add llm.txt

Author: dakixr

llm.txt

@@ -0,0 +1,30 @@
+Using xpyxl with xlsxwriter (external agents)
+- Install: dependency declared as `xlsxwriter>=3.1.0`. Import `import xpyxl as x`.
+- Build: compose nodes (no coordinates) then save with the engine you want: `x.workbook()[x.sheet("Demo")[...]].save("out.xlsx", engine="xlsxwriter")`.
+- Primitives: `x.cell[...]`, `x.row[...]`, `x.col[...]`, `x.table()[...]`, `x.vstack(..., gap=0, style=[...])`, `x.hstack(..., gap=0, style=[...])`, `x.space(rows=1, height=None)`, `x.sheet(name, background_color=None)`, `x.workbook()[...]`.
+- Table inputs: sequence of sequences, sequence of dicts (header from keys), or dict-of-lists (columns zipped). `header_style=[...]`, `style=[...]`, `column_order=[...]`.
+- Saving: engine is chosen only at save time; both engines yield equivalent output. Use `engine="xlsxwriter"` when you want speed/low memory (as in README benchmarks).
+- Backgrounds: `sheet(..., background_color="#F8FAFC")` paints the grid (render fills a large area since xlsxwriter lacks sheet fill).
+
+Precomputed styles (combine in `style=[...]`)
+- Typography: `text_xs, text_sm, text_base, text_lg, text_xl, text_2xl, text_3xl, bold, italic, mono`.
+- Text colors: `text_red, text_green, text_blue, text_orange, text_purple, text_black, text_gray, text_muted, text_primary, text_white`.
+- Backgrounds: `bg_red, bg_primary, bg_muted, bg_success, bg_warning, bg_info`.
+- Alignment & layout: `text_left, text_center, text_right, align_top, align_middle, align_bottom, wrap, nowrap, wrap_shrink, allow_overflow, row_height(n), row_width(n)`.
+- Borders: `border_all, border_top, border_bottom, border_left, border_right, border_x, border_y, border_thin, border_medium, border_thick, border_dashed, border_dotted, border_double, border_none` plus color helpers `border_red/green/blue/orange/purple/black/gray/white/muted/primary`.
+- Tables: `table_bordered`, `table_banded`, `table_compact`.
+- Number/date formats: `number_comma`, `number_precision`, `percent`, `currency_usd`, `currency_eur`, `date_short`, `datetime_short`, `time_short`.
+
+Common patterns
+- Simple sheet: `x.sheet("Summary")[x.row(style=[x.text_2xl, x.bold])["Title"], x.row()["A", 1]]`.
+- Styled table: `x.table(header_style=[x.text_sm, x.text_gray], style=[x.table_bordered, x.table_compact])[{"Region": "EMEA", "Units": 1200}, ...]`.
+- Layouts: `x.vstack(...)` for vertical stacking with optional `gap`; `x.hstack(...)` to place blocks side by side; sprinkle `x.space()` for padding.
+- Multi-sheet: collect `SheetNode`s (each from `x.sheet(...)[...]`) into `x.workbook()[sheet1, sheet2]`.
+
+Real-world examples (ready to reuse)
+- Sales dashboard:
+  `x.workbook()[x.sheet("Sales", background_color="#F8FAFC")[x.row(style=[x.text_2xl, x.bold, x.text_blue])["Q3 Overview"], x.space(), x.table(header_style=[x.text_sm, x.text_gray], style=[x.table_bordered, x.table_banded, x.table_compact])[{"Region": "EMEA", "Units": 1200, "Price": 19.0}, {"Region": "APAC", "Units": 900, "Price": 21.0}, {"Region": "AMER", "Units": 1500, "Price": 18.5}], x.space(), x.hstack(x.row(style=[x.text_sm, x.text_gray])["Generated with xpyxl"], x.cell(style=[x.text_sm, x.text_gray])["Engine: xlsxwriter"], gap=2)]]`. Save with `engine="xlsxwriter"`.
+- Two-column layout:
+  `x.workbook()[x.sheet("Metrics")[x.hstack(x.vstack(x.row(style=[x.text_lg, x.bold])["KPI"], x.row()["MRR", 125000], x.row()["Churn", "2.1%"], gap=1, style=[x.border_all, x.bg_muted]), x.vstack(x.row(style=[x.text_lg, x.bold])["Notes"], x.row(style=[x.wrap])["Performance improved; keep monitoring churn drivers across regions."], gap=1, style=[x.border_all, x.bg_info]), gap=2)]]`. Save with `engine="xlsxwriter"`.
+- Minimal CSV-like export:
+  `x.workbook()[x.sheet("Data")[x.table(style=[x.table_compact])[{"id": 1, "name": "alpha"}, {"id": 2, "name": "beta"}]]]`. Save with `engine="xlsxwriter"` for faster large writes.