- Add a
unit_mappingattribute to show a variable-unit dictionary. - Support plotting of continuous-time data by automatically using the time domain as x-axis in many plots.
- Improve the performance of the
rename()method.
PR #550 harmonized the behavior
of the rename() function when the renaming creates overlapping timeseries data.
Before, this raised an error in case there was a conflict with existing data,
but automatically performed a groupby-sum when mapping to the same coordinates.
This was marked as deprecated; in the future, an error will be raised
for any overlapping coordinates after the rename operation.
PR #548 added a unit_mapping attribute.
This will replace the variables(include_units=True) method.
- #551 Default to IamDataFrame-time-col on the x-axis of plots
- #550 Refactor the
rename()method for performance improvement - #549 Make
plotlyan optional dependency - #548 Add a
unit_mappingattribute to show a variable-unit dictionary - #546 Fixed logging for recursive aggregation
- Add algebraic operations working directly on the timeseries data including automatic handling of units.
- Drop negative weights when performing weighted regional aggregation by default.
- Allow recursive aggregation when (some) intermediate variables exist and perform validation of the existing intermediate variables.
PR #534 changed the behavior
of the aggregate_region() method when performing weighted region aggregation.
By default, if any weights are negative, the corresponding aggregated values
are dropped from the resulting IamDataFrame.
The previous behavior can be achieved by setting drop_negative_weights=False.
- #543 Add a tutorial for working with algebraic operations
- #541 Support units in binary operations
- #538 Add option to set defaults in binary operations
- #537 Enhance binary ops to support numerical arguments
- #534 Add feature to drop negative weights
- #532 Add an option to skip existing intermediate variables when aggregating recursively
- #533 Add an
apply()function for custom mathematical operations - #527 Add an in-dataframe basic mathematical operations
subtract,add,multiply,divide - #523 Add feature to compute weights for de-biasing using count
- #519 Enable explicit
labeland fix for non-string items in plot legend
- Easily order data in the line plot feature.
- Add a module for reading data from the UNFCCC Data Inventory.
- Improved integration with any IIASA Scenario Explorer instance: read non-default versions from the connected database and obtain the "audit" info (scenario upload/edit timestamp and user).
- Performance improvements when aggregating or concatenating data.
- Refactor the entire code base to the Black code style.
PR #507 harmonizes the behavior of
the aggregate() and aggregate_region() methods when performing "empty" aggregation,
i.e., no components exist to perform the requested aggregation.
In the new implementation, an empty IamDataFrame is returned if append=False
(instead of None).
PR #488 changes the default
behavior when initializing an IamDataFrame from xlsx: now, all sheets names
starting with data will be parsed for timeseries data.
- #510 Improve performance of
pyam.concat() - #508 Bugfix for non-empty but invisible header and no rows in 'meta' sheet
- #507 Refactor aggregation-functions to increase performance
- #502 Switch to Black code style
- #499 Implement
orderfeature in line plot - #497 Add a module for reading data from the UNFCCC Data Inventory
- #496 Enable loading meta from csv file
- #494 Small performance improvements
- #491 Allow loading meta from file without exclude columns or with empty rows
- #488 Read all sheets starting with
datawhen reading from xlsx - #486 Enable reading non-default scenario versions from IIASA API
- #483 Add a tutorial for integration with R
- #481 Enable custom index columns
- #477 Add a nightly test suite
- #476 Add docstrings to plotting functions
df.plot.<kind>() - #471 Add a
iiasa.Connection.properties()function to retrieve scenario audit data
- Refactor the plotting library for better UX and consistency with pandas, matplotlib and seaborn.
- Add a
sankey()feature to the plotting library. - Rework the documentation and tutorials.
Several PRs in this release changed the implementation of the plotting library for better UX and consistency with pandas, matplotlib and seaborn.
Replace the calls to plotting features by the following:
plot(...)(orplot(kind='line', ...)) forline_plot()plot.stack(...)forstack_plot()plot.bar()forbar_plot()- ...
These PRs also add an order arg to the plotting functions, and the levels
are ordered based on the run_control()['order'] dictionary by default.
- #464 Add
orderarg tostackplot() - #466 Add a
sankey()function to create sankey diagrams - #475 Add a scatter plot example to the gallery
- #473 Refactor to plotting API following pandas/matplotlib implementation
- #472 Add a
sankey()example to the plotting gallery - #470 Add two types of
orderarg tobarplot() - #467 Refactor the GAMS-pyam tutorial to use the gamstransfer module
- #466 Add a
sankey()function to create sankey diagrams - #464 Add
orderarg tostackplot() - #463 Clarification of initialization-error message if file does not exist
- Extend the data format to work with dimensionless variables (i.e., accept "empty" units).
