* ComfyUI: removed `_modelinfo` variable because it is not needed. Only `_modelfullname` can be set from the prompt.
* Some log changes.
* Improved documentation.
* Added combinatorial shuffle.
* A1111: Improved combinatorial with hiresfix.
* Some refactoring.
* Update of AI test instructions.
Co-authored-by: Copilot <copilot@github.com>
* Updated pylint settings.
* Fixed a bug with wildcard or extranetwork mappings folders in a different drive than the extension.
Co-authored-by: Copilot <copilot@github.com>
* Support for additional operators: any_in and contains_any.
* Support for starred initialization in the set command.
* New option to set strict operations for some operators.
* Fixes detecting integers and added float support.
* Fixed echoing system variables.
* A1111: fixed prompt processing when there are more than 500 prompts.
Co-authored-by: Copilot <copilot@github.com>
* Separated tests in multiple files.
* Added pydantic validation for configuration.
* Improved backup/restore on visit of nodes.
* Echoing of variables with default no longer visits the default when not used (and no longer creates variables that were not actually used).
* Updated `ppp_config.yaml.defaults` to enhance documentation and structure for model detection and configuration.
* Improved validation of configuration files.
* Added a new `tests` application in `SUPPORTED_APPS` for testing purposes.
* Support for variables in list of values comparisons.
* Enhanced the processing of "BREAK" and "AND" constructs to support replacing with a comma. Set both to comma as default in ComfyUI.
* Added a new command `setwcdeffilter` to set default filters for a wildcard.
* Fixed bug with echoed variables.
* ComfyUI: Individual help documentation for the nodes.
* Can now test for output variables.
* Improved documentation.
* Implemented handling for attention, scheduling, alternation, AND, and BREAK operators in the new configuration file.
* Moved variant definitions to the new configuration file.
* Improved tests speed by creating the needed PPP object in the process method.
Add 4 TDD tests for the var-vs-var comparison feature (S1-TT2):
- test_cmd_if_var_vs_var_eq
- test_cmd_if_var_vs_var_ne
- test_cmd_if_var_vs_var_contains
- test_cmd_if_var_vs_var_not_contains
All tests fail because the grammar only accepts SIMPLEVALUE on RHS,
not IDENTIFIER. Implementation in S1-T2 will make them pass.
Add 5 tests covering the ValueError crash in __eval_basiccondition()
when int() is called on empty/non-numeric variable values during
integer comparisons. Tests verify expected behavior for warn mode
(comparison evaluates to false) and stop mode (proper PPPInterrupt).
* added "$$" to end of parameters as first choice in a wildcard, to avoid confusion with a choice that could be considered parameters.
* additional fixes in the parser.
* Mapping of extranetworks through the new command. This allows you to use a "virtual" extranetwork that is converted to a real one depending on conditions, like the model kind/variant.
* Fixed a parsing bug with the if command.
* Some refactoring and additional checks.
* Improved timing logging.
* Refactor application support handling.
* Enhanced regex patterns in `ppp.py` for wildcard and choice detection to ensure proper escaping.
* Included icon source files.
* New option on_warning to stop on invalid content.
* New option cleanup_extra_separators_include_eol to choose whether to also remove adjacent EOLs when removing extra separators or not. Default is not to do it. In previous versions it was forced.
* Refactoring of the cleanup method.
* Only simplify attention modifiers when the merge attention option is set.
* A1111: avoid logging about HR prompts and avoid saving them in extra metadata when not different.
* A1111: consider options and environment info in the caching of results.
* Some refactorings.
