Skip to content

evtn/soda

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

79 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

soda - a fast SVG generation tool

GitHub Workflow Status GitHub Workflow Status PyPI PyPI - Downloads PyPI - Python Version Coveralls License wordstreamer badge Badge Count

Here's some basic usage:

Note

Since soda 2.0.0, a Tag instance is also a wordstreamer.Renderable, which makes it possible to use things like:

from wordstreamer import Renderer
from soda import Tag

byte_stream = Renderer().byte_stream(Tag.rect()) # Iterable[bytes]
from soda import Tag, Root

# root is a custom svg tag
root = Root(viewBox="0 0 10 10")(
    Tag.rect(width=10, height=10, fill="#ff5"),
    Tag.circle(cx=5, cy=5, r=4, fill="#222")
)

print(root.render(pretty=True))

Installation

Install soda-svg from PyPI, like python -m pip install soda-svg.

Note that soda on PyPI is a different package.

Tag construction

The main class of the module is Tag. You can create it with a constructor:

Tag("g")

or with a shorthand:

Tag.g

You can also pass children and attributes into the constructor:

Tag("g", child1, child2, attr1=1, attr2=2)

or as call arguments (this would change tag in place, no additional copies):

Tag("g")(child1, attr1=1)(child2, child3, attr2=10, attr3=3)
# or
Tag.g(child1, attr1=1)(child2, child3, attr2=10, attr3=3)

# '_' to '-' conversion
Tag.g(child1, attr_1=1)(child2, child3, attr_2=10, attr_3=3) # <g attr-1="1" attr-2="10" attr-3="3" >child1child2child3</g>

Float rounding

As floats can be used as attribute values (even though in resulting SVG every value is a string), module will round floats to 3 digits after the decimal point:

from soda import Tag

g = Tag.g(x=1/3)
print(g.render()) # '<g x="0.333"/>'

To change this behaviour, edit soda.config.decimal_length before value is assigned:

from soda import Tag, config as soda_config

soda_config.decimal_length = 4

g = Tag.g(x=1/3)
print(g.render()) # '<g x="0.3333"/>'

soda_config.decimal_length = 2
g(y=1/3)

print(g.render()) # '<g x="0.3333" y="0.33"/>'

Attribute conversion

For convenience, leading and trailing underscores are removed by default, and underscores in the middle of words are replaced with hyphens:

from soda import Tag

g = Tag.g(cla_ss_="test")

print(g.render()) # <g cla-ss="test"/>

To disable replacing behavior, use config class:

from soda import Tag, config as soda_config

soda_config.replace_underscores = False

g = Tag.g(cla_ss_="test")

print(g.render()) # <g cla_ss="test"/>

...and to disable stripping of leading/traililng underscores:

from soda import Tag, config as soda_config

soda_config.strip_underscores = False

g = Tag.g(cla_ss_="test")

print(g.render()) # <g cla-ss_="test"/>

It's important to do that before tag creation, as all conversions are happening at the tag creation time:

from soda import Tag, config as soda_config

g1 = Tag.g(cla_ss_="test") # g.attributes == {"cla-ss": "test"}

soda_config.replace_underscores = False

g2 = Tag.g(cla_ss_="test") # g.attributes == {"cla_ss_": "test"}

print(g1.render()) # <g cla-ss="test"/>
print(g2.render()) # <g cla_ss_="test"/>

Creating a Tag from XML string

new in 1.1.0

You can use Tag.from_str(xml_string) to parse an XML document in that string.
Note that currently this doesn't preserve any comments or declarations of original document.

from soda import Tag, Root

root = Root(viewBox="0 0 10 10")(
    Tag.rect(width=10, height=10, fill="#ff5"),
    Tag.circle(cx=5, cy=5, r=4, fill="#222")
)

rendered_root = root.render(pretty=True)
new_root = Tag.from_str(rendered_root)

assert rendered_root == new_root.render(pretty=True)

Text

Basic text handling is pretty straightforward:

from soda import Tag

Tag.text("Hello, World") # just pass a string as a children

This code is roughly equivalent to:

from soda import Tag, Literal

Tag.text(Literal("Hello, World"))

...except that first piece doesn't create a Literal object.

