Problem
The module is for facilitating random string generation. The example I use to explain it is the following:
Instead of writing
#vars
age = random.randint(18, 20)
#categories
names = ('Mark', 'Sally')
adjs = ('wacky', 'cool')
adjs_probability_map = (0.1, 0.99)
hellos = ('Hi!', 'Hey!', 'Good morning.')
hellos_probability_map = (0.5, 0.25, 0.25)
#calculation
name = random.choice(names)
if name is 'Mark':
age += 2
adj = random.choices(adjs, adjs_probability_map)[0]
hello = random.choices(hellos, hellos_probability_map)[0]
#string
print(f'{hello} My name is {adj} {name}. I am {age} years old')
#Example output: Hey! My name is cool Mark. I am 21 years old.
with this module you can write
import rrsg
string = rrsg.generate("""
&age = (18, 20)
$name
Mark {age += 2}
Sally
$adj
wacky {1%}
cool
>[Hi! | Hey! | Good morning.] My name is $adj $name. I am &age years old.
""").strings()[0]
print(string)
which improves (at least that’s the purpose) the readability and, at the same time, makes it easy to have string templates in text files to use for random generation.
I’ve posted all the documentation on Github, and it might be useful to check it out to understand the generator syntax (those documents themselves might need some reviewing, as it is the first time I’m writing documentation).
Anyway, the code in __init__ is:
import re, random as rnd
class RandomGenerationError(Exception):
"""General exception for rrsg"""
def __init__(self, message):
super().__init__(message)
class _Compiler:
def __init__(self, generator):
self.generator = generator
def _compile_var(self, var_b):
"""
var_b = 'var = 1' TURNS INTO var_c = (var, (1, 1))
var_b = 'var = (3,4)' TURNS INTO var_c = (var, (3, 4))
"""
var_name = re.search('.+(?==)', var_b).group()
range_value, absolute_value = re.search('(?<==)((d+),(d+))', var_b), re.search('(?<==)((d+))', var_b)
if range_value: # range declaration
var_c = (var_name, tuple(int(num) for num in range_value.group(1, 2)))
elif absolute_value: # absolute declaration
var_c = (var_name, tuple(int(num) for num in absolute_value.group(1, 2)))
else:
raise RandomGenerationError('Faulty variable atribution.')
return var_c
def _compile_datatag(self, datatag_b):
"""
datatag_b = '{75%,var+=2}' TURNS INTO datatag_c = {'prob': 0.75, 'var': ('+', 2)}
"""
datatag_c = {}
for prop in datatag_b[1:-1].split(','): # e.g. ['50%', 'var+=3']
if '%' in prop: # e.g. 'prob': 0.5
datatag_c['prob'] = int(re.search('[0-9]{1,2}?(?=%)', prop).group()) * 0.01
elif '=' in prop: # e.g. 'var': ('+', 3)
datatag_c[re.search('.*?(?=+=|-=|=)', prop).group()] = (re.search('+|-|=', prop).group(), int(re.search('(?<==)[0-9]+', prop).group()))
else:
raise RandomGenerationError('Faulty datatag.')
return datatag_c
def _compile_items(self, items_b):
"""
items_b = ['a', 'b{75%,var+=2}'] TURNS INTO items_c = {'a': {'prob': 0.25}, 'b': {'prob': 0.75, 'var': ('+', 2)}}
"""
items_c = {}
for item_b in items_b:
datatag = re.search('{.*?}', item_b)
if datatag:
items_c[re.search('.+?(?={)', item_b).group()] = self._compile_datatag(datatag.group()) # e.g. b: {'prob': 0.75, 'var': ('+', 2)}
else:
items_c[item_b] = {} # e.g. a: {}
#complete probabilities
unassigned_prob = (1 - sum(datatag.get('prob', 0) for datatag in items_c.values())) / [datatag.get('prob') for datatag in items_c.values()].count(None)
for item, datatag in items_c.items():
if not datatag.get('prob'):
items_c[item]['prob'] = unassigned_prob
return items_c
def _compile_string(self, string_b, categories_c, vars_c):
"""
string_b = [Hi! | Hey! | Good morning.] My name is $adj $name. I am &age years old.
TURNS INTO
string_c = [
{'Hi!': {'prob': 0.33}, 'Hey!': {'prob': 0.33}, 'Good morning.': {'prob': 0.33}},
' My name is ',
{'cool': 'prob': 0.99, 'wacky': 'prob': 0.01},
' ',
{'Sally': {'prob': 0.5}, 'Mark': {'prob': 0.5, 'age': ('+', 2)}},
'. I am ',
('age', (18, 20)),
' years old.'
