exprlens 0.0.4

Lenses from python expressions.

Python Expression Lenses

Python expressions to/from lenses. This library focuses on constructing lenses from python expressions, targeting function bindings as the primary container to get/set from/to. Such lenses can be serialized into strings which are valid python expressions (and thus can be deserialized using the python expression parser).

Alternatives

This package has very specific goals which are not considerations in the theory of lenses. Consider these alternatives for more general lens usage:

• lenses - more general and feature rich lenses implementation based on lenses in Haskell.
• pylens - simpler
• simplelens - unknown functionality

Basic usage

from exprlens import arg

# Lens for getting the first item in the first argument from a call:
first = arg[0]

# Getter can be accessed using the `get` method or by calling (`__call__`):
assert first.get([0, 1, 2, 3]) == first([0, 1, 2, 3]) == 0

# Setter can be accessed using the `set` method.
assert first.set([0, 1, 2, 3], val=42) == [42, 1, 2, 3]

Predefined basic lenses

from exprlens import arg, kwargs, args, argskwargs, ArgsKwargs, ident, lens, arguments, all

# All positional arguments:
assert args.get(42) == (42, )

# All keyword arguments:
assert kwargs.get(hello="there") == {'hello': 'there'}

# All arguments, stored as `ArgsKwargs` object:
assert argskwargs.get(42, hello="there") \
    == ArgsKwargs(args=(42,), kwargs={'hello': 'there'})

# argskwargs, arguments, and all are aliases.
assert argskwargs is arguments is all

# Only positional or keyword argument:
assert arg.get(42) == 42
assert arg.get(hello="there") == 'there'
# arg.get(1,2) # error if more than one argument is given

# lens, argm and ident are aliases.
assert arg is ident is lens

Lense iterators

from exprlens.valid import ValidDict, ValidMapping, ValidTuple, ValidList, ValidSequence, ValidJSON

seq = [1, 2, 3, 4]
# Lenses for each element in a sequence:
value_lenses = list(ValidSequence.elements(seq))
assert value_lenses[0].get(seq) == 1
assert value_lenses[1].get(seq) == 2
assert value_lenses[2].get(seq) == 3

# Elements can be set:
# Set the value at index 1 to 42:
assert value_lenses[1].set(seq, val=42) == [1, 42, 3, 4]

# Indices in a sequence:
seq = [1, 2, 3, 4]
index_lenses = list(ValidSequence.indices(seq))
assert index_lenses[0].get(seq) == 0
assert index_lenses[1].get(seq) == 1
assert index_lenses[2].get(seq) == 2

# Indices can be set:
# Index of element 1 to 3 thus copying the value at index 1 to index 3:
# Items at old index is spliced out.
assert index_lenses[1].set(seq, val=3) == [1, 3, 2]

# Values in a mapping:
mapping = {'a': 1, 'b': 2}
value_lenses = list(ValidMapping.values(mapping))
assert value_lenses[0].get(mapping) == 1
assert value_lenses[1].get(mapping) == 2

# Values can be set:
# Set the value of key 'a' to 42:
assert value_lenses[0].set(mapping, val=42) == {'a': 42, 'b': 2}

# Keys in a mapping:
key_lenses = list(ValidMapping.keys(mapping))
assert key_lenses[0].get(mapping) == 'a'
assert key_lenses[1].get(mapping) == 'b'

# Keys can be set:
# Key 'a' to 'c' thus renaming the key:
assert key_lenses[0].set(mapping, val='c') == {'c': 42, 'b': 2}

# JSON-like objects:
obj = {'baseint': 1, 'seqofints': [1, 2, 3], 'seqofstrs': ['a', 'b', 'c']}

# Get all lenses to JSON values including sequences and dictionaries:
for l in ValidJSON.all_values(obj):
    print(l, l.get(obj))

# Get lenses only to base JSON values (int, str, bool, None):
for l in ValidJSON.base_values(obj):
    print(l, l.get(obj))

# Get lenses only to values of the specified type, here int:
for l in ValidJSON.of_type(obj, int):
    print(l, l.get(obj))

Expressions

from exprlens import arg

# Lens that first grabs the first two items in the first argument and adds them:
plusfirsts = arg[0] + arg[1]

# Note that once expression lenses are constructed, `set` can no longer be used on them.
# plusfirsts.set([1, 2, 3, 4], val=42)  # Raises an exception.

assert plusfirsts([1, 2, 3, 4]) == 3

# Literals/constants: Expression lenses can involve literals/constants.
plusone = arg + 1
assert list(map(plusone, [1, 2, 3])) == [2, 3, 4]

Validation

from exprlens import args, kwargs

# Lenses can be validate on construction:
args[0]  # ok
# args["something"] # error

kwargs["something"]  # ok
# kwargs[0] # error

# Lenses can be validated on use:

# Lens that gets a key from the first argument, thus the first argument must be a mapping:
firstkey = arg["something"]
firstkey.get({"something": 42})  # ok

# Will fail if the first argument is not a mapping:
# firstkey.get([1,2,3]) # error

Boolean expressions

Python does not allow overriding boolean operators (and, or, not) (see relevant rejected PEP) so lenses corresponding to expressions with boolean operators cannot be created by writing python directly, i.e. lens[0] and lens[1]. Instead you can make use of Lens.conjunction, Lens.disjunction, and Lens.negation static methods to construct these. Alternatively you can use Lens.of_string static method to construct it from python code, e.g. Lens.of_string("lens[0] and lens[1]").

from exprlens import Lens, lens

# Conjunction:
both = Lens.of_string("lens[0] and lens[1]")

assert Lens.conjunction(lens[0], lens[1]) \
    == both

assert both([1, 2, 3, 4]) == 2

# Disjunction:
either = Lens.of_string("lens[0] or lens[1]")

assert Lens.disjunction(lens[0], lens[1]) \
    == either

assert either([1, 2, 3, 4]) == 1

# Logical negation:
not_second = Lens.of_string("not lens[1]")

assert Lens.negation(lens[1]) \
    == not_second

assert not_second([1, 2, 3, 4]) == False

Serialization

from exprlens import Expr, Lens, arg
from ast import parse

plusone = arg + 1
assert repr(plusone) == "(lens + 1)"

# assert plusone.pyast == parse(str(plusone), mode="eval")
assert plusone == Lens.of_string(str(plusone))