You're reading an old version of this documentation. For up-to-date information, please have a look at v0.11.
Changelog
This page contains a summary of changes between the official ASpecD releases. Only the biggest changes are listed here. A complete and detailed log of all changes is available through the GitHub Repository Browser.
Version 0.6.4
Released 2021-11-25
Changes
New attribute
comment
inaspecd.tasks.Task
,aspecd.report.Reporter
,aspecd.plotting.Plotter
,aspecd.io.DatasetExporter
allowing for storing user-supplied comments
Fixes
Warnings issued during cooking of a recipe are now log messages.
Version 0.6.3
Released 2021-11-24
Changes
aspecd.report.Reporter
adds template loader for package ifaspecd.report.Reporter.package
is provided, allowing to override templates from the ASpecD framework within derived packages.aspecd.tasks.ReportTask
passes through the default package from the recipe to the reporter for overriding templates.aspecd.infofile.Infofile
: Comment gets converted into a single stringDependency change: Jinja >= 3.0
aspecd.io.TxtImporter
handles decimal separator different than dot
Fixes
aspecd.tasks.Task
warns if key in dict (recipe) is no property of the task.aspecd.processing.DatasetAlgebra
returns shape in error message if shapes differ.Processing and analysis tasks issue warning if result name is identical to dataset label
Ensure window length for Savitzky-Golay filter in
aspecd.processing.Filtering
to always be oddaspecd.processing.CommonRangeExtraction
ignores unit of last axis (i.e., intensity) when checking for identical unitsaspecd.utils.ToDictMixin
: Added superclass call to preserve mro in dependent subclassesTasks properly handle non-dataset results from recipe
aspecd.plotting.MultiPlotter
sets axis labels when units are emptyaspecd.processing.Normalisation
removes unit from last axisaspecd.processing.BaselineCorrection
handles zero values in range properlyaspecd.analysis.AggregateAnalysisStep
no longer addsdatasets
andresult
to output ofto_dict()
aspecd.tasks.AggregatedAnalysisTask
sets correct type in output ofto_dict()
aspecd.tasks.ReportTask
does not add empty figure filenames to includesaspecd.Tasks.PlotTask
preserves labels of drawingsRecipe history does not contain path to current directory in dataset source
Version 0.6.2
Released 2021-11-16
Changes
New parameter
ytickcount
foraspecd.plotting.SinglePlotter2DStacked
to control maximum number of yticksNew parameter
tight_layout
foraspecd.plotting.Plotter
to prevent labels from getting clipped
Fixes
Recipe containing a MultiplotTask does not contain datasets as dicts
PlotTask with automatically generated filenames and >1 datasets writes correct filenames to figure record in recipe
CompositePlotter sets plot style of plotters
Grammar in
dataset.tex
templateColophon of report via ReportTask contains default package set in recipe
CompositePlotter does not add plotters of subfigures to list of dataset representations and list of dataset tasks
Escaping of “_” in LaTeX templates
Version 0.6.1
Released 2021-11-10
Fixes
ReportTask works with output directory set in recipe
LatexReporter finds templates with relative path
Yaml handles numpy floats and ints
ProcessingTask handles multiple datasets, SinglePlotTask automatically generated figure filenames with multiple datasets
ProcessingTask no longer tries to deep-copy matplotlib objects
Replacing dataset labels in recipes works with dataset ids/source
Version 0.6.0
Released 2021-11-05
New features
Reports
Templates for reporting information contained in datasets come bundled with ASpecD.
Context contains
templates_dir
allowing to include sub-templates.New class
TxtReporter
for plain text reports
Tasks/Recipe-driven data analysis
YAML representation of recipe and tasks via
aspecd.tasks.Recipe.to_yaml()
andaspecd.tasks.Task.to_yaml()
Figure labels can be set in plotters; otherwise a default label will be set and can be accessed from within reports.
Utils
change_working_dir()
can be used as context manager to temporarily change the working directory.
General
aspecd.utils.ToDictMixin.to_dict()
can optionally remove keys with empty values.to_dict()
method inaspecd.processing.ProcessingStep
,aspecd.analysis.AnalysisStep
,aspecd.annotation.Annotation
,aspecd.plotting.Plotter
,aspecd.table.Table
,aspecd.report.Reporter
,aspecd.model.Model
Models
Axes quantities and units can be explicitly set on model creation.
Changes
Dataset labels do not contain source path.
Recipe dataset_source and output directories are no longer converted to absolute paths.
