Refactor command execution and buffer management.
- `CommitFormula` now defaults to targeting `_Measure` .
- `CommitFormula` no longer automatically clears the buffer; buffer
clearing is now handled by keymap sequences.
- Added `ClearBufferCmd` to the command registry.
- Updated `AddFormulaCmd` to support optional target category.
- Added `SetBuffer` effect to allow clearing buffers.
Co-Authored-By: fiddlerwoaroof/git-smart-commit (unsloth/gemma-4-26B-A4B-it-GGUF:UD-Q5_K_XL)
Implement keymap inheritance and sequence bindings.
- Added `parent` field to `Keymap` to support Emacs-style inheritance.
- Implemented `lookup` in `Keymap` to fall through to parent keymaps.
- Added `bind_seq` to allow multiple commands to be bound to a single key
pattern.
- Refactored existing keymaps to use sequences for common patterns like
Esc/Enter/Tab to clear buffers and change modes.
Co-Authored-By: fiddlerwoaroof/git-smart-commit (unsloth/gemma-4-26B-A4B-it-GGUF:UD-Q5_K_XL)
Refactor formula handling to use a virtual '_Measure' category.
- Formulas now default to targeting '_Measure' if no category is specified.
- '_Measure' items are dynamically computed from existing data items and
formula targets, preventing redundant item storage.
- Updated persistence to handle '_Measure' formulas and items correctly in
Markdown.
- Updated UI and model to use 'effective_item_names' for layout and item
resolution.
- Updated tests to reflect the new '_Measure' behavior.
Co-Authored-By: fiddlerwoaroof/git-smart-commit (unsloth/gemma-4-26B-A4B-it-GGUF:UD-Q5_K_XL)
Replace "Measure" with "Foo" as target_category in formula parser
tests so they don't depend on a specific category name convention.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Executed-By: spot
- Document PEG grammar as single source of truth for .improv format
- Update file format section with v2025-04-09 syntax: version line,
Initial View, pipe quoting, Views→Formulas→Categories→Data order
- Add pipe quoting convention and grammar-driven testing principles
- Update file inventory (persistence: 124+2291 lines, 83 tests)
- Add pest/pest_meta to dependency table
- Update persistence testing guidance for grammar-walking generator
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Executed-By: spot
- Add improv.pest PEG grammar as the single source of truth for the
.improv file format (v2025-04-09)
- Replace hand-written line scanner with pest-derived parser that walks
the grammar's parse tree
- Add grammar-walking test generator that reads improv.pest at test time
via pest_meta and produces random valid files from the AST
- Fix 6 parser bugs: newlines in text, commas in names, brackets in
names, float precision, view name ambiguity, group brackets
- New format: version line, Initial View header, pipe quoting (|...|),
Views→Formulas→Categories→Data section order, comma-separated items
- Bare names restricted to [A-Za-z_][A-Za-z0-9_-]*, everything else
pipe-quoted with \| \\ \n escapes
- Remove all unwrap() calls from production code, propagate errors
with Result throughout parse_md
- Extract shared escape_pipe/unescape_pipe/pipe_quote helpers, deduplicate
hidden/collapsed formatting, add w!() macro for infallible writeln
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Executed-By: spot
Move tests from the monolithic tests.rs into #[cfg(test)] mod tests
blocks in each command module. Shared test helpers live in
mod.rs::test_helpers.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Executed-By: fido
Add VirtualMeasure to CategoryKind description, document _Measure as
third always-present virtual category, add section on fixed-point
formula evaluation strategy.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
_Measure is now a virtual category (CategoryKind::VirtualMeasure),
created automatically in Model::new() alongside _Index and _Dim.
Formula targets are added as items automatically by add_formula.
Formula evaluation uses a fixed-point cache: recompute_formulas()
iterates evaluation of all formula cells until values stabilize,
resolving refs through the cache for formula values and raw data
aggregation for non-formula values. This fixes formulas that
reference other measures when hidden dimensions are present.
evaluate_aggregated now checks the formula cache instead of
recursively evaluating formulas, breaking the dependency between
formula evaluation and aggregation.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Model::add_formula now ensures the formula target exists as an item in
the target category (same fix as AddFormula effect, but at the model
level — covers file loading and headless paths).
