fix(face-validator): reject images exceeding 8000x8000 resolution

[singhanurag0317-bit]

Fix: Missing image resolution boundary check in face quality check

Closes #811

Problem

The assess_face_quality endpoint loads user-provided images directly into OpenCV via cv2.imread without validating their resolution. An extremely large upload (pixel-bomb) could exhaust memory and crash the process.

Change

• Added check_image_resolution() in python-ai-service/app/services/face_validator.py that reads dimensions cheaply via PIL header metadata (Image.open(path).size) — no full pixel decode.
• validate_image_bytes() now rejects any image whose width or height exceeds MAX_DIMENSION = 8000 before the OpenCV decode step.
• New python-ai-service/app/services/test_face_validator_resolution.py covering normal-size accept plus over-limit rejection (width, height, and both).

Tests

• New pytest cases: test_accepts_normal_size_image, test_rejects_image_with_both_dimensions_over_limit, test_rejects_image_wider_than_limit, test_rejects_image_taller_than_limit.

Verification

• CI will run pytest on the python-ai-service.

[vercel]

@singhanurag0317-bit is attempting to deploy a commit to the souma9830's projects Team on Vercel.

A member of the Team first needs to authorize it.