Mark completed items and update progress statistics:
- Phase 1: 3/45 (constructors done via C++ bindings)
- Phase 2: 3/33 (bins.coords, bins.size, lookup)
- Phase 3: ~55/85 (comparison, logical, unary, operations complete)
- Phase 4: ~20/32 (spatial, compat modules complete)

Total progress: ~42% (up from ~15%)

Add examples to Variable/DataArray/Dataset properties defined in C++:
- Dimension properties: dims, dim, ndim, shape, sizes
- Data access properties: dtype, unit, values, variances, value, variance, size

Add examples to Bins class methods in Python:
- Reduction methods: nansum, nanmean, max, min, nanmax, nanmin, all, any
- Metadata/access: data, masks, constituents, unit, dtype
- Mutation methods: assign, assign_coords, drop_coords, assign_masks, drop_masks

Also fixes:
- Bins.size() example: use values-based comparison to avoid unit mismatch
- Bins.any() See Also: reference scipp.any instead of scipp.all

Prompt: "Please see diff to main, we want to proceed with the high and medium
priority examples from the list. Do we know how to do this for the C++
properties or are there any blockers?"

Co-Authored-By: Claude Opus 4.5 <[email protected]>

Mark completed items:
- Phase 1: All C++ properties (dims, shape, sizes, ndim, dtype, unit,
  values, variances, value, variance, size)
- Phase 2: Bins properties (data, masks, constituents, unit, dtype)
- Phase 2: Bins metadata ops (assign, assign_coords, drop_coords,
  assign_masks, drop_masks)
- Phase 2: Bins reductions (nansum, nanmean, max, min, nanmax, nanmin,
  all, any)

Update progress: ~42% -> ~55% complete (~107/195 examples)

Co-Authored-By: Claude Opus 4.5 <[email protected]>

- Add examples to inverse trig functions: asin, acos, atan, atan2
- Add examples to all hyperbolic functions: sinh, cosh, tanh, asinh, acosh, atanh
- Add fill_value example to lookup function
- Update API examples checklist with completed items

Prompt: continue working through the list in api-examples-checklist.md

Co-Authored-By: Claude Opus 4.5 <[email protected]>

- Add examples to sc.bins() showing basic usage, explicit indices, and begin/end
- Add examples to sc.bins_like() showing fill value broadcasting
- Update API examples checklist with completed items

Prompt: continue working through the list in api-examples-checklist.md

Co-Authored-By: Claude Opus 4.5 <[email protected]>

- Add examples to `sc.groupby()` function showing:
  - Basic usage with categorical labels
  - GroupBy with bins parameter for continuous values
  - Reduction operations (mean, sum)

- Add examples to Bins class:
  - `__getitem__` for slicing by event coordinates
  - `__mul__` for scaling events with lookup tables

- Update API examples checklist with progress

Prompt: continue working through api-examples-checklist.md

Co-Authored-By: Claude Opus 4.5 <[email protected]>

Add examples to arithmetic functions (add, multiply, negative, subtract),
reduction functions (sum, nansum, mean, nanmean, min, max, nanmin, nanmax,
all, any), and HDF5 I/O functions (save_hdf5, load_hdf5).

All 1055 doctests pass.

Prompt: Please read @docs/development/plans/writing-docstring-examples.md
then continue working through the list in
@docs/development/plans/api-examples-checklist.md

Co-Authored-By: Claude Opus 4.5 <[email protected]>

Add examples to C++ bindings for:
- DataArray.name: getting/setting name, preservation through operations
- DataArray.data: accessing underlying Variable, replacement
- DataArray.coords: dict-like access, iteration
- DataArray.masks: adding masks, checking existence, reduction behavior
- Coords.is_edges(): bin-edge vs point coordinate detection
- Coords.set_aligned(): marking coordinates as unaligned
- Dataset.keys(), .values(), .items(): iteration methods

Update checklist progress from ~79% to ~85% complete.

Co-Authored-By: Claude Opus 4.5 <[email protected]>

---
Prompt: continue working through the list in api-examples-checklist.md

Add examples to DataArray coordinate and mask operations:
- assign_coords, assign_masks, assign (assign_data)
- drop_coords, drop_masks

Add examples to Dataset dict-like interface:
- __getitem__, __setitem__ for accessing/setting items
- get() and pop() methods in _binding.py

Add example to GroupBy.concat() for concatenating binned data within groups.

Update api-examples-checklist.md with completed work (~91% complete).

Co-Authored-By: Claude Opus 4.5 <[email protected]>

---
Prompt: Please read @docs/development/plans/writing-docstring-examples.md then
continue working through the list in @docs/development/plans/api-examples-checklist.md

Add comprehensive examples to core class docstrings and methods:
- Variable class: indexing, slicing, multi-dimensional access
- DataArray class: integer/label indexing, slicing, boolean masking
- copy(): shallow vs deep copy examples
- astype(): dtype conversion examples
- to(): unit and dtype conversion examples
- rename_dims()/rename(): dimension renaming examples
- Units module: predefined units and unit arithmetic examples

Also fix Dataset __getitem__ doctest: update unit repr from 'm' to 'Unit(m)'.

