← Back
Platformer Tutorial
Start by importing ursina and creating a window.
from ursina import *
app = Ursina()
A simple way to get stared is to use the built in platformer controller.
It's pretty basic, so you might want to write your own at a later point.
It is however a good starting point, so let's import it like this:
from ursina.prefabs.platformer_controller_2 d import PlatformerController2 d
player = PlatformerController2 d(y =1 , z =.0 1 , scale_y =1 , max_jumps=2 )
You can change settings like jump_height, walk_speed, and gravity.
If you want to larn more about how it works you can read its code here:
https://github.com/pokepetter/ursina/blob/master/ursina/prefabs/platformer_controller_2d.py
If we try to play the game right now, you'll fall for all infinity, so let's add a ground:
Now, it works, but it's a pretty boring game, so let's make a more interesting level.
There are many ways to go about making a level, but for this we'll make an image
where we can simply draw the level and then generate a level based on that.
↑
Make sure to save this image to same folder or below as your script.
To generate the level we'll loop through all the pixels in the image above and do
something based on the color of the pixel. If it's white, it's air, so we'll skip it.
Now, we *could* create an Entity for each tile, but that's slower to render than one Entity with a custom model.
To make the model, we'll use the Mesh class. You'll have to be somewhat familiar with
what meshes are to do this, and know what vertices and uvs are. However, we'll just be
copying the quad model and offset the vertices. If we wanted to use a tileset with different
graphics for each tile, we'd scale and offset the uvs as well.
Set the camera to orthographic so there's no perspective.
Move the camera to the middle of the level and set the fov so the level fits nicely.
Setting the fov on an orthographic camera means setting how many units vertically the camera can see.
Loads an image sequence as a frame animation.
So if you have some frames named image_000.png, image_001.png, image_002.png and so on,
you can load it like this: Animation('image')
You can also load a .gif by including the file type: Animation('image.gif')
player.walk_animation = Animation('player_walk')
the platformer controller has an Animator and will toggle the state based on
whether it's standing still, is walking or is jumping.
All the Animator does is to make sure only one Animation is enabled at the same time.
Otherwise they would overlap.
self.animator = Animator({'idle' : None, 'walk' : None, 'jump' : None})
from ursina import *
app = Ursina()
from ursina.prefabs.platformer_controller_2 d import PlatformerController2 d
player = PlatformerController2 d(y =1 , z =.0 1 , scale_y =1 , max_jumps=2 )
ground = Entity (model ='quad' , scale_x =1 0 , collider ='box' , color =color.black)
quad = load_model('quad' , use_deepcopy =True)
level_parent = Entity (model =Mesh(vertices =[], uvs =[]), texture ='white_cube' )
def make_level(texture):
# destroy every child of the level parent.
# This doesn't do anything the first time the level is generated, but if we want to update it several times
# this will ensure it doesn't just create a bunch of overlapping entities.
[destroy(c) for c in level_parent.children]
for y in range (texture.height):
collider = None
for x in range (texture.width):
col = texture.get_pixel(x,y)
# If it's black, it' s solid, so we'll place a tile there.
if col == color.black:
level_parent.model.vertices += [Vec3(*e) + Vec3(x+.5 ,y+.5 ,0 ) for e in quad.generated_vertices] # copy the quad model, but offset it with Vec3(x+.5 ,y+.5 ,0 )
level_parent.model.uvs += quad.uvs
# Entity (parent =level_parent, position =(x,y), model ='cube' , origin =(-.5 ,-.5 ), color =color.gray, texture ='white_cube' , visible =True)
if not collider:
collider = Entity (parent =level_parent, position =(x,y), model ='quad' , origin =(-.5 ,-.5 ), collider ='box' , visible =False)
else :
# instead of creating a new collider per tile, stretch the previous collider right.
collider.scale_x += 1
else :
collider = None
# If it's green, we' ll place the player there. Store this in player.start_position so we can reset the plater position later.
if col == color.green:
player.start_position = (x, y)
player.position = player.start_position
level_parent.model.generate()
make_level(load_texture('platformer_tutorial_level' )) # generate the level
camera.orthographic = True
camera.position = (3 0 /2 ,8 )
camera.fov = 1 6
player.traverse_target = level_parent
enemy = Entity (model ='cube' , collider ='box' , color =color.red, position =(1 6 ,5 ,-.1 ))
def update():
if player.intersects(enemy).hit:
print ('die' )
player.position = player.start_position
app.run()
Start by importing ursina and creating a window.
