E1920 Fix Code Climate Issues - Revision history

Ngtitov: /* Existing Rspec Testing */

2019-04-02T01:38:57Z

Existing Rspec Testing

← Older revision		Revision as of 01:38, 2 April 2019
Line 428:		Line 428:
	''However'', these accounted for all the models, while the models of our interest are named starting with ''Se'' through ''Z''. Therefore, we had to look at all 50 files individually, record their code coverage and calculate the average on all of them. Based on our results, the models our interest represented 1510 lines of code and 58.54% code coverage. Among our models, these numbers included 10 models that had 100% code coverage.		''However'', these accounted for all the models, while the models of our interest are named starting with ''Se'' through ''Z''. Therefore, we had to look at all 50 files individually, record their code coverage and calculate the average on all of them. Based on our results, the models our interest represented 1510 lines of code and 58.54% code coverage. Among our models, these numbers included 10 models that had 100% code coverage.

	~~However, while~~ working with the existing RSpec tests, we noticed that some models have relatively low coverage (> 40%). Therefore, we could not rely strictly on existing RSpec tests and guarantee correctness of our code changes by running them. We decided to write additional Rspec tests that directly test all our code changes and add them into existing spec files by modifying existing tests and boost the test coverage on files that we modified. For example, we fixed multiple Code Climate issues in the [https://github.com/gshugh/expertiza/commit/b9aac55b6eb331c1db295b7e7dd45fbbdba19abc app/models/team.rb] file. Specifically, complications of such issues as "Use a guard clause instead of wrapping the code inside a conditional expression", "Use ''find_by'' instead of ''where.first''", and others were resolved in ''import'', ''export'', and ''handle_duplicate'' method calls. Most of the changes are listed below:		While working with the existing RSpec tests, we noticed that some models have relatively low coverage (> 40%). Therefore, we could not rely strictly on existing RSpec tests and guarantee correctness of our code changes by running them. We decided to write additional Rspec tests that directly test all our code changes and add them into existing spec files by modifying existing tests and boost the test coverage on files that we modified. For example, we fixed multiple Code Climate issues in the [https://github.com/gshugh/expertiza/commit/b9aac55b6eb331c1db295b7e7dd45fbbdba19abc app/models/team.rb] file. Specifically, complications of such issues as "Use a guard clause instead of wrapping the code inside a conditional expression", "Use ''find_by'' instead of ''where.first''", and others were resolved in ''import'', ''export'', and ''handle_duplicate'' method calls. Most of the changes are listed below:

	[[File: app_models_teams_1.png]]		[[File: app_models_teams_1.png]]

Ngtitov: /* Existing Rspec Testing */

2019-04-02T01:37:40Z

Existing Rspec Testing

← Older revision		Revision as of 01:37, 2 April 2019
Line 422:		Line 422:
	817 examples, 0 failure, 16 pending		817 examples, 0 failure, 16 pending

	Hence, prior completion of our project, the code coverage on models was as follows:		Hence, prior completion of our project, the code coverage on all the models was as follows:

	[[File: code_coverage_before.png]]		[[File: code_coverage_before.png]]

	~~These tests~~ represented 1510 lines of code and 60.34% code coverage ~~for the models~~. Among our models, these numbers included 10 models that had 100% code coverage.		''However'', these accounted for all the models, while the models of our interest are named starting with ''Se'' through ''Z''. Therefore, we had to look at all 50 files individually, record their code coverage and calculate the average on all of them. Based on our results, the models our interest represented 1510 lines of code and 58.54% code coverage. Among our models, these numbers included 10 models that had 100% code coverage.

