Refactor and minor fixes

[panyuenlau]

• Change handler name from "check_url" to "check_prompt" so it matches what the handler is doing
• Break down the functionalities within the handler to 4 different utility functions:
  • check_url_status()
  • check_openai_prompt_status()
  • check_gemini_prompt_status()
  • check_gemini_prompt_status()
• Fix minor bug that results in having Threats: D, L, P, , D, i, s, a, b, l, e, d in the response when umbrella is disabled

Example output with the changes in this PR:

• When DLP is enabled and there are threats:

  {'prompt_injection': {'openai': 'OpenAI Prompt Injection Detection disabled', 'gemini': 'Gemini Prompt Injection Detection disabled'}, 'url_verdict': 'No Malware URL(s)', 'threats': 'Medical Prescriptions (US), Email Address'}

• When DLP is disabled or no threats:

  {'prompt_injection': {'openai': 'OpenAI Prompt Injection Detection disabled', 'gemini': 'Gemini Prompt Injection Detection disabled'}, 'url_verdict': 'Malware URL(s)', 'threats': 'No Threats Detected or DLP Disabled'}

[vhosakot]

great fixes @panyuenlau! @BenderScript, lmk when you bump version of this repo so I'll pull it in and refresh internal version, thanks