- refactor HydroErr code
- bring HydroErr in as submodule
- removed pyaf
- add tidal filters to the filter command
- remove unused SciencePlots
- removed unused skill_metrics, sourcery refactor, update submodules
- finished move to git submodule toolbox_utils
- read: the columns content should be passed to the pick keyword, not usecols
- date_offset: enforce intervals as an int and improve processing of annual, monthly, and weekly offsets
- finally fixed the cltoolbox functions showing up in descendent projects
- removed unused utility code
- removed skipna and append="columns" that actually weren't doing anything
- refactor with refurb and pylint
- fill.py: force type to be float so that linear (and other mathematical methods) work in fill
- fixed type hint for plot
- another try to get plot tests to pass on github
- plot tests are failing on github that I can't reproduce locally
- plot: reinstated plot tests which are almost identical to plottoolbox tests but use tstoolbox.plot which uses plottoolbox in the background
- moved to the breaking change move in toolbox_utils from typic to pydantic and misc. modernizations
- moved from tstoolbox.tsutils to toolbox_utils.tsutils and finished pyproject.toml
- fixed skiprows option along with updates to read documentation
- cltoolbox rather than mando and plottoolbox.* instead of "tstoolbox plot ..."
- missed that had to import pint_pandas even though not used directly
- reversed start of using numpy.typing because needs a too recent numpy
- changed .format() string to f strings
- allow for a period Datetime index in printiso
- calculate_fdc: added recurrence (return) interval
- _date_slice: fixed some situations where needed tz_localize rather than tz_convert
- copy docs using decorator instead of simple assignment
- shifted to subprocess because os.system wasn't working
- input_ts: changed gof to not need input_ts instead can specify obs_col and sim_col
- gof: shifted to using HydroErr library
- rolling_window.py: window keyword was transformed to wrong type if default
- replace.py: support for None
- clip.py: use pandas clip
- typing versions greater than 2.6.4 adds a subclass test that made many tests fail
- minor refactors
- minor fixes reverting changes from autofix
- replace.py: for type checking needed additional types
- replace.py: None needed to be included in type for typical
- pip_requirements.txt to reduce vulnerabilities
- add upload of coverage results to coveralls
- fixed the way
start_dateandend_datewere handled by adding workingporkeyword - tsutils.py: print_iso changed to correctly print out no spaces table formats *_nos
- plot.py: plottoolbox needed additional keywords to correctly plot time-series
- plot input_ts, cli values
- units, Float64, flexible input
- Correctly include units.
- Correctly including units in name.
- Can work with "A" and "M" frequencies.
- remove pint units for calculations.
- Correct column names to allow for empty units.
- Added "--multi-line 3" to isort
- rearrange to allow "import tstoolbox"
- Reorganized to allow for "import tstoolbox"
- work to add units to column headers
- .pre-commit-config.yaml
- Added the beginnings of a forecast feature style: Added several features to .pre-commit-config.yaml
- Returned the ability to use comma separated file names.
- Added the capability of using the index (usually datetimeindex) as
a
x_train_colsin the regression. fix: Now detect is DataFrame or Series and skip further processing in 'read_iso_ts' - Added "min_count" keyword to
aggregatedocs: Misc. build: Added "extra_requires" to setup.py for development
- New read from xlsx, wdm files.
- rework new multiple sources read feature.
- Continue type hint tests.
- read_iso_ts is now part of common_kwds.
- Minor fixes.
- Docs, fixes because of changed memory_optimize
- memory_optimize is less aggressive
- read --columns filters output.
- "Typical" to coerce args/kwargs.
- added butterworth filter
- Fixed months_across_years.
- Matched Paul Tor's color order.
- Needed matplotlib styles in MANIFEST.in.
- Incorporated SciencePlots mpl styles.
- Pandas implemented new tolerance approach.
- Better working with time zones.
- Lossless compression of dataframes.
- Added plot styles, new default is "bright".
- Correctly uses multiple input to 'regression'
- Added regression sub-command.
- Added 'output_names' keyword to rename columns on output.
- Added hatch patterns to bar plots.
- Allow for index,y to by used for "xy" type.
- Fix x,y plots to allow for index,y plotting
- Added autocorrelation for each column.
- Corrected columns selection when 0
- xy plot incorrectly inserted a column.
- Allow fit to work across multiple columns.
- Added autocorrelation if lags==0.
- Just added some vertical space.
- Improved description of r statistic.
- Added scikit-learn scaling functions.
- Added method keyword.
- lowess and linear fit
- Added coefficient of determination to gof.
- join_axes is deprecated.
- Added force_freq throughout.
- Added "groupby='all' to aggregate.