docs: explain custom analyzer Docker language config

[ynachiket]

Change

Adds a focused installation docs section for customizing the analyzer Docker image when deploying additional NLP languages. The new section covers:

• which analyzer/NLP/recognizer YAML files map to Docker build arguments
• a minimal Spanish spaCy configuration example
• building the image so referenced NLP models are installed
• runtime config mounting when models already exist in the image
• common pitfalls around supported_languages, NLP recognizer warnings, missing model downloads, and memory growth

Fixes #1663.

Validation

• ruby -ryaml ... parsed the new YAML snippet and checked the Spanish model and recognizer language lists
• rg content check for the new Docker language guidance, build arguments, Spanish model, and health/recognizer endpoints
• local link target check for the referenced analyzer configuration docs
• git diff --check

I did not run the full MkDocs build because the docs dependency set is not installed in this local checkout; this is a single-page Markdown documentation update.

[Copilot]

Pull request overview

Adds Docker-focused installation guidance for customizing Presidio Analyzer images to support additional NLP languages, addressing the linked request for clearer YAML and build instructions.

Changes:

• Documents analyzer Docker build arguments for analyzer, NLP, and recognizer registry configuration.
• Adds a minimal Spanish spaCy combined-configuration example and custom image build/run commands.
• Lists common pitfalls around language alignment, NLP recognizers, missing model downloads, and runtime validation endpoints.