- Refactor the
interpolate()feature to accept a list of years ordatetimeinstances. - Add the list of authors/contributors to the docs.
PR #456 changed the interface
of the interpolate() feature to do the operation inplace (previous behaviour)
or return a new instance (future default).
A deprecation warning is written if a user does not provide the arg inplace
to warn of the future change.
- #461 Add list of authors to repo and docs pages
- #459 Add a
get_variable_components()function to retrieve or join variable components - #458 Enable
Pathfor IamDataFrame initialization - #456 Speed up interpolation and support interpolation for multiple time-points
- #454 Enable dimensionless units and fix
info()if IamDataFrame is empty - #451 Fix unit conversions from C to CO2eq
- #450 Defer logging set-up to when the first logging message is generated
- #445 Prevent conflicts between attributes and data/meta columns
- #444 Use warnings module for deprecation warnings
- Add a
boxplot()visualization feature to the plotting toolbox. - Implement an API to read data from World Bank Open Data Catalogue.
- Write a tutorial illustrating how to read model results from a GAMS gdx file.
- Define
index,model,scenario, ... attributes and show a summary of the index dimensions onprint(). - Refactor the timeseries data backend for improved performance.
PR #432 added attributes to
access the list of (unique) items of each index dimension
(model, scenario, ...).
The PR also marked as deprecated the equivalent functions
(models(), scenarios(), ...). The new behaviour is closer
(though still different) to what a pandas user would expect.
PR #420 added
an object IamDataFrame._data to handle timeseries data internally.
This is implemented as a pandas.Series (instead of the previous long-format
pandas.DataFrame) to improve performance.
The previous behaviour with IamDataFrame.data is maintained
via getter and setter functions.
- #440 Add
boxplot()visualization feature - #438 Add an
indexattribute of model-scenario combinations - #437 Improved test for appending mismatched timeseries
- #436 Raise an error with appending mismatching timeseries index dimensions
- #432 Add attributes to access index dimensions
- #429 Fix return type of
validate()after data refactoring - #427 Add an
info()function and use inprint(IamDataFrame) - #424 Add a tutorial reading results from a GAMS model (via a gdx file).
- #420 Add a
_dataobject (implemented as a pandas.Series) to handle timeseries data internally. - #418 Read data from World Bank Open Data Catalogue as IamDataFrame.
- #416 Include
metain new IamDataFrames returned by aggregation functions.
- Add new features for aggregating and downscaling timeseries data.
- Update the plotting library for compatibility with the latest matplotlib release.
- Refactor the feature to read data directly from an IIASA scenario database API.
- Migrate the continuous-integration (CI) infrastructure from Travis & Appveyor to GitHub Actions and use CodeCov.io instead of coveralls.io for test coverage metrics.
PR #413 changed the
return type of pyam.read_iiasa() and pyam.iiasa.Connection.query()
to an IamDataFrame (instead of a pandas.DataFrame)
and loads meta-indicators by default.
Also, the following functions were deprecated for package consistency:
index()replacesscenario_list()for an overview of all scenariosmeta_columns(attribute) replacesavailable_metadata()meta()replacesmetadata()
PR #402 changed the default
behaviour of as_pandas() to include all columns of meta in the returned
dataframe, or only merge columns given by the renamed argument meta_cols.
The feature to discover meta-columns from a dictionary was split out into
a utility function pyam.plotting.mpl_args_to_meta_cols().
- #413 Refactor IIASA-connection-API and rework all related tests.
- #412 Add building the docs to GitHub Actions CI.
- #411 Add feature to pass an explicit weight dataframe to
downscale_region(). - #410 Activate tutorial tests on GitHub Actions CI (py3.8).
- #409 Remove travis and appveyor CI config.
- #408 Update badges on the docs page and readme.
- #407 Add Codecov to Github Actions CI.
- #405 Add ability for recursivley aggregating variables.
- #402 Refactor
as_pandas()and docs for more consistent description ofmeta. - #401 Read credentials for IIASA-API-Connection by default from known location.
- #396 Enable casting to
IamDataFramemultiple times. - #394 Switch CI to Github Actions.
- #393 Import ABC from collections.abc for Python 3.10 compatibility.
- #380 Add compatibility with latest matplotlib and py3.8
- Add feature to aggregate timeseries at sub-annual time resolution
- Refactored the iam-units utility from a submodule to a dependency
- Clean up documentation and dependencies
- #386 Enables unit conversion to apply to strings with "-equiv" in them.
- #384 Add documentation for the pyam.iiasa.Connection class.