]
"""
string_c = []
while string_b:
category, bitesized, var, text = re.search('(?<=$)[a-zA-Z]+', string_b), re.search('[.*?]', string_b), re.search('(?<=&)[a-zA-Z]+', string_b), re.search('[^&[]|$]+', string_b)
if category and category.start() is 1:
string_c += [categories_c[category.group()]]
string_b = string_b.replace('$' + category.group(), '', 1)
elif bitesized and bitesized.start() is 0:
string_c += [self._compile_items(bitesized.group()[1:-1].split('|'))]
string_b = string_b.replace(bitesized.group(), '', 1)
elif var and var.start() is 1:
string_c += [[var_c for var_c in vars_c if var_c[0] == var.group()][-1]]
string_b = string_b.replace(f'&{var.group()}', '', 1)
elif text:
string_c += [text.group()]
string_b = string_b.replace(text.group(), '', 1)
return string_c
def _compile(self):
"""
Compiles a generator. Returns a list of compiled strings (strings_c).
"""
vars, categories, strings, querying = [], {}, [], False
for line in self.generator.split('n'):
# remove comments
line = re.sub('//.*', '', line)
# remove unnecessary spaces
# region
line = line.lstrip(' ').rstrip(' ') # remove spaces before and after line
line = re.sub('&(?: +)?', '&', line) # remove spaces after '&'
line = re.sub('$(?: +)?', '$', line) # remove spaces after '$'
line = re.sub('>(?: +)?', '>', line) # remove spaces after '>'
line = re.sub('(?: +)?=(?: +)?', '=', line) # remove spaces before and after '='
line = re.sub('(?: +)?(?:+=)(?: +)?', '+=', line) # remove spaces before and after '+='
line = re.sub('(?: +)?(?:-=)(?: +)?', '-=', line) # remove spaces before and after '-='
line = re.sub('(?: +)?|(?: +)?', '|', line) # remove spaces before and after '|'
line = re.sub('(?: +)?,(?: +)?', ',', line) # remove spaces before and after ','
line = re.sub('[(?: +)?', '[', line) # remove spaces after '['
line = re.sub('(?: +)?]', ']', line) # remove spaces before ']'
line = re.sub('(?: +)?{(?: +)?', '{', line) # remove spaces before and after '{'
line = re.sub('(?: +)?}(?: +)?', '}', line) # remove spaces before and after '}'
# endregion
# ignore empty lines
if line is '': continue
# iterpret line
if line[0] is '&':
# stop querying for category values and compile last category
if querying:
querying = False
categories[category_name] = self._compile_items(categories[category_name])
# create a new compiled var
var_s = re.search('(?<=&).+', line).group()
vars += [self._compile_var(var_s)]
elif line[0] is '$':
# start querying for category values and compile last category
if querying:
categories[category_name] = self._compile_items(categories[category_name])
else:
querying = True
# add a new empty category
category_name = re.search('(?<=$)[^ n]+', line).group()
categories[category_name] = []
elif line[0] is '>':
# stop querying for category values and compile last category
if querying:
querying = False
categories[category_name] = self._compile_items(categories[category_name])
# add phrase to results
string_b = re.search('(?<=>).+', line).group()
strings += [self._compile_string(string_b, categories, vars)]
elif querying:
# add raw item to category being queried
categories[category_name] += [line]
return Compiled(strings)
pass
class Compiled:
def __init__(self, strings_c):
self.strings_c = strings_c
def _generate_value(self, value_c):
"""
value_c = (1,4) TURNS INTO, E.G. value_f = 3
"""
return rnd.randint(*value_c)
def _generate_item(self, category_c):
"""
category_c = {'Sally': {'prob': 0.5}, 'Mark': {'prob': 0.5, 'age': ('+', 2)}} TURNS INTO, E.G. item_f = ('Mark', ('age', ('+', 2)))
"""
item = rnd.choices(list(category_c.keys()), [datatag['prob'] for datatag in category_c.values()])[0]
return (item, list(attr for attr in category_c[item].items() if attr[0] is not 'prob'))
def _generate_string(self, string_c):
"""
string_c = [
{'Hi!': {'prob': 0.33}, 'Hey!': {'prob': 0.33}, 'Good morning.': {'prob': 0.33}},
' My name is ',
{'cool': 'prob': 0.99, 'wacky': 'prob': 0.01},
' ',
{'Sally': {'prob': 0.5}, 'Mark': {'prob': 0.5, 'age': ('+', 2)}},
'. I am ',
('age', (18, 20)),
' years old.'
]
TURNS INTO
string_f = 'Hi! My name is cool Mark. I am 21 years old.'