Formula panel now shows raw formula text instead of debug-formatted
string with redundant target name and quotes.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
When there are more tiles than fit in the available width, the tile
bar now auto-scrolls to ensure the selected tile is always visible.
Overflow indicators (◀ ▶) show when tiles exist beyond the visible
area. Scroll offset is computed fresh each frame from tile_cat_idx.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Circular or self-referencing formulas now return CellValue::Error
instead of stack overflowing. eval_expr uses Result<f64, String>
internally so errors (circular refs, div/0, missing refs) propagate
immediately through the expression tree via ?. The depth limit (16)
is checked per evaluate_depth call — normal 1-2 level chains are
unaffected.
Also adds CellValue::Error variant for displaying ERR:reason in the
grid, and handles it in format, persistence, and search.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
When adding a formula interactively, the target (e.g. "Margin") was
registered as a formula but never added as an item to the target
category. The grid layout never created cells for it, making the
formula invisible. Now AddFormula::apply calls add_item before
registering the formula.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Four recorded casts at 100x30 with 2s idle cap:
- pivot.cast: axis reassignment with tile mode
- drill.cast: drill-down into aggregated cells
- formulas.cast: formula panel and adding formulas
- import.cast: CSV import wizard walkthrough
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Create docs/demo.tape scripting a ~15s pivot axis reassignment demo.
Hide shell startup and quit so the GIF opens directly on the TUI.
Set up git-lfs for docs/*.gif. Merge LFS hooks with existing beads hooks.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add demo recording tools (asciinema, vhs) and release tooling
(cargo-dist) to nativeBuildInputs. Include a cargo-dist wrapper so
`cargo dist` works as a subcommand. Add scripts/record-demo.sh for
consistent asciinema cast recording at 100x30.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
40-row CSV with obviously-fake sales data (fictional companies like
Acme Corp, Wonka Industries, Cyberdyne Systems). demo.improv generated
via headless import with Profit formula and a default view showing
Region+Product on rows, Date_Month+Measure on columns. Added the
import command to the README quick-start section.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add dedicated sections for the data model (formula examples), views and
axes (tile mode, records mode, drill-down), the .improv file format
(annotated example), import wizard and headless scripting, and the
command/effect architecture. Link docs/design-notes.md from the "Why"
section. Update build instructions to use `nix build .`.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Audited SPEC.md against code — mostly accurate but redundant with
repo-map.md and design-principles.md. Minor drift in storage internals,
wizard step count, and mode representation. Salvaged product vision and
non-goals into docs/design-notes.md with staleness disclaimer.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add CL/SQL-style symbol quoting using pipe delimiters for formula
identifiers. This allows category and item names that collide with
keywords (WHERE, SUM, IF, etc.) or contain special characters
(parens, operators, spaces) to be used unambiguously in formulas:
|WHERE| + |Revenue (USD)|
SUM(|Net Revenue| WHERE |Region Name| = |East Coast|)
Pipes produce Token::Ident (same as bare identifiers), so they work
everywhere: expressions, aggregates, WHERE clauses. Double-quoted
strings remain Token::Str for backward compatibility.
Also updates split_where and parse_where to skip/strip pipe delimiters.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add repository, homepage, readme, keywords, and categories fields.
Update description to match project vision. Use GitHub URLs for
public discoverability.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add repository, homepage, readme, keywords, and categories fields.
Update description to match project vision.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add virtual category boundary rule: use regular_category_names() for
user-facing logic, never expose _Index/_Dim
- Document formula tokenizer keyword-aware identifier breaking
- Update repo-map test counts (356 → 510) and add regular_category_names
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The tokenizer already broke multi-word identifiers when the NEXT word
was a keyword, but not when the identifier collected SO FAR was a
keyword. This meant "WHERE Region" was merged into one token when
tokenizing "SUM(Revenue WHERE Region = East)".
