logseq-table-formatter

Logseq table formatter

Use when a table lives inside an LFM bullet tree (pages/, journals/) and must render in Logseq and read cleanly in the raw file.

Default path

1. Confirm the table should stay a table (not a bullet label/value list). If more than two columns and the content is not a compact comparison, consider bullets per skill logseq-convert-md-to-lfm.
2. Build row data (header + separator + body). Prefer short cell text; put detail in child bullets under the table.
3. Use wikilinks in cells when a page exists ([[mise/unuse]], not `mise unuse`). Run skill logseq-link-hygiene when unsure a page exists.
4. Format with aligned columns:
   • First row: - | col1 | col2 | … | at the correct tab depth.
   • Following rows: same tab depth, two spaces instead of - , then | … | so every | shares the same column as row 1.
   • Pad cell text with spaces so pipes line up across all rows (verify in plain text).
5. Prefer the formatter script for padding; hand-edit only small tables.

Script

From repo root:

python3 .rulesync/skills/logseq-table-formatter/scripts/format-table.py \
  --tabs 2 \
  --rows='Command|Primary effect|Cache' \
  --rows='----|----|----' \
  --rows='[[mise/unuse]]|remove config and binary|✓'

• --tabs N — tab count before the table (default 2).
• --rows='a|b|c' — one row per flag (use = so separator rows like ----|---- are not parsed as CLI flags). First row is emitted with - |, rest with aligned |.
• Or pipe one row per line on stdin: printf '%s\n' 'H|H' '---|---' 'a|b' | python3 …/format-table.py --tabs 2

Trust script output for column widths; fix wikilinks and wording before running.

When to read references

• references/lfm-table-layout.md — LFM nesting, garden examples, wikilink vs backticks, checkmarks, edge cases.

Related

• rule: logseq-core — bullets, tabs, no blank lines between blocks (advanced detail: skill logseq-lfm).
• skill: logseq-convert-md-to-lfm — when not to keep a table.
• skill: logseq-link-hygiene — resolve [[wikilinks]] before filing.