(x, y, lower_bound)
+                    x = max(x, lower_bound)
+                    if pixel == skip_pixel:
+                        yield x, y, 0
+                    else:
+                        yield x, y, pixel
+                        pixel_chain.append([x, last_x, pixel])
+                    last_x = x - 1
+            if pixel_chain:
+                self._consume_pixel_chains(pixel_chain, y, True)
+            if next_y is None:
+                # remaining image is blank, we stop right here.
+                return
+            yield x, next_y, 0
+            if x != next_x:
+                yield next_x, next_y, 0
+            x = next_x
+            y = next_y
+            dx = -dx
+
+    def _plot_greedy_neighbour(self, horizontal: bool = True):
+        """
+        Distance Matrix Function: The distance_matrix function calculates the squared distances from the
+        current point to all other points, applying a penalty to the y-distances to prefer movements in the x-direction.
+
+        Initialization: We initialize the visited array to keep track of visited segments, the path list
+        to store the order of segments along with the relevant point (start or end),
+        and the current_point as the starting point.
+
+        Sliding Window: We use a sliding window to limit the number of segments we need to consider at each step.
+        The window size is controlled by the window_size parameter and is applied to both x and y coordinates.
+
+        Main Loop: In the main loop, we calculate the distances from the current point to all unvisited points
+        within the sliding window using the distance_matrix function.
+        We then use np.argmin to find the index of the smallest distance.
+        We update the current_point to the next point and mark the segment as visited.
+        The path list is updated with the segment index and whether the start or end point is relevant.
+        """
+
+        def walk_segments(segments, horizontal=True, xy_penalty=1, width=1, height=1):
+            n:int = len(segments)
+            visited = np.zeros(n, dtype=bool)
+            path = []
+            window_size = 10
+            current_point = np.array(segments[0][0], dtype=float)
+            segment_points = np.array([point for segment in segments for point in segment], dtype=float)
+            mask = ~visited.repeat(2)
+            while len(path) < n:
+                # Find the range of segments within the x- and y-window
+                x_min = current_point[0] - window_size
+                x_max = current_point[0] + window_size
+                y_min = current_point[1] - window_size
+                y_max = current_point[1] + window_size
+                unvisited_indices = np.where(
+                    (segment_points[:, 0] >= x_min) &
+                    (segment_points[:, 0] <= x_max) &
+                    (segment_points[:, 1] >= y_min) &
+                    (segment_points[:, 1] <= y_max) &
+                    mask
+                )[0]
+                if len(unvisited_indices) == 0:
+                    # If no segments are within the window, expand the window
+                    window_size *= 2
+                    # print (f"Did not find points: now window: {window_size}")
+                    if window_size <= 2* height or window_size <= 2 * width: # Safety belt
+                        continue
+
+                unvisited_points = segment_points[unvisited_indices]
+
+                # distances = distance_matrix(unvisited_points, current_point, y_penalty)
+                diff = unvisited_points - current_point
+                if horizontal:
+                    diff[:, 1] *= xy_penalty # Apply penalty to y-distances
+                else:
+                    diff[:, 0] *= xy_penalty # Apply penalty to x-distances
+
+                distances = np.sum(diff ** 2, axis=1) # Return squared distances
+
+                min_distance_idx = np.argmin(distances)
+                next_segment = unvisited_indices[min_distance_idx] // 2
+
+                if not visited[next_segment]:
+                    visited[next_segment] = True
+                    # mask = ~visited.repeat(2)
+                    mask[2 * next_segment] = False
+                    mask[2 * next_segment + 1] = False
+                    if min_distance_idx % 2 == 0:
+                        path.append((next_segment, 'end'))
+                        current_point = segment_points[next_segment * 2 + 1]  # Move to the other endpoint
+                    else:
+                        path.append((next_segment, 'start'))
