Welcome to the official test runner for DoCoreAI — a framework that dynamically optimizes LLM responses based on intelligence parameters like reasoning, creativity, precision, and temperature.

This repo helps you evaluate the performance of intelligence_profiler — a single-step API endpoint from DoCoreAI — by comparing generated temperature values against expected ranges for diverse test prompts.

• Sends test prompts (user_content + role) to /intelligence_profiler.
• Parses the response.
• Extracts the predicted temperature.
• Validates it against the expected range.
• Prints a clean report with ✅ passes and ❌ failures.

.
├── run_tests.py          # Main test runner
├── test_cases.json       # Test prompts + expected temperature ranges
├── .env                  # API keys and model info (not committed)
└── ...

No need for local setup. You can run all tests directly inside GitHub Actions with just a few clicks.

✅ Just follow these simple steps:

1. Fork this repository (top-right corner → Fork).
2. Go to your fork → Settings → Secrets and variables → Actions.
3. Click New repository secret:

• Name: OPENAI_API_KEY
• Value: Your OpenAI secret key
• Then click Add secret.

4. Now, go to the Actions tab in your fork.
5. If prompted, click "I understand my workflows" → Enable workflows.
6. Under "All Workflows", click on "DoCoreAI PyPI Test" (our sample test workflow).
7. Click "Run workflow" → Choose main branch → Hit Run.
8. Wait a few seconds for the test to begin.
9. Once it turns green ✅, click into it → Run tests → View live execution of each test case.

1. Clone the Repo

git clone https://github.com/SajiJohnMiranda/DoCoreAI-Pulse.git
cd DoCoreAI-Pulse

2. Install Dependencies

pip install -r requirements.txt

3. Create a .env file

OPENAI_API_KEY=your-openai-key
GROQ_API_KEY=your-groq-key
MODEL_PROVIDER=openai
MODEL_NAME=gpt-3.5-turbo

4. Start the DoCoreAI Server

uvicorn api.main:app --reload

run in a new terminal window

python run_tests.py

Make sure the /intelligence_profiler endpoint is running locally at http://localhost:8000.

You'll see each test input, actual response, and whether the predicted temperature falls in the expected range.

Example output:

▶️ Running Test Case #1
✅ Passed: temperature=0.250 is within range (0.1, 0.3)

▶️ Running Test Case #2
❌ Failed: temperature=0.600 not in expected range (0.8, 1.0)

🔍 Final Report: 4 passed, 1 failed

Output video on Linkedin

💡 If you run this in GitHub Actions, the test will fail the workflow if any test fails. You can modify it to soft-fail with only a warning.

Edit test_cases.json to add your own inputs and expected temperature ranges:

[
  {
    "input": {
      "user_content": "Invent a brand-new type of sport.",
      "role": "Creative Thinker"
    },
    "expected_range": [0.9, 1.0]
  }
]

This runner is independent of the main DoCoreAI repo, as long as:

• The /intelligence_profiler API is running.

• Your .env has valid API keys.

• You have a valid test_cases.json.

Have an idea to improve the test coverage? Spot an issue? Feel free to:

• Open an issue

• Submit a PR

• Or suggest improvements to the test logic!

Creative Commons Attribution-NonCommercial (CC BY-NC 4.0) — do what you want, just don’t forget to give credit ✨