Now the tokenizer also checks if the identifier built up to the current
space IS a keyword (WHERE, SUM, AVG, MIN, MAX, COUNT, IF), which
correctly produces separate tokens for "Revenue", "WHERE", "Region".
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Updates `CommitFormula` and `ImportPipeline` to use
`regular_category_names` instead of `category_names` . This ensures that
these components do not target or default to virtual categories (_Index,
_Dim) when no regular categories are present.
Includes updated tests for `CommitFormula` to verify it correctly handles
cases with no regular categories.
Co-Authored-By: fiddlerwoaroof/git-smart-commit (unsloth/gemma-4-26B-A4B-it-GGUF:UD-Q5_K_XL)
Updates `Model` to include `regular_category_names` , which returns
category names excluding virtual categories (_Index, _Dim). This allows
other parts of the application to distinguish between user-defined
categories and system-internal ones.
Co-Authored-By: fiddlerwoaroof/git-smart-commit (unsloth/gemma-4-26B-A4B-it-GGUF:UD-Q5_K_XL)
The tokenizer was greedily consuming spaces and potentially merging
identifiers with subsequent keywords. This change improves the tokenizer
by:
- Peeking ahead past spaces to find the next word/token.
- Breaking the identifier if the next word is a known keyword (WHERE, SUM,
AVG, MIN, MAX, COUNT, IF).
- Adding support for more delimiter characters (<, >, =, !, ").
This fixes a regression where "Revenue WHERE" was treated as a single
identifier instead of an identifier followed by a WHERE clause.
Includes a new regression test for inline WHERE filters in aggregate
functions.
Co-Authored-By: fiddlerwoaroof/git-smart-commit (unsloth/gemma-4-26B-A4B-it-GGUF:UD-Q5_K_XL)
Add unit tests for ImportWizard and ImportPipeline, covering:
- Wizard step transitions
- Proposal and formula editing
- Date configuration
- Edge cases for building the data model
Co-Authored-By: fiddlerwoaroof/git-smart-commit (unsloth/gemma-4-26B-A4B-it-GGUF:UD-Q5_K_XL)
Add a comprehensive suite of unit tests for the formula parser, covering:
- Aggregate functions
- WHERE clauses
- Comparison operators
- Arithmetic operations
- Various error handling scenarios
Co-Authored-By: fiddlerwoaroof/git-smart-commit (unsloth/gemma-4-26B-A4B-it-GGUF:UD-Q5_K_XL)
Add unit tests for various commands including PasteCell, TransposeAxes,
ViewNavigate, and MovePanelCursor to ensure correct command execution and
state changes.
Co-Authored-By: fiddlerwoaroof/git-smart-commit (unsloth/gemma-4-26B-A4B-it-GGUF:UD-Q5_K_XL)
Improve feedback for axis operations:
- `CycleAxisAtCursor` : Now provides a status message if no category is at
the cursor.
- `TileAxisOp` :
- Now provides a status message showing the new axis (e.g., "Category →
Row").
- No longer automatically switches to `AppMode::Normal` , allowing for
multiple consecutive adjustments.
- Provides a status message if no category is at the cursor.
Co-Authored-By: fiddlerwoaroof/git-smart-commit (unsloth/gemma-4-26B-A4B-it-GGUF:UD-Q5_K_XL)
Add unit tests for model precision and category tree rendering:
- `src/model/types.rs` : Added `formula_chain_preserves_full_precision` to
ensure formulas use full `f64` precision for calculations, even when
display is rounded.
- `src/ui/cat_tree.rs` : Added comprehensive tests for `build_cat_tree` ,
covering empty models (virtual categories), expanded/collapsed states,
and item rendering.
Co-Authored-By: fiddlerwoaroof/git-smart-commit (unsloth/gemma-4-26B-A4B-it-GGUF:UD-Q5_K_XL)