+                        current_point = segment_points[next_segment * 2]  # Move to the other endpoint
+                    window_size = 10 # Reset window size
+
+            return path
+
+        t0 = perf_counter()
+        # An experimental routine
+        if horizontal:
+            dy = 1 if self.start_minimum_y else -1
+            lower = min(self.initial_y, self.final_y)
+            upper = max(self.initial_y, self.final_y)
+            y = lower if self.start_minimum_y else upper
+        else:
+            dx = 1 if self.start_minimum_x else -1
+            lower = min(self.initial_x, self.final_x)
+            upper = max(self.initial_x, self.final_x)
+            x = lower if self.start_minimum_x else upper
+
+        line_parts = []
+        on_parts = []
+        if self.debug_level > 2:
+            print (f"{'horizontal' if horizontal else 'Vertical'} for {self.width}x{self.height} image. {'y' if horizontal else 'x'} from {lower} to {upper}")
+        if horizontal:
+            while lower <= y <= upper:
+                segments = self._get_pixel_chains(y, True)
+                self._consume_pixel_chains(segments, y, True)
+                for seg in segments:
+                    # Append (xstart, y), (xend, y), on
+                    line_parts.append( ( (seg[0], y), (seg[1], y) ) )
+                    on_parts.append(seg[2])
+                y += dy
+        else:
+            while lower <= x <= upper:
+                segments = self._get_pixel_chains(x, False)
+                self._consume_pixel_chains(segments, x, False)
+                for seg in segments:
+                    # Append (xstart, y), (xend, y), on
+                    line_parts.append( ( (x, seg[0]), (x, seg[1]) ) )
+                    on_parts.append(seg[2])
+                x += dx
+        if self.debug_level > 2:
+            print (f"Created {len(line_parts)} segments")
+        t1 = perf_counter()
+        penalty = 3 if self.special.get("gantry", False) else 1
+        path = walk_segments(line_parts, horizontal=horizontal, xy_penalty=penalty, width=self.width, height=self.height)
+        # print("Order of segments:", path)
+        t2 = perf_counter()
+        if horizontal:
+            last_x = self.initial_x
+            last_y = lower
+        else:
+            last_x = lower
+            last_y = self.initial_y
+        for idx, mode in path:
+            if mode == "end":
+                # end was closer
+                (ex, ey), (sx, sy) = line_parts[idx]
+            else:
+                (sx, sy), (ex, ey) = line_parts[idx]
+            on = on_parts[idx]
+            if horizontal:
+                dx = ex - sx
+                if dx >= 0:
+                    # from left to right
+                    edge_start = -0.5
+                    edge_end = 0.5
+                else:
+                    edge_start = 0.5
+                    edge_end = -0.5
+                sx += edge_start
+                ex += edge_end
+                if sy != last_y:
+                    last_y = sy
+                    yield last_x, last_y, 0
+                if last_x != sx:
+                    yield sx, last_y, 0
+            else:
+                dy = ey - sy
+                if dy >= 0:
+                    # from left to right
+                    edge_start = -0.5
+                    edge_end = 0.5
+                else:
+                    edge_start = 0.5
+                    edge_end = -0.5
+                sy += edge_start
+                ey += edge_end
+                if sx != last_x:
+                    last_x = sx
+                    yield last_x, last_y, 0
+                if last_y != sy:
+                    last_y = sy
+                    yield sx, sy, 0
+
+            yield ex, ey, on
+            last_x = ex
+            last_y = ey
+        t3 = perf_counter()
+        if self.debug_level > 1:
+            print (f"Overall time for {'horizontal' if horizontal else 'vertical'} consumption: {t3-t0:.2f}s - created: {len(line_parts)} segments")
+            print (f"Computation: {t2-t0:.2f}s - Chain creation:{t1 - t0:.2f}s, Walk: {t2 - t1:.2f}s")
+        self.final_x = last_x
+        self.final_y = last_y
+
+    def _plot_spiral(self):
+
+
+        rows = self.height
+        cols = self.width
+        center_row, center_col = rows // 2, cols // 2
+        self.initial_x = center_col
+        self.initial_y = center_row
+
+        directions = [(1, 0), (0, 1), (-1, 0), (0, -1)]  # Right, Down, Left, Up
+        edges = [(0.5, 0), (0, 0.5), (-0.5, 0), (0, -0.5)]
+        direction_index = 0
+        steps = 1
+
+        row, col = center_row, center_col
+        # is the very first pixel an on?
