What is this?

Simple automated tests; for grug devs like me who don't have time to test.

How do I use this?

pip install grug_test

Super Barebones Example (Not Recommended):

In your main.py equivlent:

from grug_test import grug_test

@grug_test(should_record_io=True, should_run_tests=True)
def repeat(a,times):
    for _ in range(times):
        a += f"{a}"
    return a

### [[Using your function like you normally would]] ###
print("hello world")
print(repeat("hello world", times=10))

Thats it!

• Run the main.py and grug_test will record the input/output of repeat (those are the tests)
• Change the code of the repeat function, then run main.py again
  • If the output files change AND you like the changes, ✨volia✨ those are your auto-updated test cases
  • If the output files change and you DON'T like the changes, then it failed the test case ❌
  • If the output files don't change and they should've, then it also failed the test case ❌
  • If the output files don't change and you're okay with that, then your tests passed ✅

Normal Example (Recommended)

from grug_test import GrugTest, grug_test
import os

# use ENV flags
GrugTest.force_fully_disable = os.environ.get("PROD")!=None
GrugTest.default_record_io = os.environ.get("GRUG_RECORD")!=None
GrugTest.default_run_tests = os.environ.get("GRUG_TEST")!=None

@grug_test(max_io=20) # max_io is basically "max number of recorded tests"
def repeat(a,times):
    for _ in range(times):
        a += f"{a}"
    return a

### [[Using your function like you normally would]] ###
print("hello world")
print(repeat("hello world", times=10))

Now here is a more respectable setup:

• PROD=True python ./main.py totally disable grug_test
• GRUG_RECORD=True python ./main.py will record the inputs/outputs of the function (commit the files to git!)
• GRUG_TEST=True python ./main.py will load your saved inputs, then generate/save the new outputs
  • If the output files change AND you like the changes, ✨volia✨ those are your auto-updated test cases
  • If the output files change and you DON'T like the changes, then it failed the test case ❌
  • If the output files don't change and they should've, then it also failed the test case ❌
  • If the output files don't change and you're okay with that, then your tests passed ✅

Big Project Example

from grug_test import GrugTest, grug_test
import os

# enable groups
GrugTest.force_fully_disable = os.environ.get("PROD")!=None
GrugTest.default_record_io = [ "string_tools", ]
GrugTest.default_run_tests = [ "string_tools", ]

@grug_test(tags=["string_tools",])
def repeat(a,times):
    for _ in range(times):
        a += f"{a}"
    return a

@grug_test(tags=["feature_branch_1",])
def helper1_for_new_feature(something):
    return something + 1

### [[Using your function like you normally would]] ###
print("hello world")
print(repeat("hello world", times=10))
print(repeat("hello world", times=helper1_for_new_feature(10)))

Same as the normal example above, except only string_tools (like the repeat function) will have their inputs/outputs recorded and tested. NOTE:

• Doing GrugTest.default_record_io = [ "string_tools", "feature_branch_1" ] would run record both the repeat function AND the helper1_for_new_feature function.

Full API Options

from grug_test import GrugTest, grug_test
import math

# Defaults
GrugTest.default_max_io         = math.inf
GrugTest.force_fully_disable    = False # will override EVERYTHING and disable
GrugTest.default_record_io      = False  
GrugTest.default_run_tests      = False  
GrugTest.default_rewrite_inputs = False # (rarely needed)
GrugTest.force_record_io        = False # will override the local grug_test(should_record_io=False) 
GrugTest.force_run_tests        = False # will override the local grug_test(should_run_tests=False) 
GrugTest.force_rewrite_inputs   = False # will override the local grug_test(should_rewrite_inputs=False) 
GrugTest.verbose                = 1
GrugTest.overflow_strat         = "keep_old" # whenever the max_io limit is reached
# GrugTest.overflow_strats = [ 'keep_old', 'delete_random', ]

# Possbilities
GrugTest.default_max_io         = 30
GrugTest.force_fully_disable    = True 
GrugTest.default_run_tests      = [ "some_tag", "some_other_tag" ]
GrugTest.default_record_io      = [ "some_tag", "some_other_tag" ]
GrugTest.default_rewrite_inputs = [ "some_tag", "some_other_tag" ]  
GrugTest.force_run_tests        = [ "some_tag", "some_other_tag" ]
GrugTest.force_record_io        = [ "some_tag", "some_other_tag" ]
GrugTest.force_rewrite_inputs   = [ "some_tag", "some_other_tag" ]
GrugTest.verbose                = 0
GrugTest.overflow_strat         = "delete_random" # whenever the max_io limit is reached
     
@grug_test(
    max_io=None,                  # how many tests to save for this function (default=infinite)
    additional_io_per_run=1,      # if you want to gradually add tests (1 per run)
    save_to=None,                 # defaults to a local folder using the function name
    tags=["string_tools", "bla"], # tags are used to create groups of tests
                                  #     E.g. GrugTest.default_record_io = ["string_tools"] would record
                                  #     this function because its part of the group
    # for local debugging:
    should_record_io=None,      # when true, it will record_id uses GrugTest.default_record_io if not given
    should_run_tests=None,      # uses GrugTest.default_run_tests if not given
    should_rewrite_inputs=None, # uses GrugTest.default_rewrite_inputs if not given
    soft_skip=False,            # this is a quick-toggle for debugging, it overrides the other direct args
    func_name="repeat_string",  # override the actual function name (rarely needed)
                                # can be useful for comparing two functions
)
def some_function(arg):
    pass

Q&A

Does this work with @staticmethod?

• Yes but you have to put the decorator on the line BELOW @staticmethod

Does this work with ANY pure function?

• Almost, the arguments need to be seralizable. For example, if you pass a lambda function as an argument then grug_test can't really save/load that lambda function when replay_inputs=True. However, you can make almost any normal class seralizable, just checkout a tutorial on making a class work with python-pickle, or (even better) do from grug_test import yaml and make your class be yaml-seralizable (tutorial/example here)