	However, while working with the existing RSpec tests, we noticed that some models have relatively low coverage (> 40%). Therefore, we could not rely strictly on existing RSpec tests and guarantee correctness of our code changes by running them. We decided to write additional Rspec tests that directly test all our code changes and add them into existing spec files by modifying existing tests and boost the test coverage on files that we modified. For example, we fixed multiple Code Climate issues in the [https://github.com/gshugh/expertiza/commit/b9aac55b6eb331c1db295b7e7dd45fbbdba19abc app/models/team.rb] file. Specifically, complications of such issues as "Use a guard clause instead of wrapping the code inside a conditional expression", "Use ''find_by'' instead of ''where.first''", and others were resolved in ''import'', ''export'', and ''handle_duplicate'' method calls. Most of the changes are listed below:		However, while working with the existing RSpec tests, we noticed that some models have relatively low coverage (> 40%). Therefore, we could not rely strictly on existing RSpec tests and guarantee correctness of our code changes by running them. We decided to write additional Rspec tests that directly test all our code changes and add them into existing spec files by modifying existing tests and boost the test coverage on files that we modified. For example, we fixed multiple Code Climate issues in the [https://github.com/gshugh/expertiza/commit/b9aac55b6eb331c1db295b7e7dd45fbbdba19abc app/models/team.rb] file. Specifically, complications of such issues as "Use a guard clause instead of wrapping the code inside a conditional expression", "Use ''find_by'' instead of ''where.first''", and others were resolved in ''import'', ''export'', and ''handle_duplicate'' method calls. Most of the changes are listed below:

Ngtitov: /* Conclusion */

2019-04-02T01:32:08Z

Conclusion

← Older revision		Revision as of 01:32, 2 April 2019
Line 501:		Line 501:
	Code Climate is a powerful tool that finds flaws in the source code of the Expertiza project. We would like to acknowledge that all issues reported by Code Climate on the files beginning with ''Se'' thru ''Z'' under ''app/models'' directory were fixed (except [http://wiki.expertiza.ncsu.edu/index.php?title=E1920_Fix_Code_Climate_Issues#Ignored_Code_Climate_Smells Ignored Code Climate Smells])! There are no more Code Climate issues found!		Code Climate is a powerful tool that finds flaws in the source code of the Expertiza project. We would like to acknowledge that all issues reported by Code Climate on the files beginning with ''Se'' thru ''Z'' under ''app/models'' directory were fixed (except [http://wiki.expertiza.ncsu.edu/index.php?title=E1920_Fix_Code_Climate_Issues#Ignored_Code_Climate_Smells Ignored Code Climate Smells])! There are no more Code Climate issues found!

	Although, in our project we were not asked to write additional Rspec tests, we dedicated some of our time to do so. The goal of writing additional spec test cases was to validate our code changes, make sure our changes did not cause any problems in the app, and boost the test coverage on the models we worked on. We achieved ~~all our goals~~ and ~~tested~~ all our changes with the help of Rspec tests that we wrote. We ~~also ensured~~ that ~~our additional spec tests boost~~ the code coverage.		Although, in our project we were not asked to write additional Rspec tests, we dedicated some of our time to do it. The goal of writing additional spec test cases was to validate our code changes, make sure our changes did not cause any problems in the app, and boost the test coverage on the models we worked on. All the goals were achieved and all our changes were successfully tested with the help of Rspec tests that we wrote.

			We concluded that code coverage increased to '''68.19%''' for the models of interest (models with names starting with ''Se'' through ''Z''). The number of lines of code decreased by 4.64% (from 1510 to 1440), a code coverage improvements is by 9.65% (from 58.54% to 68.19%) and an increase in test examples by 8.45% (from 817 to 886). Included in the new test numbers was a 70% increase in the number of models that had 100% coverage (from 10 to 17)!

	We encourage software developers working on the Expertiza project to maintain clean, readable, well-commented and efficient code. We suggest to achieve at least 85% of test coverage on all of the existing files by writing more Rspec tests and considering more possible scenarios (i.e, edge-case scenarios).		We encourage software developers working on the Expertiza project to maintain clean, readable, well-commented and efficient code. We suggest to achieve at least 85% of test coverage on all of the existing files by writing more Rspec tests and considering more possible scenarios (i.e, edge-case scenarios).

