Skip to main content

​
Documentation project instructions

​
About this project

  • Documentation for Jetxl, a fast Excel writer for Python built in Rust
  • Built on Mintlify. Pages are MDX with YAML frontmatter; configuration lives in docs.json
  • Use the Mintlify MCP server, https://mcp.mintlify.com, to edit content and settings
  • Use the Mintlify docs MCP server, https://www.mintlify.com/docs/mcp, to query Mintlify usage

​
Sources of truth

Ranked, most reliable first:
  1. Running the library. pip install jetxl polars pyarrow openpyxl, write a file, read it back with openpyxl. This is the only way to catch behavior that contradicts both the stub and the source comments.
  2. The Rust source in the Jetxl repository.
  3. jetxl.pyi. Signatures only. Treat defaults and requiredness as unverified.
Files that matter most:

​
Terminology

  • “Jetxl” capitalized in prose, jetxl in code
  • Examples import as import jetxl as jet
  • “Arrow functions” means those ending in _arrow; “dictionary functions” means write_sheet and write_sheets
  • “Bytes functions” means those ending in _to_bytes
  • Say “spreadsheet” or “workbook”, not “Excel file”, except where Excel’s own behavior is the subject

​
Style preferences

Follow the Google developer documentation style guide.
  • Active voice, second person
  • Sentence case headings, one idea per sentence
  • American spelling
  • Bold for UI elements: choose Format Cells
  • Code formatting for file names, commands, paths, parameters, and values
  • No manual H1 — Mintlify generates it from the title frontmatter
  • Avoid “simply”, “just”, “please”, “in order to”, “e.g.”, “i.e.”, “via”, and “allows you to”

​
Content boundaries

Document silent failures explicitly. These cost users the most time and none of them raises:
  • Unusable colors are dropped and the file still writes
  • NaN and infinity become empty cells
  • Control characters are stripped from text
  • An unrecognized conditional-format operator falls back to greater_than
  • An unrecognized rule_type drops the whole rule
  • A table missing a required key is dropped
Never invent example output, benchmark numbers, or API behavior. If a number can’t be traced to the repository, a published benchmark, or a run you performed, leave it out. Performance figures come from the project’s own benchmarks and should be attributed as such. Keep the experimental status visible. Jetxl is pre-1.0 and warns that combining formatting parameters may hit bugs.

​
Audience

Two groups, served by content type rather than by page:
  • Guides and recipes for analysts who want a formatted spreadsheet. Lead with the outcome, put mechanics lower down.
  • Reference for developers who know what they’re looking for. Dense and complete, no tutorial framing.