Open mgaber2024 opened 2 months ago
@mgaber2024 Thanks for filing the issue. PHP is a great template to add to the arsenal. We'll get on that ASAP.
In the meantime, lower performing LLMs like llama3-8b
won't cut it. You'll need to try something higher end such at gpt-4o
. Have you tried gpt-4o
yet with Cover Agent?
The tests are shown in the console output when running the Codium coverage agent with the following command, but no tests are written to the specified file. it repeats the same outputs over and over until I got this " Reached maximum iteration limit without achieving desired coverage. Current Coverage: 0.0% " I got the same result even when I lower the desired-coverage value
Command used:
export GROQ_API_KEY="api-key" ./cover-agent-ubuntu \ --source-file-path "./app/Http/Controllers/LogoutController.php" \ --test-file-path "./tests/Feature/LogoutTest.php" \ --code-coverage-report-path "./coverage.xml" \ --test-command "./vendor/bin/pest --coverage --coverage-clover=coverage.xml" \ --test-command-dir "." \ --coverage-type "cobertura" \ --desired-coverage 20 \ --max-iterations 3 \ --model "groq/llama3-8b-8192" \ --api-base "https://api.groq.com/openai/v1/chat/completions"
Expected Behavior: The tests should be written to the specified test file (LogoutTest.php) as indicated in the command.
Actual Behavior: The tests appear to run successfully, and coverage is reported in the console output, but no changes are made to the test file (LogoutTest.php). It remains unchanged after execution.
Steps to Reproduce:
Run the above command with a valid API key and file paths. Observe the test output in the console. Check the test file (LogoutTest.php), which should have been modified or written with tests but is not.
Environment:
OS: Ubuntu Test framework: Pest Model: groq/llama3-8b-8192 API base: https://api.groq.com/openai/v1/chat/completions Coverage type: Cobertura
Additional Information:
I am using Pest for testing in a Laravel project. The console shows successful test output, but no tests are written to the file as expected. Please look into this issue. Let me know if you need additional information or logs.
Thank you!