Parser

Python package for quick creation/ combination of value parser object for muly-puposes

Install

> pip install valueparser

Usage

from valueparser.parsers import Bounded

value_parser = Bounded(min=0, max=1)

value_parser.parse(4.0)

# ParseError: 4.0 is higher than 1.0

Several parsers can be combined to one single parser object, they share however the same name space for configuration parameters

from valueparser import parser, Clipped, Rounded

ratio_parser = parser( (float, Clipped, Rounded),  min=0, max=1.0, ndigits=2 )

assert ratio_parser.parse( 0.231234) == 0.23 
assert ratio_parser.parse( 4.5) == 1.0 
assert ratio_parser.parse( "0.12345") == 0.12

Equivalent can be done by creating a new class

from valueparser import parser_class, Clipped, Rounded

MyParser = parser_class( (float, Clipped, Rounded), "MyParser" )
ratio_parser = MyParser( min=0, max=1, ndigits=2)

conparser works the same way than parser except it construct a typing object to be use inside pydantic BaseModel in a compact way.

from valueparser import conparser, Bounded
from pydantic import BaseModel 

Pixel =  conparser( (int, Bounded), min=0, max=1023 ) 
class Data(BaseModel):
    x: Pixel = 512
    y: Pixel = 512
   
Data(x=-200)

# 
#pydantic.error_wrappers.ValidationError: 1 validation error for Data
#x
#    -200.0 is lower than 0.0 (type=value_error.parse; error_code=Errors.OUT_OF_BOUND)

to make any function a parser (e.g. an object with parse method) one can use the parser function as well :

from valueparser import parser

float_parser = parser(float)
assert float_parser.parse("1.234") == 1.234

force_int_parser = parser( (float, int)) # parse to float then int 
assert force_int_parser.parse( "1.234") == 1

Actually the parser function accepts :

• A Parser Class iddentified as a class with the parse method
• A callable
• An instance of a Parser Class
• an mix inside an iterable

Plus any kwargs accepted by the combination of parsers

Builtin Parsers

Create a custom parser

To create a parser one need to create a new class from BaseParser, declare any configurable argument inside the child class Config and define the static or classmethod __parse__

For instance a parser adding some noise to a value ( can be usefull for e.i. a simulator)

from valueparser import BaseParser
import random 

class Noisier(BaseParser):
    class Config:
        noise_scale = 1.0
    @staticmethod
    def __parse__( value, config):
        return value + (random.random()-0.5) * config.noise_scale

Usage :

noisier = Noisier( noise_scale=100)
x = noisier.parse(0.0)
x
36.700125482238036

Or use conparser in pydantic Model:

from valueparser import conparser 
from pydantic import BaseModel  

class MyData(BaseModel):
    x: conparser(Noisier, noise_scale=100)
    y: conparser(Noisier, noise_scale=100)

my_data = MyData(x=0, y=0)
my_data
MyData(x=32.819723479459284, y=-25.95893228872207)

Parser and Systemy

Parsers are beased on :class:systemy.System class. One can include a parser factory in a systemy class and expose the parser configuration to user.

from valueparser import ParserFactory , Bounded
from systemy import BaseSystem 
from pydantic import AnyUrl 

dummy = lambda x:x 

class Node(BaseSystem):
    class Config:
        url: AnyUrl = "http://localhost:4840"
        parser: ParserFactory = ParserFactory(type=dummy)

    def set(self, value):
        value = self.parser.parse(value) 
        # set value on server 
        return value

node = Node( parser={'type':(float,Bounded), 'min':0, 'max':10} )