Improve performance of fake_username

[caioariede]

Description

This PR introduces a backwards-incompatible change by removing the rand_range argument from fake_username. With our focus being on performance, getting rid of expensive calls, such as random, is essential. The same/old behavior can still be accomplished using a custom function.

The impact of this change, being backwards-incompatible, is still minimal since as mentioned above the same behavior can still be accomplished using a custom function, and, the rand_range argument was optional.

Profiling results

Before changes

Total time: 35.0136 s
File: anon/utils.py
Function: fake_username at line 260

Line #      Hits         Time  Per Hit   % Time  Line Contents
==============================================================
   260                                           @profile
   261                                           def fake_username(max_size=10, separator="", rand_range=(1000, 999999)):
   262                                               """ Returns fake username
   263
   264                                               :max_size: Maximum number of chars
   265                                               :separator: Word separator
   266                                               :rand_range: Range to use when generating random number
   267
   268                                               """
   269   1000000    1481137.0      1.5      4.2      rand_start, rand_end = rand_range
   270   1000000    1197990.0      1.2      3.4      if not rand_end > rand_start:
   271                                                   raise ValueError(
   272                                                       "rand_range start ({}) must be > end ({})".format(rand_start, rand_end)
   273                                                   )
   274
   275   1000000   18277722.0     18.3     52.2      random_number = str(random.randint(rand_start, rand_end))
   276   1000000    1983915.0      2.0      5.7      min_size_allowed = _min_word_size + len(random_number)
   277
   278   1000000    1242765.0      1.2      3.5      if max_size < min_size_allowed:
   279                                                   raise ValueError("username must be >= {}".format(min_size_allowed))
   280                                               else:
   281   1000000    1369452.0      1.4      3.9          max_size -= len(random_number)
   282
   283   1000000    9460604.0      9.5     27.0      return fake_text(max_size, separator=separator) + random_number

After changes

Total time: 15.1933 s (2.3 times faster)
File: anon/utils.py
Function: fake_username at line 290

Line #      Hits         Time  Per Hit   % Time  Line Contents
==============================================================
   290                                           @profile
   291                                           def fake_username(max_size=10, separator=""):
   292                                               """ Returns fake username
   293
   294                                               :max_size: Maximum number of chars
   295                                               :separator: Word separator
   296                                               :rand_range: Range to use when generating random number
   297
   298                                               """
   299   1000000    2576025.0      2.6     17.0      random_number = str(next(_small_int_generator))
   300   1000000    2003946.0      2.0     13.2      min_size_allowed = _min_word_size + len(random_number)
   301
   302   1000000    1035719.0      1.0      6.8      if max_size < min_size_allowed:
   303                                                   raise ValueError("username must be >= {}".format(min_size_allowed))
   304                                               else:
   305   1000000    1324213.0      1.3      8.7          max_size -= len(random_number)
   306
   307   1000000    8253427.0      8.3     54.3      return fake_text(max_size, separator=separator) + random_number

Todos

• Tests
• Documentation
• Changelog

[github-actions]

File	Coverage
All files	95%	✅
anon/init.py	63%	✅
anon/base.py	97%	✅
anon/utils.py	87%	✅
tests/compat.py	50%	✅

Minimum allowed coverage is 50%

Generated by 🐒 cobertura-action against d644cd2