* fix broken contributing link

* add tzdata to fix readthedocs build

* Revert "add tzdata to fix readthedocs build"

This reverts commit 9d7b89e.

* Fix interval keyword argument in oedi_9068 gallery example

Fixes pvlib#2791.

The oedi_9068 gallery example calls get_nsrdb_psm4_conus() with
interval=5, but the function's parameter is named time_step (the
interval name comes from the NSRDB API query string, not the Python
signature). This was missed when PR pvlib#2582 migrated the example from
get_psm3 (which used interval) to get_nsrdb_psm4_conus (which uses
time_step).

Changes interval=5 to time_step=5 in the example call.

No test changes needed — test_get_nsrdb_psm4_conus_5min already
exercises time_step=5 against the function directly.

* Apply review feedback: update PSM3→PSM4 references and add contributor

- Update narrative comments from 'NSRDB PSM3' to 'NSRDB PSM4' to
  match the actual function called (get_nsrdb_psm4_conus)
- Rename psm3/psm3_metadata variables to psm4/psm4_metadata for
  consistency with the PSM4 data source
- Add Karl Hill to Contributors in v0.15.3 whatsnew

Addresses Copilot suggestions (+1 from @cwhanse) on PR pvlib#2793.

---------

Co-authored-by: Kevin Anderson <kevin.anderso@gmail.com>

* docs: fix parameter name mismatches in docstrings

- pvsystem.py: `aoi_model` -> `iam_model` in PVSystem.get_iam and
  Array.get_iam (signature parameter is `iam_model`)
- solarposition.py: `time` -> `times` in sun_rise_set_transit_ephem
- irradiance.py: _liujordan documented a nonexistent `pressure`
  parameter; replace with the actual `airmass` parameter
- bifacial/infinite_sheds.py: remove duplicated stale Parameters block
  in _poa_sky_diffuse_pv (listed `f_x` and `npoints` which are not
  parameters of this function)

* Add contributor entry to v0.15.3 whatsnew

* Update pvlib/irradiance.py

Co-authored-by: Adam R. Jensen <39184289+AdamRJensen@users.noreply.github.com>

---------

Co-authored-by: Kevin Anderson <kevin.anderso@gmail.com>
Co-authored-by: Adam R. Jensen <39184289+AdamRJensen@users.noreply.github.com>

)

Thanks @yonryzhu and @echedey-ls

* add tzdata to fix readthedocs build

* fix whatsnew and add clarifying tzdata comment

* Remove unused clearsky._is_leap_year (closes pvlib#2768)

* Add contributor entry to whatsnew

* Address review comments: fix pull number and contributor entry

* Address review comments: fix pull number, contributor entry, restore blank lines

* Remove empty horizon brightening component from haydavies

* Change OrderedDict to Dict

* Apply suggestion from @AdamRJensen

* Add what's new entry

* Update docs/sphinx/source/whatsnew/v0.16.0.rst

Co-authored-by: Adam R. Jensen <39184289+AdamRJensen@users.noreply.github.com>

* Change Dict to dict

* Update docs/sphinx/source/whatsnew/v0.16.0.rst

Co-authored-by: Cliff Hansen <cwhanse@sandia.gov>

* Apply suggestion from @cwhanse

Co-authored-by: Cliff Hansen <cwhanse@sandia.gov>

* Apply suggestion from @cwhanse

Co-authored-by: Cliff Hansen <cwhanse@sandia.gov>

---------

Co-authored-by: Adam R. Jensen <39184289+AdamRJensen@users.noreply.github.com>
Co-authored-by: Cliff Hansen <cwhanse@sandia.gov>

* Change OrderedDict to Dict in perez and perez_driesse

* Add what's new entry

* Update docs/sphinx/source/whatsnew/v0.16.0.rst

Co-authored-by: Adam R. Jensen <39184289+AdamRJensen@users.noreply.github.com>

* Update docs/sphinx/source/whatsnew/v0.16.0.rst

Co-authored-by: Echedey Luis <80125792+echedey-ls@users.noreply.github.com>

* Change Dict to dict

---------

Co-authored-by: Adam R. Jensen <39184289+AdamRJensen@users.noreply.github.com>
Co-authored-by: Echedey Luis <80125792+echedey-ls@users.noreply.github.com>

The meteonorm iotools tests pinned specific values, the full column
list, and exact column counts returned by the live API. Meteonorm
regularly updates their weather dataset and adds parameters to the
'all' set, neither of which is a breaking API change, but both cause
these tests to fail spuriously (e.g. new
'surface_tilt'/'surface_azimuth'
columns broke the exact-list assertion).

Assert only what pvlib controls, not what meteonorm's dataset controls:

- realtime: check the known columns are a subset of the response and
  that the column count is at least as large, instead of exact equality.
- training/tmy: drop the golden-value assert_frame_equal comparisons
  (and their fixtures); check columns, numeric dtype, and that at most
  20% of values are NaN (the guarantee meteonorm provides).

Row counts and index checks are kept since they are determined by the
requested time range, not the dataset.

* Exclude gallery files that start with NX_

* Rename oedi_9068.py example to NX_... to exclude from sphinx-gallery execution

* Fix exclusion regex

* ¿? Add warning into oedi example just in case

* Comment editorial

* Flake8 doing its thing. Once again.

* Flake8 x2

* Flake8 x3

* Pin sphinx-gallery to 0.21

* Add whatsnew entry

* One last ammend to regex

* This is the last ammend to the regex 100% real don't doubt it

* Guess what. I'm updating the regex once again

* Minor edit to whatsnew

* regex escape separator - Win build was broken

* Document regex on different OSes

* Revive import distutils - IDK why it's there, thou

* Remove distutil leftover

* Change prefix NX_ -> noplot_

* Forgot to save conf.py

* flake8 needed love

* Maint: Normalize line endings and add .gitattributes

* Update .gitattributes

Co-authored-by: Echedey Luis <80125792+echedey-ls@users.noreply.github.com>

---------

Co-authored-by: Echedey Luis <80125792+echedey-ls@users.noreply.github.com>

…vlib#2820)

sun_rise_set_transit_spa normalized timestamps to midnight *after*
converting to UTC (introduced in v0.11.1 by pvlib#2055). An evening-local
timestamp is the next calendar day in UTC, so its day of interest was
shifted forward and the function returned the following day's
sunrise, sunset, and transit.

Use the input's local calendar date as the day of interest, restoring
the pre-v0.11.1 behavior. The SPA routine requires 00:00 UTC input, so
the local date is re-labeled (not converted) to UTC. The timezone label
is stripped before normalize() so truncation runs on a naive index:
normalizing a tz-aware index would build local midnight, which does not
exist (spring-forward) or is ambiguous (fall-back) on DST-transition
days and would raise. This yields unixtimes byte-for-byte identical to
the v0.11.0 implementation for all valid inputs. Adds an inline comment
explaining the timezone handling (requested in the issue) and a
regression test covering negative and positive offsets and a
midnight-DST zone.

Fixes pvlib#2238

This change was drafted with AI assistance (Claude); all AI-generated
material has been vetted for accuracy and is license-compliant.

Co-authored-by: Claude <noreply@anthropic.com>

* Add NSRDB PSM4 polar and polar TMY iotools

* Fix linter

* Set utc=True

* Add whatsnew, update tests

* Fix tests

* Set map_variables=False in tests

* Add get_nsrdb_psm4_polar_tmy

* Fix linter

* Add tests for erroneous inputs

* Fix linter

* Minor doc updates based on review

* Change economic zone to plural