+        last_x = col
+        last_y = row
+        count = 1
+        pixel = self.px(col, row)
+        if pixel == self.skip_pixel:
+            pixel = 0
+        last_pixel = pixel
+        if pixel:
+            yield col - 0.5, row, 0
+            last_x = col + 0.5
+            yield last_x, row, pixel
+        while count < rows * cols:
+            for _ in range(2):
+                segments = []
+                dx, dy = edges[direction_index]
+                edge_start_x = -dx
+                edge_start_y = -dy
+                edge_end_x = dx
+                edge_end_y = dy
+                # msg = f"[({col}, {row}) - {steps}] "
+                for _ in range(steps):
+                    row += directions[direction_index][1]
+                    col += directions[direction_index][0]
+                    if 0 <= row < rows and 0 <= col < cols:
+                        pixel = self.px(col, row)
+                        on = 0 if pixel == self.skip_pixel else pixel
+                        # msg = f"{msg} {'X' if on else '.'}"
+                        if on:
+                            if on == last_pixel and len(segments):
+                                segments[-1][1] = (col, row)
+                            else:
+                                segments.append ([(col, row), (col, row), on])
+
+                        last_pixel = on
+                        count += 1
+                        if count == rows * cols:
+                            break
+                # Deal with segments
+                # print (msg)
+                # print (segments)
+                for (sx, sy), (ex, ey), pixel in segments:
+                    sx += edge_start_x
+                    sy += edge_start_y
+                    ex += edge_end_x
+                    ey += edge_end_y
+                    if last_y != sy:
+                        yield last_x, sy, 0
+                        last_y = sy
+                    if last_x != sx:
+                        yield sx, last_y, 0
+                    last_x = ex
+                    last_y = ey
+                    yield last_x, last_y, pixel
+                    self.final_x, self.final_y = last_x, last_y
+                # Now we need to empty overlapping pixels...
+                if self.overlap > 0:
+                    BLANK = 255
+                    for (start_x, start_y), (end_x, end_y), on in segments:
+                        sx = min(start_x, end_x)
+                        ex = max(start_x, end_x)
+                        sy = min(start_y, end_y)
+                        ey = max(start_y, end_y)
+
+                        if direction_index in (0, 2): # horizontal
+                            for y_idx in range(-self.overlap, self.overlap + 1):
+                                ny = sy + y_idx
+                                for nx in range(sx, ex + 1):
+                                    if 0 <= nx < self.width and 0 <= ny < self.height:
+                                        self.data[nx, ny] = BLANK
+                        else:
+                            for x_idx in range(-self.overlap, self.overlap + 1):
+                                nx = sx + x_idx
+                                for ny in range(sy, ey + 1):
+                                    if 0 <= nx < self.width and 0 <= ny < self.height:
+                                        self.data[nx, ny] = BLANK
+
+
+                direction_index = (direction_index + 1) % 4
+            steps += 1
+
+
+    def _plot_crossover(self):
+        """
+        This algorithm scans through the image looking for the row or the column with the most pixels.
+        It will hand back this information together with a precompiled list of
+        state changes (start, end, pixel), ie a non-blank pixel with a different on value
+        than the previous pixel
+        It will then clean the row / column and start again.
+        This happens until no more non-empty rows / cols can be found
+
+        The receiving routine takes these values, orders them according to
+        type (row/col) and row/col number and generates plot instructions
+        that will be yielded.