- #382 Streamline dependencies and implementation of xlsx-io
- #373 Extends the error message when initializing with duplicate rows.
- #370 Allowed filter to work with np.int64 years and np.datetime64 dates.
- #369
convert_unit()supports GWP conversion of same GHG species without context, lower-case aliases for species symbols. - #361 iam-units refactored from a Git submodule to a Python dependency of pyam.
- #322 Add feature to aggregate timeseries at sub-annual time resolution
- Improved feature for unit conversion using the pint package and the IAMconsortium/units repository, providing out-of-the-box conversion of unit definitions commonly used in integrated assessment research and energy systems modelling; see this tutorial for more information
- Increased support for operations on timeseries data with continuous-time resolution
- New tutorial for working with various input data formats; take a look
- Rewrite and extension of the documentation pages for the API; read the new docs!
PR #341 changed
the API of IamDataFrame.convert_unit() from a dictionary to explicit kwargs
current, to and factor (now optional, using pint if not specified).
PR #334 changed the arguments
of IamDataFrame.interpolate() and pyam.fill_series() to time. It can still
be an integer (i.e., a year).
With PR #337, initializing
an IamDataFrame with n/a entries in columns other than value raises an error.
- #354 Fixes formatting of API parameter docstrings
- #352 Bugfix when using
interpolate()on data with extra columns - #349 Fixes an issue with checking that time columns are equal when appending IamDataFrames
- #348 Extend pages for API docs, clean up docstrings, and harmonize formatting
- #347 Enable contexts and custom UnitRegistry with unit conversion
- #341 Use
pintand IIASA-ene-units repo for unit conversion - #339 Add tutorial for dataframe format io
- #337 IamDataFrame to throw an error when initialized with n/a entries in columns other than
value - #334 Enable interpolate to work on datetimes
This is a patch release to enable compatibility with pandas v1.0.
It also adds experimental support of the frictionless datapackage format.
- #324 Enable compatibility with pandas v1.0
- #323 Support import/export of frictionless
datapackageformat
- New feature: downscale regional timeseries data to subregions using a proxy variable
- Improved features to support aggregation by sectors and regions: support weighted-average, min/max, etc. (including a reworked tutorial)
- Streamlined I/O: include
metatable when reading from/writing to xlsx files - Standardized logger behaviour
PR #305 changed the default
behaviour of aggregate_region() regarding the treatment of components at the
region-level. To keep the previous behaviour, add components=True.
PR #315 changed the return
type of aggregate[_region]() to an IamDataFrame instance.
To keep the previous behaviour, add timeseries().
The object returned by [check_]aggregate[_region]() now includes both the
actual and the expected value as a pd.DataFrame instance.
The function check_internal_consistency() now returns a concatenated dataframe
rather than a dictionary and also includes optional treatment of components
(see paragraph above). To keep the previous behaviour, add components=True.
- #315 Add
equals()feature, change return types of[check_]aggregate[_region](), rework aggregation tutorial - #314 Update IPCC color scheme colors and add SSP-only colors
- #313 Add feature to
downscaletimeseries data to subregions using another variable as proxy - #312 Allow passing list of variables to
aggregatefunctions - #305 Add
methodandweightoptions to the (region) aggregation functions - #302 Rework the tutorials
- #301 Bugfix when using
to_excel()with apd.ExcelWriter - #297 Add
emptyattribute, better error fortimeseries()on empty dataframe - #295 Include
metatable when writing to or reading fromxlsxfiles - #292 Add warning message if
datais empty at initialization (after formatting) - #288 Put
pyamlogger in its own namespace (see here) - #285 Add ability to fetch regions with synonyms from IXMP API
- Streamlined generation of quantitative metadata indicators from timeseries data using
set_meta_from_data() - Better support for accessing public and private IIASA scenario explorer databases via the REST API
- More extensive documentation of the
pyamdata model and the IAMC format - Compatible with recent
pandasv0.25
- #277 Restructure and extend the docs pages, switch to RTD-supported theme
- #275 Completely removes all features related to region plotting, notably
region_plot()andread_shapefile() - #270 Include variables with zeros in
stack_plot(see #266) - #269 Ensure append doesn't accidentally swap indexes
- #268 Update
aggregate_regionso it can find variables below sub-cateogories too - #267 Make clear error message if variable-region pair is missing when
check_aggregate_regionis called - #261 Add a check that
keepinfilter()is a boolean - #254 Hotfix for aggregating missing regions and filtering empty dataframes
- #243 Update
pyam.iiasa.Connectionto support all public and private database connections. DEPRECATED: the argument 'iamc15' has been deprecated in favor of names as queryable directly from the REST API. - #241 Add