Ngtitov: /* New RSpec Testing */

2019-04-02T01:27:31Z

New RSpec Testing

← Older revision		Revision as of 01:27, 2 April 2019
Line 477:		Line 477:
	[[File: code_coverage_after.png]]		[[File: code_coverage_after.png]]

	However, the above results generated for all the models (not only for models with names starting with ''Se'' through ''Z''). We had to manually look at each file and determine the code coverage increase and other statistics.		''However'', the above results generated for all the models (not only for models with names starting with ''Se'' through ''Z''). We had to manually look at each file and determine the code coverage increase and other statistics.

	After looking at each model of our interest individually, we concluded that 1440 lines of code and '''68.19%''' of code coverage for the models of interest. The number of lines of code decreased by 4.64% (from 1510 to 1440), an increase in code coverage of 9.65% (from 58.54% to 68.19%) and an increase in test examples of 8.45% (from 817 to 886). Included in the new test numbers was a 70% increase in the number of models that had 100% coverage (from 10 to 17). The increase in the percentage of code coverage was due to a combination of the decrease in the number of lines and the increase in the number of test examples.		After looking at each model of our interest individually, we concluded that 1440 lines of code and '''68.19%''' of code coverage for the models of interest. The number of lines of code decreased by 4.64% (from 1510 to 1440), an increase in code coverage of 9.65% (from 58.54% to 68.19%) and an increase in test examples of 8.45% (from 817 to 886). Included in the new test numbers was a 70% increase in the number of models that had 100% coverage (from 10 to 17). The increase in the percentage of code coverage was due to a combination of the decrease in the number of lines and the increase in the number of test examples.

Ngtitov: /* New RSpec Testing */

2019-04-02T01:26:54Z

New RSpec Testing

← Older revision		Revision as of 01:26, 2 April 2019
Line 477:		Line 477:
	[[File: code_coverage_after.png]]		[[File: code_coverage_after.png]]

	~~This represented 1440 lines of code and 63.4% of code coverage~~ for the models~~. The number of lines of code decreased by 4.64%~~ (~~from 1510 to 1440~~)~~, an increase in code coverage of 3~~.~~06% (from 60.34%~~ to ~~63.4%)~~ and ~~an increase in test examples of 8.45% (from 817 to 886). Included in~~ the ~~new test numbers was a 70% increase in the number of models that had 100%~~ coverage ~~(from 10 to 17). The~~ increase ~~in the percentage of code coverage was due to a combination of the decrease in the number of lines~~ and ~~the increase in the number of test examples~~.		However, the above results generated for all the models (not only for models with names starting with ''Se'' through ''Z''). We had to manually look at each file and determine the code coverage increase and other statistics.

	The code coverage increase of models ( due to our additional Rspec tests on models with names starting with ''Se'' through ''Z'') was determined by running Rspec tests against the latest master branch and our development beta branch.		After looking at each model of our interest individually, we concluded that 1440 lines of code and '''68.19%''' of code coverage for the models of interest. The number of lines of code decreased by 4.64% (from 1510 to 1440), an increase in code coverage of 9.65% (from 58.54% to 68.19%) and an increase in test examples of 8.45% (from 817 to 886). Included in the new test numbers was a 70% increase in the number of models that had 100% coverage (from 10 to 17). The increase in the percentage of code coverage was due to a combination of the decrease in the number of lines and the increase in the number of test examples.

			The code coverage increase of models (due to our additional Rspec tests on models with names starting with ''Se'' through ''Z'') was determined by running Rspec tests against the latest master branch and our development beta branch.