+
+        Yields:
+            list of tuples with (x, y, on)
+        """
+        ROW=0
+        COL=1
+
+        def process_image(image):
+            # We will modify the image to keep track of deleted rows and columns
+            # Get the dimensions of the image
+            rows, cols = image.shape
+            # We prefer cols (which is the x-axis and that is normally
+            # slightly advantageous for gantry lasers)
+            if self.special.get("gantry", False):
+                colfactor = 1.0
+                rowfactor = 0.8
+            else:
+                colfactor = 1.0
+                rowfactor = 1.0
+
+            # Initialize a list to store the results
+            results = []
+
+            # Iterate through the matrix, we cover all rows and cols
+            colidx = 0
+            rowidx = 0
+            covered_row = [None] * rows
+            covered_col = [None] * cols
+            stored_row = np.zeros(rows)
+            stored_col = np.zeros(cols)
+            stored_row_len = np.zeros(rows)
+            stored_col_len = np.zeros(cols)
+            recalc_row = True
+            recalc_col = True
+            while True:
+                if recalc_col:
+                    for i in range(cols):
+                        col_len = cols
+                        if covered_col[i] is None:
+                            nonzero_indices = np.nonzero(image[:, i])[0]
+                            count = len(nonzero_indices)
+                            if count == 0:
+                                covered_col[i] = True
+                            else:
+                                col_len = nonzero_indices[-1] - nonzero_indices[0] + 1
+                                covered_col[i] = False
+                            stored_col[i] = count
+                            stored_col_len[i] = col_len
+                if recalc_row:
+                    for i in range(rows):
+                        row_len = rows
+                        if covered_row[i] is None:
+                            nonzero_indices = np.nonzero(image[i, :])[0]
+                            count = len(nonzero_indices)
+                            if count == 0:
+                                covered_row[i] = True
+                            else:
+                                row_len = nonzero_indices[-1] - nonzero_indices[0] + 1
+                                covered_row[i] = False
+                            stored_row[i] = count
+                            stored_row_len[i] = row_len
+
+                colidx = np.argmax(stored_col)
+                rowidx = np.argmax(stored_row)
+
+                col_count = stored_col[colidx]
+                col_len = stored_col_len[colidx]
+                row_count = stored_row[rowidx]
+                row_len = stored_row_len[rowidx]
+
+                if row_count == col_count == 0:
+                    break
+                # Determine whether there are more pixels in the row or column
+
+                row_ratio = row_count * row_count / row_len * rowfactor
+                col_ratio = col_count * col_count / col_len * colfactor
+                # print (f"Col #{rowidx}: {int(row_count):3d} pixel over {int(row_len):3d} length, ratio: {row_ratio:.3f} {'winner' if row_ratio >= col_ratio else 'loser'}")
+                # print (f"Row #{colidx}: {int(col_count):3d} pixel over {int(col_len):3d} length, ratio: {col_ratio:.3f} {'winner' if row_ratio < col_ratio else 'loser'}")
+                # if row_count >= col_count:
+                if row_ratio >= col_ratio:
+                    last_pixel = None
+                    segments = []
+                    # msg = ""
+                    for idx in range(cols):
+                        on = image[rowidx, idx]
+                        # msg = f"{msg}{'X' if on else '.'}"
+                        if on:
+                            if not covered_col[idx]:
+                                covered_col[idx] = None # needs recalc
+                            if on == last_pixel:
+                                segments[-1][1] = idx
+                            else:
+                                segments.append ([idx, idx, on])
+                        last_pixel = on
+                    results.append((COL, rowidx, segments)) # Intentionally so, as the numpy array has x and y exchanged
+                    # print (f"Col #{rowidx}: {msg} -> {segments}")
+
+                    # Clear the column
+                    image[rowidx,:] = 0
+                    covered_row[rowidx] = True
+                    stored_row[rowidx] = 0
+                    for rc in range(self.overlap):
+                        r = rowidx - rc
+                        if 0 <= r < rows:
+                            image[r,:] = 0
+                            covered_row[r] = True
+                            stored_row[r] = 0
+                        r = rowidx + rc
+                        if 0 <= r < rows:
+                            image[r,:] = 0
+                            covered_row[r] = True
+                            stored_row[r] = 0
+                    recalc_col = True
+                else:
+                    last_pixel = None
+                    segments = []
+                    # msg = ""
+                    for idx in range(rows):
+                        on = image[idx, colidx]
+                        # msg = f"{msg}{'X' if on else '.'}"
+                        if on:
+                            if not covered_row[idx]:
+                                covered_row[idx] = None # needs recalc
+                            if on == last_pixel:
+                                segments[-1][1] = idx
+                            else:
+                                segments.append ([idx, idx, on])
+                        last_pixel = on
+                    results.append((ROW, colidx, segments))
+                    # print (f"Row #{colidx}: {msg} -> {segments}")
+                    # Clear the row
+                    image[:, colidx] = 0
+                    covered_col[colidx] = True
+                    stored_col[colidx] = 0
+                    for rc in range(self.overlap):
+                        r = colidx - rc
+                        if 0 <= r < cols:
+                            image[:, r] = 0
+                            covered_col[r] = True
+                            stored_col[r] = 0
+                        r = rowidx + rc
+                        if 0 <= r < cols:
+                            image[:, r] = 0
+                            covered_col[r] = True
+                            stored_col[r] = 0
+                    recalc_row = True
+                if self.debug_level > 1:
+                    for cidx in range(cols):
+                        msg = ""
+                        for ridx in range(rows):
+                            on = image[ridx, cidx]
+                            msg = f"{msg}{'X' if on else '.'}"
+                        print (f"{cidx:3d}: {msg}")
+
+            return results
+
+        t0 = perf_counter()
+        # initialize the matrix
+        image = np.empty((self.width, self.height))
+        # Apply filter and eliminate skip_pixel
+        for x in range(self.width):
+            for y in range(self.height):
+                px = self.px(x, y)
+                if px==self.skip_pixel:
+                    px = 0
+                image[x, y] = px
+        t1 = perf_counter()
+        results = process_image(image)
+        results.sort()
+        t2 = perf_counter()
+        dx = +1
+        dy = +1
+        first_x = 0
+        first_y = 0
+        last_x = self.initial_x
+        last_y = self.initial_y
+        yield last_x, last_y, 0
+        for mode, xy, segments in results:
+            # eliminate data and swap direction
+            if self.special.get("mode_filter", "") == "ROW" and mode != ROW:
+                continue
+            if self.special.get("mode_filter", "") == "COL" and mode != COL:
+                continue
+
+            if not segments:
+                continue
+            # NB: Axis change indication is no longer required,
+            # the m2nano does not support it, the other devices do not need it...
+            if mode == ROW:
+                if xy != last_y:
+                    last_y = xy
+                    yield last_x, last_y, 0
+                if dx > 0:
+                    # from left to right
+                    idx = 0
+                    start = 0
+                    end = 1
+                    edge_start = -0.5
+                    edge_end = 0.5
+                else:
+                    idx = len(segments) - 1
+                    end = 0
+                    start = 1
+                    edge_start = 0.5
+                    edge_end = -0.5
+                while 0 <= idx < len(segments):
+                    sx = segments[idx][start] + edge_start
+                    ex = segments[idx][end] + edge_end
+                    on = segments[idx][2]
+                    if last_x != sx:
+                        yield sx, last_y, 0
+                    last_x = ex
+                    yield last_x, last_y, on
+                    idx += dx
+            else:
+                if xy != last_x:
+                    last_x = xy
+                    yield last_x, last_y, 0
+                if dy > 0:
+                    # from top to bottom
+                    idx = 0
+                    start = 0