"""
var_changes = []
for i, category_c in (obj_c for obj_c in list(enumerate(string_c)) if type(obj_c[1]) is dict):
item_f = self._generate_item(category_c)
string_c[i] = item_f[0]
if item_f[1]: var_changes += item_f[1]
for i, var_c in (obj_c for obj_c in enumerate(string_c) if type(obj_c[1]) is tuple):
var_f = self._generate_value(var_c[1])
for var, operation in (var_change for var_change in var_changes if var_change[0] == var_c[0]):
var_f = {'=': lambda x, y: y,
'+': lambda x, y: x + y,
'-': lambda x, y: x - y}[operation[0]](var_f, operation[1])
string_c[i] = str(var_f)
return ''.join(string_c)
def generate(self):
"""
Returns a Results object containing all final strings.
"""
strings_f = []
for string_c in self.strings_c:
strings_f += [self._generate_string(string_c)]
return Results(strings_f)
class Results:
def __init__(self, strings_f):
self.strings_f = strings_f
def strings(self):
return self.strings_f
def compile(generator):
"""Returns a Compiled object useful when generating multiple times from the same generator."""
return _Compiler(generator)._compile()
def generate(generator):
"""Returns a Results object containing the generation results and the output data."""
return compile(generator).generate()
I would like some notes and tips on how my code could be improved, as I am quite new to Python (tested the module on Python 3.7) and I have never tried this type of project.
Solution
1. Quick fixes in the first part
I wouldn’t do that:
names = ('Mark', 'Sally')
adjs = ('wacky', 'cool')
adjs_probability_map = (0.1, 0.99)
hellos = ('Hi!', 'Hey!', 'Good morning.')
hellos_probability_map = (0.5, 0.25, 0.25)
adj = random.choices(adjs, adjs_probability_map)[0]
hello = random.choices(hellos, hellos_probability_map)[0]
instead:
names = {'Mark', 'Sally'}
adjs = {'wacky': 0.1, 'cool': 0.99}
hellos = ('Hi!': 0.5, 'Hey!': 0.25, 'Good morning.': 0.25)
adj = random.choices(*adjs.items())[0]
hello = random.choices(*hellos.items())[0]
Also:
sum(adjs_probability_map) # 1.09
I don’t think that was intended
2. Problems in second part:
re.search('(?<==)((d+),(d+))', var_b), re.search('(?<==)((d+))', var_b)
You have to escape your backslashes
re.search('(?<==)\((\d+),(\d+)\)', var_b), re.search('(?<==)((\d+))', var_b)
def compile(generator):
"""Returns a Compiled object useful when generating multiple times from the same generator."""
return _Compiler(generator)._compile()
compile is a builtin, don’t overload it
class RandomGenerationError(Exception):
"""General exception for rrsg"""
def __init__(self, message):
super().__init__(message)
Equivalent to :
class RandomGenerationError(Exception):
"""General exception for rrsg"""
pass
Which can be replaced by :
Because they are plenty of more descriptive built-in exceptions and you probably don’t need to create more:
BaseException
+-- SystemExit
+-- KeyboardInterrupt
+-- GeneratorExit
+-- Exception
+-- StopIteration
+-- StopAsyncIteration
+-- ArithmeticError
| +-- FloatingPointError
| +-- OverflowError
| +-- ZeroDivisionError
+-- AssertionError
+-- AttributeError
+-- BufferError
+-- EOFError
+-- ImportError
| +-- ModuleNotFoundError
+-- LookupError
| +-- IndexError
| +-- KeyError
+-- MemoryError
+-- NameError
| +-- UnboundLocalError
+-- OSError
| +-- BlockingIOError
| +-- ChildProcessError
| +-- ConnectionError
| | +-- BrokenPipeError
| | +-- ConnectionAbortedError
| | +-- ConnectionRefusedError
| | +-- ConnectionResetError
| +-- FileExistsError
| +-- FileNotFoundError
| +-- InterruptedError
| +-- IsADirectoryError
| +-- NotADirectoryError
| +-- PermissionError
| +-- ProcessLookupError
| +-- TimeoutError
+-- ReferenceError
+-- RuntimeError
| +-- NotImplementedError
| +-- RecursionError
+-- SyntaxError
| +-- IndentationError
| +-- TabError
+-- SystemError
+-- TypeError
+-- ValueError
| +-- UnicodeError
| +-- UnicodeDecodeError
| +-- UnicodeEncodeError
| +-- UnicodeTranslateError
+-- Warning
+-- DeprecationWarning
+-- PendingDeprecationWarning
+-- RuntimeWarning
+-- SyntaxWarning
+-- UserWarning
+-- FutureWarning
+-- ImportWarning
+-- UnicodeWarning
+-- BytesWarning
+-- ResourceWarning
Also :
generate runs a function generate on the result of
compile
which calls _compile
on a _Compiler object which returned a Compiled object
That might make sense to you, but I doubt people will take the time to read your program if you name your variables like this.