Photo Watermark Generator

Problem

When photos need consistent overlays for time, location, and notes, manual editing often produces uneven results: font size may not scale with image width, long addresses can overflow, wrapped lines may misalign, and mask padding can vary. Photo Watermark Generator turns this into a fixed layout pipeline: after receiving a photo and watermark fields, it renders a reusable watermark image using OpenCV and Pillow.

How It Works

The default style uses a transparent background:
- The watermark appears in the bottom-left corner with a fixed 6px yellow vertical bar.
- Text is white, scaled to 4.2% of the photo width, helping different resolutions stay legible.
- The notes field is always rendered; text top aligns with the yellow bar.
- If a line exceeds the reference width, it wraps automatically. The reference line is 地点：XXXX省XXXX市XXXX区·XXXXXXX小区, and wrapped content indents to the first character after the label.

If the input says “add mask,” it switches to a semi-transparent black background with 18px rounded corners, 15px left/right padding, and 12px top/bottom padding.

Limits and Notes

This skill handles photo watermark layout rather than general OCR or mapping. Typical inputs include time, location, lat/lng, and notes; notes are always shown, while project is optional. Missing or malformed fields will affect the rendered text, and coordinates are treated as display strings rather than location-service inputs. It assumes a Python 3.6+ environment with opencv-python and Pillow.