set_meta_from_datafeature - #236 Add
swap_time_for_yearmethod and confirm datetime column is compatible with pyam features - #273 Fix several issues accessing IXMP API (passing correct credentials, improve reliability for optional fields in result payload)
- the
filtersargument inIamDataFrame.filter()has been deprecated pd.date_timenow has experimental supported for time-related columns- plots now support the official IPCC scenario color palatte
- native support for putting legends outside of plot axes
- dataframes can now be initialized with default values, making reading raw datasets easier
- #228 Update development environment creation instructions and make pandas requirement more specific
- #219 Add ability to query metadata from iiasa data sources
- #214 Tidy up requirements specifications a little
- #213 Add support for IPCC colors, see the new tutorial "Using IPCC Color Palattes"
- #212 Now natively support reading R-style data frames with year columns like "X2015"
- #207 Add a
aggregate_region()function to sum a variable from subregions and add components that are only defined at the region level - #202 Extend the
df.rename()function with acheck_duplicates (default True)validation option - #201 Added native support for legends outside of plots with
pyam.plotting.OUTSIDE_LEGENDwith a tutorial - #200 Bugfix when providing
cmapandcolorarguments to plotting functions - #199 Initializing an
IamDataFrameaccepts kwargs to fill or create from the data any missing required columns - #198 Update stack plot functionality and add aggregation tutorial. Also adds a
copymethod toIamDataFrame. - #197 Added a
normalizefunction that normalizes all data in a data frame to a specific time period. - #195 Fix filtering for
time,dayandhourto use genericpattern_match()(if such a column exists) in 'year'-formmatted IamDataFrames - #192 Extend
utils.find_depth()to optionally return depth (as list of ints) rather than assert level tests - #190 Add
concat()function - #189 Fix over-zealous
dropna()inscatter() - #186 Fix over-zealous
dropna()inline_plot(), reworkas_pandas()to (optionally) discover meta columns to be joined - #178 Add a kwarg
appendto the functionrename(), change behaviour of mapping to only apply to data where all given columns are matched - #177 Modified formatting of time column on init to allow subclasses to avoid pandas limitation (https://stackoverflow.com/a/37226672)
- #176 Corrected title setting operation in line_plot function
- #175 Update link to tutorial in readme.md
- #174 Add a function
difference()to compare two IamDataFrames - #171 Fix a bug when reading from an
ixmp.TimeSeriesobject, refactor to mitigate circular dependency - #162 Add a function to sum and append timeseries components to an aggregate variable
- #152 Fix bug where scatter plots did not work with property metadata when using two variables (#136, #152)
- #151 Fix bug where excel files were not being written on Windows and MacOSX (#149)
- #145 Support full semantic and VCS-style versioning with
versioneer - #132 support time columns using the
datetimeformat and additionalstrcolumns indata
- #114 extends
append()such that data can be added to existing scenarios - #111 extends
set_meta()such that it requires a name (not None) - #109 add ability to fill between and add data ranges in
line_plot() - #104 fixes a bug with timeseries utils functions, ensures that years are cast as integers
- #101 add function
cross_threshold()to determine years where a timeseries crosses a given threshold - #98 add a module to compute and format summary statistics for timeseries data (wrapper for
pd.describe() - #95 add a
scatter()chart in the plotting library using metadata - #94
set_meta()can take pd.DataFrame (with columns['model', 'scenario']) asindexarg - #93 IamDataFrame can be initilialzed from pd.DataFrame with index
- #92 Adding
$to the pseudo-regexp syntax inpattern_match(), adds override option - #90 Adding a function to
set_panel_label()as part of the plotting library - #88 Adding
check_aggregate_regionsandcheck_internal_consistencyto help with database validation, especially for emissions users - #87 Extending
rename()to work with model and scenario names - #85 Improved functionality for importing metadata and bugfix for filtering for strings if
nanvalues exist in metadata - #83 Extending
filter_by_meta()to work with non-matching indices betweendfand `data - #81 Bug-fix when using
set_meta()with unnamed pd.Series and nonamekwarg - #80 Extend the pseudo-regexp syntax for filtering in
pattern_match() - #73 Adds ability to remove labels for markers, colors, or linestyles
- #72 line_plot now drops NaNs so that lines are fully plotted
- #71 Line plots
legendkeyword can now be a dictionary of legend arguments - #70 Support reading of both SSP and RCP data files downloaded from the IIASA database.
- #66 Fixes a bug in the
interpolate()function (duplication of data points if already defined) - #65 Add a
filter_by_meta()function to filter/join a pd.DataFrame with an IamDataFrame.meta table