If you need to add unescaped text (such as prerendered XML), you should pass escape=False to a Literal constructor:

XML Declarations and comments

To insert an XML declaration (i.e. <?xml version="1.0" encoding="UTF-8"?>), use XMLDeclaration:

from soda import XMLDeclaration


print(XMLDeclaration(version="2.0", encoding="UTF-8").render()) # '<?xml version="2.0" encoding="UTF-8"?>'

Default values for version and encoding are "1.0" and "UTF-8" respectively

XML comments are used similarly:

from soda import XMLComment


print(XMLComment("comment text!!").render()) # '<!-- comment text!! -->'
from soda import Tag, Literal

Tag.g(Literal('<path d="M0 0 L10 0 Z"/>', escape=False))

Accessing data

tag[attr] syntax can be used to manage tag attributes (where attr should be a string).

from soda import Tag

tag = Tag.g
tag["id"] = "yes-thats-an-id" # sets attribute
tag["cool"] = None # deletes attribute if exists, otherwise does nothing
print(tag["id"]) # prints attribute
print(tag["non-existent-attribute"]) # prints None

tag[index] syntax can be used to manage tag children (where index should be either integer or slice).

from soda import Tag

tag = Tag.g(Tag.a)
tag[0]["href"] = "https://github.com/evtn/soda"
print(tag[1]) # IndexError
print(tag[0]) # prints <a href="https://github.com/evtn/soda" />

Children can also be accessed directly through tag.children attribute.

This is not necessary for most tasks as of 1.1.6 with new methods and iterator protocol:

  • Tag.insert(int, Node) inserts a node on an index. Be aware that tag.children is not flattened: Tag.g("test", ["test1", "test2"]).insert(2, elem) will insert elem after the array.
  • Tag.append(Node) appends one node to the tag.
  • Tag.extend(*Node) appends several nodes to the tag. *This is not the same as .append([*Node])*
  • Tag.pop(int?) pops one node from specified index. If index is not provided, pops the last one.
  • Tag.iter_raw() returns an iterable to get every Node of the tag. This doesn't dive into nested arrays, for that behaviour iterate over Tag
  • You can also iterate over the Tag itself to get every flat node of it (no arrays)

Fragments

Fragments use concept similar to React's fragment. It renders just it's children:

from soda import Tag, Fragment

tag = Tag.g(
    Fragment(Tag.a, Tag.a)
)
print(tag) # <g><a/><a/></g>

Paths

new in 0.1.7

There is a builder for SVG path commands in soda:

You can build a list of path commands using descriptive command names:

from soda import Tag, Root, Path

commands = (
    Path.moveto(x=10, y=30),
    Path.arc(
        radius_x=20,
        radius_y=20,
        # for convenience, omitted arguments
        # (here: x_axis_rotation and large_arc_flag) are set to 0
        sweep_flag=1,
        x=50,
        y=30,
    ),
    Path.arc(
        radius_x=20,
        radius_y=20,
        sweep_flag=1,
        x=90,
        y=30,
    ),
    Path.quadratic(
        x1=90,
        y1=60,
        x=50,
        y=90,
    ),
    Path.quadratic(
        x1=10,
        y1=60,
        x=10,
        y=30,
    ),
    Path.close()
)

...or using common SVG command names (letter case signifies if command is relative):

# or

commands = (
    Path.M(10, 30),
    Path.A(20, 20, 0, 0, 1, 50, 30),
    Path.A(20, 20, 0, 0, 1, 50, 30),
    Path.Q(90, 60, 50, 90),
    Path.Q(10, 60, 10, 30),
    Path.Z()
)

...and render it with Path.build(*commands, compact=False) method

root = Root(
    viewBox="0 0 100 100",
    use_namespace=True,
)(
    Tag.rect(width="100%", height="100%", fill="white"),
    Tag.path()(
        d=Path.build(*commands)
    )
)

print(root.render(pretty=True))

"""
yields:

<svg
  viewBox="0 0 100 100"
  version="2.0"
  xmlns="http://www.w3.org/2000/svg"
  xmlns:xlink="http://www.w3.org/1999/xlink"
>
  <rect
    width="100%"
    height="100%"
    fill="white"
  />
  <path
    d="M 10 30 A 20 20 0 0 1 50 30 A 20 20 0 0 1 50 30 Q 90 60 50 90 Q 10 60 10 30 Z"
  />
</svg>
"""

You can also optimize resulting path with compact argument:

print(Path.build(*commands, compact=True))
# prints M10 30A20 20 0 0 1 50 30A20 20 0 0 1 50 30Q90 60 50 90Q10 60 10 30Z

Points

new in 1.1.0

To work with coordinates, you can use Point class:

from soda import Point

a = Point(1, 2)
b = Point(4, 5)

In any context where a point could be used, "point-like" values can be used:

  • (1, 2) <-> Point(1, 2)
  • [1, 2] <-> Point(1, 2)
  • 1 <-> Point(1, 1)

You can use coordinates of a point:

print(a) # Point[1, 2]
print(a.x) # 1
print(a.coords) # (1, 2)
print([*a]) # [1, 2] (same as [*a.coords])

...perform mathematical operations on points:

print(a + b) # Point[5, 7]
print(a - b) # Point[-3, -3]
print(a * b) # Point[4, 10] (a.x * b.x, a.y * b.y)
print(a / b) # Point[0.25, 0.4]
print(a % b) # Point[1, 2]

...and any point-like values:

print(a + 10) # Point[11, 12]
print(a * 2) # Point[2, 4]

You also can calculate distance between points and rotate a point around some point:

from math import pi

print(a.distance(b)) # 4.242640687119285
print(a.distance()) # 2.23606797749979 (distance between a and (0, 0), basically the length of a vector)
print(a.rotate(degrees=90)) # Point[-2, 1]
print(a.rotate((10, 10), degrees=90))
print(a.rotate((10, 10), radians=pi / 2)) # Point[18, 1]

...and get a normalized vector:

print(a.normalized()) # Point[0.4472135954999579, 0.8944271909999159]

You also can get an angle (in radians) between two vectors (with specified starting point):

print(a.angle(b)) # 0.21109333322274684
print(a.angle(b, (10, 10))) # 0.03190406448501816 (second argument specifies starting point (0, 0) by default)
print(a.angle()) # 1.1071487177940904 (angle between `a` and (1, 0) vector)

print(
    a.angle(
        a.rotate(radians=2)
    )
) # 2

Using as attributes

Point.as_ provides a convenient way of using points as tag attributes:

from soda import Tag, Point

size = Point(100, 200)

print(size.as_()) # {"x": 100, "y": 200}
print(size.as_("width", "height")) # {"width": 100, "height": 200}


print(
    Tag.rect(
        **size.as_("width", "height")
    )
) # <rect width="100" height="200"/>

PointPath

A version of Path accepting Point instead of some arguments. Where Path.something(...) accepts coordinates (as two arguments) or some size (like radius_x and radius_y in arc), PointPath accepts a point-like object instead.

Custom components

You can build custom components, using different approaches:

Building a tree on init

Builds a tree on every component creation

from soda import Tag, Fragment

class CustomComponent(Fragment):
    def __init__(self):
        children = Tag.g(
            Tag.anythingother,
            Tag.lalala(
                Tag.yes,
                Tag.no
            )
        )
        super().__init__(*children)

CustomComponent().render()

Functional approach

Builds a tree on every call

from soda import Tag

def custom_component():
    return Tag.g(
        Tag.anythingother,
        Tag.lalala(
            Tag.yes,
            Tag.no
        )
    )

custom_component().render()

Speed

soda is able to render tens of thousands tags per second, but if you wanna optimize your execution, there are some tips:

Building a tree efficiently

If you using the same structure many times (especially if it's a heavy one), avoid rebuilds. Rather than building a new tree every time, consider changing specific parts of it when needed. It won't speed up the render time, though (check Prerendering right below for that)

Prerendering

If you have some static tags, you can use tag.prerender() to get a prerendered Literal. This could speed up your render significantly in some complex cases.