	=== UI Testing ===		=== UI Testing ===

Ngtitov: /* New RSpec Testing */

2019-04-02T01:19:00Z

New RSpec Testing

← Older revision		Revision as of 01:19, 2 April 2019
Line 473:		Line 473:
	886 examples, 0 failure, 16 pending		886 examples, 0 failure, 16 pending

	~~This represented 1440 lines~~ of ~~code~~ and ~~68.19% of code coverage for~~ the ~~models. The number of lines of code decreased by 4.64% (from 1510 to 1440), an increase in~~ code coverage ~~of 16.48% (from 58.54% to 68.19%) and an increase in test examples of 8.45% (from 817 to 886). Included in the new test numbers was a 70%~~ increase in the ~~number of~~ models ~~that had 100% coverage (from 10 to 17). The increase in the percentage of code coverage was due to a combination of the decrease in the number of lines and the increase in the number of test examples.~~		After completion of our project, and adding newly developed Rspec tests, the we observed code coverage increase on the models as follows:

	The code coverage increase of models ( due to our additional Rspec tests on models with names starting with ''Se'' through ''Z'') was determined by running Rspec tests before and after of the project completion.		[[File: code_coverage_after.png]]

	~~Prior completion~~ of ~~our project, the~~ code coverage on models ~~was as follows:~~		This represented 1440 lines of code and 63.4% of code coverage for the models. The number of lines of code decreased by 4.64% (from 1510 to 1440), an increase in code coverage of 3.06% (from 60.34% to 63.4%) and an increase in test examples of 8.45% (from 817 to 886). Included in the new test numbers was a 70% increase in the number of models that had 100% coverage (from 10 to 17). The increase in the percentage of code coverage was due to a combination of the decrease in the number of lines and the increase in the number of test examples.

	~~[[File: code_coverage_before~~.~~png]]~~

	~~After completion~~ of ~~our project~~, the code coverage ~~on models is as follows:~~

	~~[[File: code_coverage_after~~.~~png]]~~

	~~Therefore, we achieved 16.48%~~ increase of ~~code coverage~~ on models by ~~adding new~~ Rspec tests.		The code coverage increase of models ( due to our additional Rspec tests on models with names starting with ''Se'' through ''Z'') was determined by running Rspec tests against the latest master branch and our development beta branch.

	=== UI Testing ===		=== UI Testing ===

Ngtitov: /* Existing Rspec Testing */

2019-04-02T01:11:54Z

Existing Rspec Testing

← Older revision		Revision as of 01:11, 2 April 2019
Line 418:		Line 418:
	$ bundle exec rspec spec/controllers spec/models		$ bundle exec rspec spec/controllers spec/models

	We ran ~~this~~ command against master branch as a baseline, which produced the following results:		We ran above command against master branch as a baseline, which produced the following results:

	817 examples, 0 failure, 16 pending		817 examples, 0 failure, 16 pending

	These tests represented 1510 lines of code and 58.54% code coverage for ~~our~~ models. Among our models, these numbers included 10 models that had 100% code coverage.		Hence, prior completion of our project, the code coverage on models was as follows:

			[[File: code_coverage_before.png]]

			These tests represented 1510 lines of code and 60.34% code coverage for the models. Among our models, these numbers included 10 models that had 100% code coverage.

	However, while working with the existing RSpec tests, we noticed that some models have relatively low coverage (> 40%). Therefore, we could not rely strictly on existing RSpec tests and guarantee correctness of our code changes by running them. We decided to write additional Rspec tests that directly test all our code changes and add them into existing spec files by modifying existing tests and boost the test coverage on files that we modified. For example, we fixed multiple Code Climate issues in the [https://github.com/gshugh/expertiza/commit/b9aac55b6eb331c1db295b7e7dd45fbbdba19abc app/models/team.rb] file. Specifically, complications of such issues as "Use a guard clause instead of wrapping the code inside a conditional expression", "Use ''find_by'' instead of ''where.first''", and others were resolved in ''import'', ''export'', and ''handle_duplicate'' method calls. Most of the changes are listed below:		However, while working with the existing RSpec tests, we noticed that some models have relatively low coverage (> 40%). Therefore, we could not rely strictly on existing RSpec tests and guarantee correctness of our code changes by running them. We decided to write additional Rspec tests that directly test all our code changes and add them into existing spec files by modifying existing tests and boost the test coverage on files that we modified. For example, we fixed multiple Code Climate issues in the [https://github.com/gshugh/expertiza/commit/b9aac55b6eb331c1db295b7e7dd45fbbdba19abc app/models/team.rb] file. Specifically, complications of such issues as "Use a guard clause instead of wrapping the code inside a conditional expression", "Use ''find_by'' instead of ''where.first''", and others were resolved in ''import'', ''export'', and ''handle_duplicate'' method calls. Most of the changes are listed below:
Line 429:		Line 433:
	[[File: app_models_teams_2.png]]		[[File: app_models_teams_2.png]]