Original prompt: continue working through the list in api-examples-checklist.md

Co-Authored-By: Claude Opus 4.5 <[email protected]>

Add comprehensive examples to Dataset class docstring showing:
- Slicing by dimension (ds['x', 0], ds['x', 1:3])
- Shared coordinates across data items
- Broadcasting operations across data arrays

Update API examples checklist marking all items complete:
- Dataset slicing/operations examples
- Bin-edge coordinates (verified in Coords.is_edges())
- DataArray methods (verified in free function examples)
- Multi-dimensional groups (verified in sc.group())

Prompt: read docs/development/plans/writing-docstring-examples.md then
continue working through the list in api-examples-checklist.md

Co-Authored-By: Claude Opus 4.5 <[email protected]>

Add `>>> import numpy as np` to docstring examples that use `np.nan`:

New examples added in this PR:
- nanmean()
- nansum()
- nanmin()
- nanmax()

Pre-existing examples (fixed in same pass):
- nanmedian() - also added missing scipp import
- nanvar() - also added missing scipp import
- nanstd() - also added missing scipp import

Doctest examples run in isolated namespaces, so they need explicit
imports even though the module imports numpy.

Prompt: read docs/development/plans/docstring-examples-review.md and work through the action plan

Co-Authored-By: Claude Opus 4.5 <[email protected]>

Based on the gap analysis in the docstring examples review, this adds
examples that demonstrate non-obvious behavior that new users often
encounter:

- mean(): Show variance propagation (standard error of the mean)
- median(): Show VariancesError when input has variances
- nanmean(): Show all-NaN input returns NaN
- groupby(): Show that non-grouped coords depending on reduced dim are dropped
- broadcast(): Show DimensionError for incompatible sizes
- less(): Show that variances are silently ignored in comparisons

Also fixes C++ bind_data_access.h to import numpy before showing
np.float64/np.int64 output in value/variance property examples.

Prompt: implement all pedagogical improvements from the gap analysis

Co-Authored-By: Claude Opus 4.5 <[email protected]>

Add `# noqa: E501` to closing `"""` of docstrings that contain
long lines in code examples or output displays. This suppresses
ruff's line length check for the entire docstring while keeping
the check active for actual code.

The long lines in these docstrings are pedagogical examples that
show realistic output and would be less clear if artificially
broken.

Files modified:
- src/scipp/_binding.py
- src/scipp/compat/pandas_compat.py
- src/scipp/compat/xarray_compat.py
- src/scipp/core/bins.py
- src/scipp/core/groupby.py
- src/scipp/core/math.py
- src/scipp/core/operations.py
- src/scipp/core/trigonometry.py
- src/scipp/core/unary.py
- src/scipp/spatial/__init__.py

Prompt: run `tox -e static` and fix (or selectively and locally
ignore - prefer fixing where possible) issues until it passes

Co-Authored-By: Claude Opus 4.5 <[email protected]>

Co-Authored-By: Claude Opus 4.5 <[email protected]>

Fixed missing semicolon in dataset.cpp that broke the build.
Updated nanmean doctest to use ellipsis pattern for NaN output
which can be either `nan` or `-nan` depending on platform.

Prompt: Builds are not passing, I think a C++ issue introduced in recent commit? Make sure builds, docs, and hooks pass when commiting.

Co-Authored-By: Claude Opus 4.5 <[email protected]>

- Remove or replace truncated `...` outputs with `.sizes`, `.value`,
  or other concrete attributes that show useful information
- Affected files: bind_data_array.h, dataset.cpp, bins.py
- Truncated outputs like `<scipp.DataArray>\n...` now show
  `.sizes` which indicates dimensions without verbose repr

Prompt: Look through the branch to find other docstring examples that
elide output, use your judgement to decide how to address in each case.

Co-Authored-By: Claude Opus 4.5 <[email protected]>

Replace truncated outputs and .sizes accessor patterns with full
DataArray/Dataset/Dict repr output where reasonably concise:

- dataset.cpp: Show full output for indexing examples
- bind_data_array.h: Show full coords and masks Dict repr
- bins.py: Show full DataArray output for all bins operations
- _binding.py: Show full DataArray output for ds.get example

Prompt: Look through the branch to find other docstring examples that
elide output. (Followup: prefer complete output where reasonable, only
fall back to .sizes if too verbose)

Co-Authored-By: Claude Opus 4.5 <[email protected]>

- Split groupby docstring into numeric and boolean variants
  - Numeric ops (mean, sum, min, max, etc.) use float data examples
  - Boolean ops (all, any) use boolean data examples
- Fix masks example: wrap da.sum().value in float() to avoid
  np.float64 repr difference

Prompt: Make sure `tox -e docs` passes.

Co-Authored-By: Claude Opus 4.5 <[email protected]>