More complete recipe history for tasks, including more of their properties
Recipe-driven data analysis: Figures get added to recipe with default label if no label is provided.
aspecd.processing.Noise
: explicit noise amplitude can be given.Model can add label to created dataset.
ModelTask adds result label as id to result.
Plotter: Default figure size changed to (6., 4.) inch
Fixes
aspecd.tasks.Task.to_yaml()
serialises numpy arraysDatasets from foreign packages are correctly listed in recipe history
aspecd.utils.copy_keys_between_dicts()
properly traversesaspecd.utils.Yaml
handlesnumpy.double
Recipe-driven data analysis: automatically generated figure filenames get added to recipe figure record
Models work now correctly when based on a dataset
aspecd.model.FamilyOfCurves
sets correct values for additional axisaspecd.processing.Differentiation
works correctly for 2D datasetsaspecd.processing.Noise
: normalisation works with >1D datasetsaspecd.plotting.SinglePlotter2DStacked
: ylabel is set to third axis if offset = 0
Version 0.5.0
Released 2021-10-12
New features
Tasks/Recipe-driven data analysis
YAML representation of tasks and recipes using
aspecd.tasks.Task.to_yaml()
andaspecd.tasks.Recipe.to_yaml()
- convenience methods for later use in guided recipe generationaspecd.tasks.AggregatedanalysisTask
for performing a SingleAnalysisStep on a series of datasets, aggregating the result in a CalculatedDatasetaspecd.tasks.TabulateTask
for tabular representation of data of a dataset
Datasets
New attribute
aspecd.dataset.Axis.index
(for individual labels for each data point, similar to pandas and for tabular data)aspecd.dataset.Dataset.tabulate()
to create tables from datasets
Analysis steps
New class
aspecd.analysis.AggregatedAnalysisStep
for aggregating the results of a SingleAnalysisStep on multiple datasets in a CalculatedDataset
Tabular representation of datasets
New module
aspecd.table
Series of output formats for tables (including DokuWiki and LaTeX)
Tables can have captions that are output as well
Utils
aspecd.utils.get_package_data()
for obtaining package data (i.e., non-code files contained in distribution)
Changes
aspecd.analysis.BasicCharacteristics
always returns scalars or lists in its results and writes index (for compatibility withaspecd.analysis.AggregatedAnalysisStep
and tabular output).
Version 0.4.0
Released 2021-10-08
Note: Starting with this release ASpecD requires Python >= 3.7.
New features
Tasks/Recipe-driven data analysis
New attribute
aspecd.tasks.PlotTask.target
allows adding a plot to an already existing plot.aspecd.tasks.Task.to_dict()
adds (implicit) parameters of underlying task objectClasses from the ASpecD framework can be used without prefixing them with “aspecd” in recipes with “default_package” set to a package based on the ASpecD framework.
serve
command outputs log messages for each taskCommand-line options for
serve
setting the log level/verbosityCatching of errors, excluding the stack trace and only showing the error message (but full stack trace in verbose mode)
Switch in recipe to suppress writing history (for development/debugging, issuing warning on the command line via logging)
New structure of recipes: Move
default_package
andautosave_plots
to new dictsettings
;output_directory
anddatasets_source_directory
to new dictdirectories
Add
format
dict to recipe with fieldstype
andversion
Automatically convert old recipe formats within
aspecd.io.RecipeYamlImporter
Processing steps writing parameters during execution and applied to multiple datasetes are unpacked in the recipe history if these parameters change for each dataset
References in processing and analysis steps and models (using bibrecord package)
Version 0.3.1
Released 2021-09-21
The following bugs have been fixed:
Handling of lists as properties in recipes
Improved handling of axes labels with xkcd style
Offset in SinglePlotter2DStacked can be set to zero
Version 0.3.0
Released 2021-09-02
Note: This is the last ASpecD release with explicit support for Python 3.5.
New features
Processing steps
Adding (coloured) noise to datasets (
aspecd.processing.Noise
)Provide a new range of axis values for a dataset for correction (
aspecd.processing.ChangeAxesValues
)
Analysis steps
Power spectral density of 1D dataset (
aspecd.analysis.PowerDensitySpectrum
), e.g., for analysing noisePolynomial fit of 1D data (
aspecd.analysis.PolynomialFit
)Linear regression of 1D data without fitting the intercept (
aspecd.analysis.LinearRegressionWithFixedIntercept
)Additional methods in
aspecd.analysis.BlindSNREstimation
Class
aspecd.model.Model
New attribute
aspecd.model.Model.description
New non-public method
_sanitise_parameters
New models
aspecd.model.Polynomial
for evaluating polynomials (e.g., as obtained usingaspecd.analysis.PolynomialFit
)aspecd.model.CompositeModel
for models consisting of a (weighted) sum of individual modelsaspecd.model.FamilyOfCurves
for inspecting systematic variations of one parameter of a given model
Tasks
Comments can be added easily to processing and analysis steps using the top-level key
comment
of the respective task.
