Closes #763

What

Adds a plain text (.txt) output format alongside the existing PDF, Typst, Markdown, HTML and PNG outputs, as proposed in #763.

• New generate_text() renderer (renderer/text.py), same structure as the Markdown renderer
• New Jinja template set under templater/templates/text/ (header, section begin/end, one template per entry type)
• markdown_to_plain_text() in the Markdown parser: emphasis and code markers stripped, links rendered as label (url) when the URL adds information over the link text
• Plain text connection formatting (compute_connections_for_text)
• text_path and dont_generate_text settings, --text-path / --dont-generate-text CLI options, wired through the model builder
• Entry templates now receive entry_index and entries_count, so numbered entries can carry real numbers in formats that have no renderer doing the numbering; existing templates do not use them, so there is no behavior change
• Docs updated (user guide index and settings page), JSON schema and example YAMLs regenerated

The regenerated binary artifacts (example PDFs, PNGs, Typst files) are intentionally left out to keep the diff reviewable; the generated-files tests only require the example YAMLs and schema.json to be current.

User template overrides work exactly as for the other formats: placing a text/ directory next to the input file allows full restyling of the txt output (80-column layouts, ASCII tables, custom headers) without touching the core.

Test plan

Ran the full test suite and regenerated the reference files with just update-testdata, which produced the new tests/renderer/testdata/test_text/ references for the minimal and full models. Added unit tests for markdown_to_plain_text(). Also rendered a complete real-world CV to text and reviewed the output side by side with the Markdown and PDF versions.