	Any of the above changes may break existing functionality of the code without proper testing. Moreover, we noticed that existing ''spec/models/team_spec.rb'' Rspec file does not thoroughly tests these method calls and had very a little to none of such tests. Hence, in order to verify and test all our changes, we wrote additional Rspec tests and added them into ''spec/models/team_spec.rb'' that specifically tests our code changes. ''Note'' that in the following test cases, we cover all possible scenarios, including edge-case scenarios. Check out [https://github.com/gshugh/expertiza/commit/96cec7fa82c1c55e31e0196d3a025111a4fe134a spec/models/team_spec.rb] for full list of Rspec tests. The following is a list of Rspec we developed to test all our changes in the ''app/models/team.rb'' file:		Any of the above changes may break existing functionality of the code without proper testing. Moreover, we noticed that existing ''spec/models/team_spec.rb'' Rspec file does not thoroughly tests these method calls and had very a little to none of such tests.

			Hence, in order to verify and test all our changes, we wrote additional Rspec tests and added them into existing ''spec/models/team_spec.rb'' that now specifically tests our code changes. ''Note'' that in the following test cases, we cover all possible scenarios, including edge-case scenarios. Check out [https://github.com/gshugh/expertiza/commit/96cec7fa82c1c55e31e0196d3a025111a4fe134a spec/models/team_spec.rb] for full list of Rspec tests. The following is a list of Rspec we developed to test all our changes in the ''app/models/team.rb'' file:

	[[File: rspec_models_team_spec_1.png]]		[[File: rspec_models_team_spec_1.png]]

Ngtitov: /* New RSpec Testing */

2019-04-02T01:01:32Z

New RSpec Testing

← Older revision		Revision as of 01:01, 2 April 2019
Line 473:		Line 473:
	Prior completion of our project, the code coverage on models was as follows:		Prior completion of our project, the code coverage on models was as follows:

	[[File: ~~se_z_test_coverage_gui_before~~.png]]		[[File: code_coverage_before.png]]

	After completion of our project, the code coverage on models is as follows:		After completion of our project, the code coverage on models is as follows:

	[[File: ~~se_z_test_coverage_gui_after~~.png]]		[[File: code_coverage_after.png]]

	Therefore, we achieved 16.48% increase of code coverage on models by adding new Rspec tests.		Therefore, we achieved 16.48% increase of code coverage on models by adding new Rspec tests.

Ngtitov: /* New RSpec Testing */

2019-04-02T00:52:16Z

New RSpec Testing

← Older revision		Revision as of 00:52, 2 April 2019
Line 461:		Line 461:
	After all of the changes to the RSpec files, our beta branch produced the following results:		After all of the changes to the RSpec files, our beta branch produced the following results:

	$ bundle exec rspec spec/controllers spec/models ~~[20:26:45]~~		$ bundle exec rspec spec/controllers spec/models
	....................................................................		....................................................................
	Finished in 4 minutes 13.8 seconds (files took 11.09 seconds to load)		Finished in 4 minutes 13.8 seconds (files took 11.09 seconds to load)

