Add script to convert CTML to YAML input format #693

Correctly converts GRI-3.0 from CTML to YAML format.

Check that values for properties are not equal to zero and only include them in that case.

Can now handle undeclared element options and species arrays located in another file.

Adds surface phase types

Composition strings can be separated by comma or by space with no differences. Also, the value could be an int or a float, so check both. Use updated function to set the elemental composition of species.

In phase definitions, density can be specified for certain thermo models

Reaction entries can now be filtered as specified in the include node of the reactionArray node. Fix typo in transport key name at phase level.

reaction_id nodes that look like integers (e.g., '0001') are being read as integers by the parser and converting them to strings throws errors. The fix is to only include the reaction_id if it doesn't look like an integer.

Add metadata to the output. Add newlines between each top-level key.

Use a global flag to determine the available formatting functions from NumPy. Move some functions around for readability.

This change simplifies how classes are registered with the YAML emitter.

Species can be specified in another data source.

The XML file for this test wasn't already present in the repo, so convert the CTI file to XML first.

This commit includes some work-around for the fact that electron-cloud phases include the density in the phase definition but fixed-stoichiometry phases put the density in the species definition.

The docstring for the reaction class has some math formatting. This was causing an invalid escape sequence warning during build. The fix is to make the docstring a raw string.

Requires multipel reactionArray nodes, multiple speciesArray nodes, and the option to skip undeclared third body species.

If a speciesData node has an id other than species_data, use it to label the corresponding section in the YAML output file.

Fixes error when speciesData has an id attribute other than species_data and the phase is StoichSubstance.

The R-K EOS is moving activity coefficient data from the phase node to the species node in YAML. However, other phases are retaining this data in the phase node. This change specializes the previous processing of the activityCoefficient node to be only for the R-K phase thermo model.

Use ThermoPhase to load the input files rather than Solution. This avoids a bug in TransportFactory that would be hit by using Solution.

The implementation is not quite complete, some refactoring needs to be done to make it easier to move data from the phase node to the species nodes.

Just clarify what the input argument is

Create Species and Reaction objects before Phase objects. This allows the Phase constructor to modify the appropriate objects based on reaction filtering or various equation of state models where information is moving from the phase node to the Species.

Use error classes that better describe what is missing and print the actual node instead of its memory address

For Species and Reaction classes. This makes it easier to reference these attributes and the extra specifier is not needed because it is implied by the type of class already.

In the IonsFromNeutral phase, make sure the neutralMoleculePhase is initialized before we try to use it. This prevents a segfault if the input YAML file does not have the neutral-phase key.

In the IonsFromNeutral phase, if a special-species is not included the index is never set after initialization and remains at npos. The check here ensures that a special-species is specified. Also move building the elemVectorI outside of the jNeut loop. It doesn't appear to need to be done for every neutral species, so this would save some effort

When specified in YAML input, the neutral-phase can now be specified as being in another file by using a / to delimit the file from the phase name. This is similar to specifying species or elements in another file. This brings back parity with the XML specification of these phases.

Add Margules and ions-from-neutral-molecule phases. Margules is not complete, it is missing the interaction parameters.

If the Motz-Wise option is specified at the global reactionData node level, this is used to set the option for all the reactions in that node that don't otherwise specify the option. This method is easier than setting the equivalent phase-level option.

These changes allow parsing composition map strings where colons are allowed as characters in keys in a colon delimited map, as well as handling comma or space separated pairs where commas are also valid characters in the keys.

Extract a method to process the polynomials. Ensure that sorting of the temperature ranges correlates with the sorting of the data.

The changes now match all of the options in newReaction, including synonyms. Butler-Volmer parameters are deprecated so the user is warned. Simplify processing interface type reactions into one method, since the electrochem node is optional.

Add a docstring for get_float_or_units. Use the clean_node_text function.

As demonstrated in Cantera#683, sometimes the zero values are intentional and meaningful. Rework the SpeciesTransport class to not use this function.

The Python Expat parser requires that the <?xml version...> tag occur as the first characters in the file, even before any blank space, so lstrip is used to remove any whitespace. In addition, raw & characters are replaced with their escaped version.

If a reactionData or speciesData node has a duplicated id attribute, combine the duplicate sections together. If duplicate reaction id attributes or species names are found, either warn or raise an error.

Add intersphinx extension to auto-link to Python docs. Bump KaTeX version.

A quantity is a value that can be either just a number (i.e., a float) or a number plus a unit (i.e., a str).

Add parsing of arguments passed when the script is run from the command line. A new function is introduced to parse the arguments and is the default function when the file is executed as a script. Update the documentation to add the new function. Add a new argument to convert that takes a string containing the content of a CTML file. This allows convert to be used from within the Python interpreter more easily.

Node tag names are now quoted for clarity.

The Redlich-Kwong phase-thermo model doesn't require the activityCoefficients node. If not specified, values will be found from critProperties.xml

ctml_writer.py maintains global state between calling convert(). This change commits the resulting XML file into the test/data directory rather than converting during the test.

The index from the enumerate is not actually used anywhere.

There isn't a reason to convert gri30 every time the class is instantiated. Move the conversion into the test_gri30 function.

Better conform to PEP-8.

The YAML converters are now installed as scripts from the Python package. The zip_safe flag is needed to ensure that data files are accessible by the package. This was added to resolve a build problem with Python 3.8. I'm not sure why it hasn't been a problem until now.

Bump the minimum Python version to 3.5 to support features in ctml2yaml.

Use argparse to have the CLI match ctml2yaml.py

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add script to convert CTML to YAML input format #693

Add script to convert CTML to YAML input format #693

Commits on Dec 2, 2019