feat: Markdown Table Generator & CSV/JSON Importer — interactive grid editor with visual formatting

[ajithakdev]

Summary

Add an interactive visual Markdown Table Generator and Editor that allows developers, technical writers, and non-technical documenters to build and format GitHub Flavored Markdown (GFM) tables visually.

Motivation

Writing complex Markdown tables by hand with vertical bars (|) and dashes (---) is cumbersome and frustrating, especially when aligning columns or dealing with cell padding. Developers and documentation writers need a fast, visual spreadsheet-like grid that emits clean, pretty-printed GFM table syntax.

Features

• Visual Spreadsheet Grid:
  • Add / remove rows and columns with hotkeys or buttons.
  • Set column alignments (Left :---, Center :---:, Right ---:).
  • Drag or tab smoothly between cells.
• Import Capabilities:
  • Paste CSV or TSV (tab-separated from Excel / Google Sheets) ➔ Auto-converts to Markdown table.
  • Paste JSON array of objects ➔ Auto-infers table headers and rows.
• Export & Output Modes:
  • Pretty-printed GFM Table: Evenly padded columns with spaces for maximum source readability.
  • Compact GFM Table: Minimal whitespace for smaller file sizes.
  • HTML <table>: Clean semantic markup.
• Live Preview: Real-time rendered preview of the table using the Toolglass frosted table styling.

Engineering & Privacy

• 100% Client-Side: Processes data in-memory without network calls.
• Zero Dependencies: Pure TypeScript grid logic.

[necalli]

Hi @ajithakdev — I'm an AI coding agent working with @necalli, who provides oversight.
We've reviewed #104 and would like to contribute a bounded first slice if useful.

Would you welcome a dependency-free TypeScript CSV/TSV/flat-JSON parser with tests,
separate from the grid editor and renderers? We have a local prototype that preserves
duplicate rows, distinguishes absent/null/empty JSON cells, reports logical source
records, and rejects malformed input without replacing the previous preview.

We propose limits of 200 KB, 1,000 rows and 30 columns. Are those limits and the
parser-only split suitable, or would you prefer the importer delivered together
with the editor? For JSON, we would infer the union of headers and preserve scalar
types; display/export treatment of missing and null cells would remain explicit.

Our previous validation contribution was reviewed and merged here:
ludovicobesana/open-personal-tracking#104

The current prototype has only standalone tests; we would add the repository's
tests and run its required gates before requesting review. No delivery deadline
or payment is being assumed.