Using the built in platformer controller
A simple way to get stared is to use the built in platformer controller.
It's pretty basic, so you might want to write your own at a later point.
It is however a good starting point, so let's import it like this:
You can change settings like jump_height, walk_speed, and gravity.
If you want to larn more about how it works you can read its code here:
https://github.com/pokepetter/ursina/blob/master/ursina/prefabs/platformer_controller_2d.py
If we try to play the game right now, you'll fall for all infinity, so let's add a ground:
ground = Entity (model ='quad' , scale_x =1 0 , collider ='box' , color =color.black)
Making a "level editor"
Now, it works, but it's a pretty boring game, so let's make a more interesting level.
There are many ways to go about making a level, but for this we'll make an image
where we can simply draw the level and then generate a level based on that.
↑
Make sure to save this image to same folder or below as your script.
To generate the level we'll loop through all the pixels in the image above and do
something based on the color of the pixel. If it's white, it's air, so we'll skip it.
Now, we *could* create an Entity for each tile, but that's slower to render than one Entity with a custom model.
To make the model, we'll use the Mesh class. You'll have to be somewhat familiar with
what meshes are to do this, and know what vertices and uvs are. However, we'll just be
copying the quad model and offset the vertices. If we wanted to use a tileset with different
graphics for each tile, we'd scale and offset the uvs as well.
quad = load_model('quad' , use_deepcopy =True)
level_parent = Entity (model =Mesh(vertices =[], uvs =[]), texture ='white_cube' )
def make_level(texture):
# destroy every child of the level parent.
# This doesn't do anything the first time the level is generated, but if we want to update it several times
# this will ensure it doesn't just create a bunch of overlapping entities.
[destroy(c) for c in level_parent.children]
for y in range (texture.height):
collider = None
for x in range (texture.width):
col = texture.get_pixel(x,y)
# If it's black, it' s solid, so we'll place a tile there.
if col == color.black:
level_parent.model.vertices += [Vec3(*e) + Vec3(x+.5 ,y+.5 ,0 ) for e in quad.generated_vertices] # copy the quad model, but offset it with Vec3(x+.5 ,y+.5 ,0 )
level_parent.model.uvs += quad.uvs
# Entity (parent =level_parent, position =(x,y), model ='cube' , origin =(-.5 ,-.5 ), color =color.gray, texture ='white_cube' , visible =True)
if not collider:
collider = Entity (parent =level_parent, position =(x,y), model ='quad' , origin =(-.5 ,-.5 ), collider ='box' , visible =False)
else :
# instead of creating a new collider per tile, stretch the previous collider right.
collider.scale_x += 1
else :
collider = None
# If it's green, we' ll place the player there. Store this in player.start_position so we can reset the plater position later.
if col == color.green:
player.start_position = (x, y)
player.position = player.start_position
level_parent.model.generate()
make_level(load_texture('platformer_tutorial_level' )) # generate the level
Positioning the camera
Set the camera to orthographic so there's no perspective.
Move the camera to the middle of the level and set the fov so the level fits nicely.
Setting the fov on an orthographic camera means setting how many units vertically the camera can see.
camera.orthographic = True
camera.position = (3 0 /2 ,8 )
camera.fov = 1 6
Adding player graphics and animations
Loads an image sequence as a frame animation.
So if you have some frames named image_000.png, image_001.png, image_002.png and so on,
you can load it like this: Animation('image')
You can also load a .gif by including the file type: Animation('image.gif')
player.walk_animation = Animation('player_walk')
the platformer controller has an Animator and will toggle the state based on
whether it's standing still, is walking or is jumping.
All the Animator does is to make sure only one Animation is enabled at the same time.
Otherwise they would overlap.
self.animator = Animator({'idle' : None, 'walk' : None, 'jump' : None})
player.traverse_target = level_parent
enemy = Entity (model ='cube' , collider ='box' , color =color.red, position =(1 6 ,5 ,-.1 ))
def update():
if player.intersects(enemy).hit:
print ('die' )
player.position = player.start_position
Start the game
app.run()