| from env.environment import CodeGuardianEnv |
| from env.models import Action |
|
|
| env = CodeGuardianEnv() |
| obs = env.reset() |
|
|
| done = False |
|
|
| step_cycle = ["detect", "risk", "fix"] |
| step_index = 0 |
|
|
| while not done: |
| print("\n--- NEW STEP ---") |
| print("Code:", obs.code) |
|
|
| current_action = step_cycle[step_index] |
|
|
| if current_action == "detect": |
| action = Action( |
| action_type="detect", |
| detected_issues=["wrong operator"], |
| risk_level=None, |
| suggested_fix=None |
| ) |
|
|
| elif current_action == "risk": |
| action = Action( |
| action_type="risk", |
| detected_issues=None, |
| risk_level="low", |
| suggested_fix=None |
| ) |
|
|
| elif current_action == "fix": |
| action = Action( |
| action_type="fix", |
| detected_issues=None, |
| risk_level=None, |
| suggested_fix="return a + b" |
| ) |
|
|
| obs, reward, done, info = env.step(action) |
|
|
| print("Action:", current_action) |
| print("Reward:", reward) |
| print("Info:", info) |
|
|
| step_index = (step_index + 1) % 3 |
