[Text Extractor] Add a headless OCR CLI

Description of the new feature / enhancement

Add a supported, headless command-line interface for Text Extractor OCR.

The CLI should support OCR from an image file and, where feasible, a specified screen region. Recognized text should be written to stdout by default, with options for clipboard, text-file, and machine-readable output.

Example shape:

powertoys text-extractor image screenshot.png
powertoys text-extractor region --x 100 --y 100 --width 800 --height 600 --language en-US

Scenario when this would be used?

• Extract text from screenshots in scripts, document-processing workflows, and accessibility tools.
• Let another application or agent invoke Windows OCR without opening an interactive selection overlay.
• Batch-process image files and redirect recognized text into other terminal commands.

Supporting information

This should expose Text Extractor's OCR capability, not only launch its current interactive UI. It should provide predictable stdout/stderr behavior, documented exit codes, language selection, and follow doc/devdocs/cli-conventions.md.

Read doc/devdocs/cli-conventions.md first, then inspect the existing Text Extractor OCR entry point and how its interactive UI is launched. Define the CLI scope around image files and feasible screen regions, with stdout, clipboard, text-file, machine-readable output, language selection, documented exit codes, and predictable stdout/stderr behavior.