Ngtitov: /* New RSpec Testing */

2019-04-02T00:51:31Z

New RSpec Testing

← Older revision		Revision as of 00:51, 2 April 2019
Line 459:		Line 459:
	Again this is only one example we show here of the files that we created. More additional spec files were created to test our code changes. For example, we made extensive changes by fixing "Tagging a string as html safe may be a security risk" Code Climate issue. Most of those models did not have corresponding Rspec files, so we added those RSpec files to allow developers to have some confidence in the code before deployment. The test coverage on those models was also increased.		Again this is only one example we show here of the files that we created. More additional spec files were created to test our code changes. For example, we made extensive changes by fixing "Tagging a string as html safe may be a security risk" Code Climate issue. Most of those models did not have corresponding Rspec files, so we added those RSpec files to allow developers to have some confidence in the code before deployment. The test coverage on those models was also increased.

	After all of ~~these~~ changes to the RSpec files, our beta branch produced the following results:		After all of the changes to the RSpec files, our beta branch produced the following results:

	886 examples, 0 failure, 16 pending		$ bundle exec rspec spec/controllers spec/models [20:26:45]
			....................................................................
			Finished in 4 minutes 13.8 seconds (files took 11.09 seconds to load)

			886 examples, 0 failure, 16 pending

	This represented 1440 lines of code and 68.19% of code coverage for the models ~~of interest~~. The number of lines of code decreased by 4.64% (from 1510 to 1440), an increase in code coverage of 16.48% (from 58.54% to 68.19%) and an increase in test examples of 8.45% (from 817 to 886). Included in the new test numbers was a 70% increase in the number of models that had 100% coverage (from 10 to 17). The increase in the percentage of code coverage was due to a combination of the decrease in the number of lines and the increase in the number of test examples.		This represented 1440 lines of code and 68.19% of code coverage for the models. The number of lines of code decreased by 4.64% (from 1510 to 1440), an increase in code coverage of 16.48% (from 58.54% to 68.19%) and an increase in test examples of 8.45% (from 817 to 886). Included in the new test numbers was a 70% increase in the number of models that had 100% coverage (from 10 to 17). The increase in the percentage of code coverage was due to a combination of the decrease in the number of lines and the increase in the number of test examples.

	The ~~ratio of~~ code coverage of models ~~of interests~~ (with names starting with ''Se'' through ''Z'') ~~to all the models and to the entire code base~~ was determined by running ~~the following command~~ before and after of the project completion:		The code coverage increase of models ( due to our additional Rspec tests on models with names starting with ''Se'' through ''Z'') was determined by running Rspec tests before and after of the project completion.

	~~rspec spec/models/[se-z]*.rb~~		Prior completion of our project, the code coverage on models was as follows:

	~~We determined that code coverage ratio of models we fixed / all models and models we fixed / entire code base is increased by 3.15% and 0.95% respectively.~~

	Prior completion of our project, ~~we executed above command and determined that ''Se'' through ''Z'' models to all~~ the ~~models ratio had only 37.1% of~~ code coverage~~; the same~~ models ~~of interest to the entire code base ratio had 14.65% code coverage~~ as ~~documented below~~:

	[[File: se_z_test_coverage_gui_before.png]]		[[File: se_z_test_coverage_gui_before.png]]

	After completion of our project, ~~using~~ the ~~same command, we determined~~ code coverage ~~ration is increased to 40.25%~~ on ~~the same~~ models ~~of interest and 15.6% of the total code respectively~~ as follows:		After completion of our project, the code coverage on models is as follows:

	[[File: se_z_test_coverage_gui_after.png]]		[[File: se_z_test_coverage_gui_after.png]]

	Therefore, we achieved 3.15% increase of code coverage ~~ratio to all the~~ models ~~and 0.95% increase of code coverage ratio to the entire code base~~.		Therefore, we achieved 16.48% increase of code coverage on models by adding new Rspec tests.

	=== UI Testing ===		=== UI Testing ===