Utils
aspecd.utils.not_zero()
ensuring a float not to cause DivisionByZero errors
Changes
aspecd.processing.Differentiation
usesnumpy.gradient()
instead ofnumpy.diff()
aspecd.processing.BaselineCorrection
returns polynomial coefficients in unscaled data domain
Fixes
Axis labels without “/” if no unit is present
aspecd.metadata.Measurement
handles dates imported from YAML (implicitly converted into datetime.date object)
Version 0.2.2
Released 2021-06-19
The following bugs have been fixed:
Normalisation to minimum now divides by absolute value of minimum
Normalisation raises ValueError in case of unknown kind
Import with explicit importer when importer resides in (sub)package
Recipe history shortens dataset source if dataset_source_directory has trailing slash
Version 0.2.1
Released 2021-06-03
The following bugs have been fixed:
Report: template can have “..” in its path
aspecd.utils.copy_values_between_dicts()
cascades through source dictAdd missing template files for sphinx multiversion
Additionally, the following new helper functions appeared:
Version 0.2.0
Released 2021-05-19
New features
Singleplot and multiplot tasks automatically save results to generic file(s) when no filename is provided
Importer and importer parameters can be set in recipe
DatasetImporterFactory: importer can be set explicitly, parameters can be passed to importers
DatasetImporter with parameters property
Meaningful error messages for exceptions in ProcessingSteps, AnalysisSteps, Plotters
Method
create_dataset()
in AnalysisStepPeakFinding (for 1D data)
BlindSNREstimation (currently only with simplest method)
BasicStatistics (mean, median, std, var)
BasicCharacteristics (min, max, amplitude, area)
ProcessingStep provides non-public method
_set_defaults()
for setting default parameters before sanitising parameters.Filtering (with uniform, Gaussian, and Savitzky-Golay filter)
Interpolation (at least for 1D and 2D datasets)
Normalisation:
Act on parts of the data of a dataset
Handle noise for ND data with N>1
RangeExtraction: extract range of data from dataset (using slice notation)
MultiprocessingTask (and SingleprocessingTask aliasing ProcessingTask)
ScalarAxisAlgebra: perform scalar algebra on axis values
DatasetAlgebra: add and subtract data of second dataset to/from dataset
CommonRangeExtraction for 1D and 2D datasets
SinglePlotter2D:
Filled contour plot with additional contour lines that can be styled
Lines of contour plot can be styled
SliceExtraction now handles both, axis indices and axis values
Changes
Removed parameter
source
from method_get_importer
in DatasetImporterFactory, importer factories of derived packages now handle ASpecD-implemented importers by default.Dataset: method
aspecd.dataset.append_history_record()
made publicSystemInfo: Packages contain now full list of dependencies with version numbers of currently installed packages
SliceExtraction:
parameter “index” renamed to “position”
works for ND datasets with N>1
ProcessingStep split into SingleProcessingStep and MultiProcessingStep
All processing steps previously inheriting from aspecd.ProcessingStep need to inherit now from aspecd.SingleProcessingStep to continue working as expected.
Plots throw “NotApplicableToDataset” exceptions rather than “PlotNotApplicableToDataset”
Fixes
SingleanalysisTask assigns results of multiple (individual) datasets
Exceptions print messages
New dependencies
scipy (for interpolation in ExtractCommonRange and various analysis steps)
Version 0.1.1
Released 2021-05-03
The following bugs have been fixed:
MetadataMapper: Fix sequence of mapping operations performed
MetadataMapper: Mappings are automatically loaded from file if filename is given
CompositePlotter: Legends for subplots work
SliceExtraction: Remove correct axis from dataset
MultiPlotter1D*: Fix problem in conjunction with CompositePlotter and assigning drawings
SliceExtraction: fix problem extracting slice with index zero
CompositePlotter: more intuitive axes_positions
Fix bug with aspect ratio of 2D plots using imshow
Update intersphinx mapping for matplotlib
Version 0.1.0
Released 2021-04-24
First public release
List of generally applicable concrete processing steps
List of generally applicable concrete plotters
Recipe-driven data analysis fully working with history
Introduced ASpecD dataset format (ADF)
Version 0.1.0.dev280
Released 2019-06-14
First public pre-release on PyPI