+                    end = 1
+                    edge_start = -0.5
+                    edge_end = 0.5
+                else:
+                    idx = len(segments) - 1
+                    end = 0
+                    start = 1
+                    edge_start = 0.5
+                    edge_end = -0.5
+                while 0 <= idx < len(segments):
+                    sy = segments[idx][start] + edge_start
+                    ey = segments[idx][end] + edge_end
+                    on = segments[idx][2]
+                    if last_y != sy:
+                        yield last_x, sy, 0
+                    last_y = ey
+                    yield last_x, last_y, on
+                    idx += dy
+            if self.bidirectional:
+                if mode == ROW:
+                    dx = -dx
+                else: # column
+                    dy = -dy
+
+        # We need to set the final values so that the rastercut is able to carry on
+        self.final_x = last_x
+        self.final_y = last_y
+        t3 = perf_counter()
+        if self.debug_level > 1:
+            print (f"Overall time for crossover consumption: {t3-t0:.2f}s")
+            print (f"Computation: {t2 - t0:.2f}s - Array creation:{t1 - t0:.2f}s, Algorithm: {t2 - t1:.2f}s")
+    """
+    # Testpattern generation
+    def testpattern_generator(self):
+        def rectangle_h():
+            # simple rectangle
+            self.initial_x = 0
+            self.initial_y = 0
+            self.final_x = 0
+            self.final_y = 0
+            self.horizontal = True
+
+            yield 0, 0, off
+            yield self.width - 1, 0, on
+
+            yield self.width - 1, self.height - 1, on
+
+            yield 0, self.height - 1, on
+
+            yield 0, 0, on
+
+        def rectangle_v():
+            # simple rectangle but start with y movements first
+            self.initial_x = 0
+            self.initial_y = 0
+            self.final_x = 0
+            self.final_y = 0
+            self.horizontal = True
+
+            yield 0, 0, off
+            yield 0, self.height - 1, on
+
+            yield self.width - 1, self.height - 1, on
+
+            yield self.width - 1, 0, on
+
+            yield 0, 0, on
+
+
+        def snake_h():
+            # horizontal snake
+            self.initial_x = 0
+            self.initial_y = 0
+            x = 0
+            y = 0
+            self.horizontal = True
+            yield 0, 0, off
+            wd = self.width - 1
+            left = True
+            while y < self.height - 2:
+                x = wd if left else 0
+                self.horizontal = True
+                yield x, y, on
+                self.horizontal = False
+                yield x, y + 2, on
+                left = not left
+                y += 2
+                self.final_x = x
+                self.final_y = y
+
+        def snake_v():
+            # vertical snake
+            self.initial_x = 0
+            self.initial_y = 0
+            x = 0
+            yield 0, 0, off
+            ht = self.height - 1
+            top = True
+            while x < self.width - 2:
+                y = ht if top else 0
+                self.horizontal = False
+                yield x, y, on
+                self.horizontal = True
+                yield x + 2, y, on
+                top = not top
+                x += 2
+                self.final_x = x
+                self.final_y = y
+
+        def spiral():
+            # Spiral to inside
+            self.initial_x = 0
+            self.initial_y = 0
+            yield 0, 0, off
+
+            x = -2 # start
+            y = 0
+            width = self.width + 1
+            height = self.height - 1
+            while width > 0 and height > 0:
+                x += width
+                y += 0
+                self.horizontal = True
+                yield x, y, on
+                x += 0
+                y += height
+                self.horizontal = False
+                yield x, y, on
+                x -=(width - 2)
+                y += 0
+                self.horizontal = True
+                yield x, y, on
+                x += 0
+                y -= (height - 2)
+                self.horizontal = False
+                yield x, y, on
+                width -= 4
+                height -= 4
+                self.final_x = x
+                self.final_y = y
+
+        on = self.filter(0)
+        off = 0
+        # print (f"on={on}, off={off}")
+        method = abs(self.direction) - 1
+        methods = (rectangle_h, rectangle_v, snake_h, snake_v, spiral)
+        try:
+            yield from methods[method]()
+        except IndexError:
+            print (f"Unknown testgenerator for {self.direction}")
+    """