Great Assertions
This library is inspired by the Great Expectations library. The library has made the various expectations found in Great Expectations available when using the inbuilt python unittest assertions.
Install
pip install great-assertionsCode example Pandas
from great_assertions import GreatAssertions
import pandas as pd
class GreatAssertionTests(GreatAssertions):
def test_expect_table_row_count_to_equal(self):
df = pd.DataFrame({"col_1": [100, 200, 300], "col_2": [10, 20, 30]})
self.expect_table_row_count_to_equal(df, 3)Code example PySpark
from great_assertions import GreatAssertions
from pyspark.sql import SparkSession
class GreatAssertionTests(GreatAssertions):
def setUp(self):
self.spark = SparkSession.builder.getOrCreate()
def test_expect_table_row_count_to_equal(self):
df = self.spark.createDataFrame(
[
{"col_1": 100, "col_2": 10},
{"col_1": 200, "col_2": 20},
{"col_1": 300, "col_2": 30},
]
)
self.expect_table_row_count_to_equal(df, 3)List of available assertions
Pandas |
PySpark |
|
|---|---|---|
expect_table_row_count_to_equal |
|
|
expect_table_row_count_to_be_greater_than |
|
|
expect_table_row_count_to_be_less_than |
|
|
expect_table_has_no_duplicate_rows |
|
|
expect_column_value_to_equal |
|
|
expect_column_values_to_be_between |
|
|
expect_column_values_to_match_regex |
|
|
expect_column_values_to_be_in_set |
|
|
expect_column_values_to_be_of_type |
|
|
expect_table_columns_to_match_ordered_list |
|
|
expect_table_columns_to_match_set |
|
|
expect_date_range_to_be_more_than |
|
|
expect_date_range_to_be_less_than |
|
|
expect_date_range_to_be_between |
|
|
expect_column_mean_to_be_between |
|
|
expect_column_value_counts_percent_to_be_between |
|
|
expect_frame_equal |
|
|
expect_column_has_no_duplicate_rows |
|
|
expect_column_value_to_equal_if |
|
|
expect_column_value_to_be_greater_if |
|
|
Assertion Descriptions
For a description of the assertions see Assertion Definitions
Running the tests
Executing the tests still require unittest, the following options have been tested with the examples provided.
Option 1
import unittest
suite = unittest.TestLoader().loadTestsFromTestCase(GreatAssertionTests)
runner = unittest.TextTestRunner(verbosity=2)
runner.run(suite)Options 2
if __name__ == '__main__':
unittest.main()Pie Charts and Tables
For a more visual representation of the results, when using in Databricks or Jupyter Notebooks. The results can be outputted as tables or pie-chart.
import unittest
from great_assertions import GreatAssertionResult, GreatAssertions
class DisplayTest(GreatAssertions):
def test_pass1(self):
assert True is True
def test_fail(self):
assert "Hello" == "World"
suite = unittest.TestLoader().loadTestsFromTestCase(DisplayTest)
test_runner = unittest.runner.TextTestRunner(resultclass = GreatAssertionResult)
result = test_runner.run(suite)
result.to_barh() #Also available: result.to_pie()
result.to_results_table()
result.to_full_results_table()
Runnng with XML-Runner
To run with xml-runner, there is no difference to how it’s currently used. However you will not be able to get method like to_results_table as these use a different resultclass
import xmlrunner
suite = unittest.TestLoader().loadTestsFromTestCase(DisplayTest)
test_runner = xmlrunner.XMLRunner(output="test-results")
test_runner.run(suite)Production Monitoring
The assertions provided by GA will also allow the validation of the any environment including Production. Currently GA only supports saving the results to Spark, for example databricks.
Once the run has completed there is a save method, as seen below.
import xmlrunner
suite = unittest.TestLoader().loadTestsFromTestCase(DisplayTest)
test_runner = xmlrunner.XMLRunner(output="test-results")
result = test_runner.run(suite)
result.save(format="databricks")The image below shows a simple graph of the accumulation of tests over test run. However much more complex analysis can be performed with the extended data being generated by GA.
The extended table of results contains the following:
run_id |
timestamp |
method |
information |
test_id |
status |
extended |
20211222093029 |
2021-12-22 09:30:29 |
test_fail8 |
Traceback (most recent call last… |
13 |
Fail |
{“id”: 13, “name”: “expect_date_range_to_be_less_than”, “values”: {“expected_max_date”: “2019-05-13”, “actual_max_date”: “2019-05-13”}} |
20211222093029 |
2021-12-22 09:30:29 |
test_fail9 |
Traceback (most recent call last… |
14 |
Fail |
{“id”: 14, “name”: “expect_date_range_to_be_more_than”, “values”: {“expected_min_date”: “2015-10-01”, “actual_min_date”: “2015-10-01”}} |
From the extended column you can get further details about the type test, which was executed and the results. For example if we look at the test expect_table_row_count_to_be_less_than we should assert that the max row should not be breached.
In the code below, the expected was 100 and the actual was 205, which caused the test to fail. Therefore Analysts can query the extended data to get a picture of the size of the breach.
extended = {
"id": 2,
"name": expect_table_row_count_to_be_less_than,
"values": {
"exp_max_count": 100,
"act_count": 205,
},
}In production monitoring these types of results can allow the prevention of skewed results. For example, if you had a result, where the expected values were withing a range of 0-100 and you got an exceptionally large value.
The large value could cause business functionality to be skewed such that a defect could causes damage or loss of income or incorrect reporting to a downstream system.
Therefore, GA will allow you to provide benchmarks to the production validation and an experienced analyst can create reports on top of the data.
An example of the extended dataset:
Notes
If you get an arrows function warning when running in Databricks, this will happen because a toPandas() method is being used for many of the assertions. The plan is to remove Pandas conversion for pure PySpark code. If this is an issue, please raise an issue so this method can be prioritised. For now, it’s advisable to make sure the datasets are not too big, which cause the driver to crash.
Development
To create a development environment, create a virtualenv and make a development installation
virtualenv ve source ve/bin/activate
To run tests, just use pytest
(ve) pytest
2 maintainers
Metadata
2 maintainers
Release files for great-assertions 0.0.76
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| great_assertions-0.0.76.tar.gz | 24.5 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| great_assertions-0.0.76-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 43.0 kB
Release files / great_assertions-0.0.76.tar.gz
| Download URL | great_assertions-0.0.76.tar.gz |
|---|---|
| Size | 24.5 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
64f49ed96bc7446a7c5764922c342e246052ae2c156941c38820463111c1ddcd
|
|
BLAKE2b-256 checksum How to use checksums |
5aee3cee59d2d3a9b645dcd42c5a8ead7dacc08d416dd2ead324690c32065007
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.14.2
|
Release files / great_assertions-0.0.76-py3-none-any.whl
| Download URL | great_assertions-0.0.76-py3-none-any.whl |
|---|---|
| Size | 18.5 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
9f2fa6bc430864e3678314febd15626973b4ab7910d9104f1cd658eab59ed96f
|
|
BLAKE2b-256 checksum How to use checksums |
583ca0dcaa90d147d51caaa80f66064b8bce9b8cd1c734de8ff7525d1fd3abde
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.14.2
|
