* prefer to not wedge bonds to double bonds
* addWavyBondsForStereoAny() works
* add option to clear bond stereo
* support wavy bonds in prepareMolForDrawing()
get rid of MolDraw2D/catch_main.cpp
* add a drawing example
* name the bug correctly
* add that new functionality to the python wrapper
* remove some debugging cruft
* Fixes a variety of problems with simulating counts using the FingerprintGenerators from Python
1. the countBounds arguments were not working
2. the GetXXXGenerator() functions used inconsistent naming for the countSimulation argument
3. there were no tests to make sure that the changing countBounds actually did something
* typo
Shift `Trajectory` and `Snapshot` class constructors to methods on the classes themselves to prevent shadowing of class names in the Python wrapper module
* Changed SVG output so all coordinates are to 1 decimal place.
Adjusted tests accordingly.
* Checking test1 SVGs via hashcode.
* Added testing of all generated test files by comparing hash-codes.
The few cases where this doesn't work are mentioned in the code.
* close streams
* Close files after writing - Windows won't delete them otherwise.
* Changes post-review.
* Proposed change post-review for further discussion.
* Use MolDraw2D_details::doubleFormat.
* Copied improved check_file_hash() from catch_tests.cpp.
Updated a few hash codes.
* Copied improved check_file_hash() from catch_tests.cpp.
Updated a few hash codes.
Added hash codes for no-Freetype version.
Co-authored-by: David Cosgrove <david@cozchemix.co.uk>
Co-authored-by: Greg Landrum <greg.landrum@gmail.com>
* Allow int->int64 conversions
* Add original python test
* Add self to class method
* Quick cout to see what the error really is
* Add overflow checks to integer conversions in getProp
* Fix conversions, convert to ValueError in python for bad conversions
* Trigger tests to run again
* Revert "Quick cout to see what the error really is"
This reverts commit 70feea687e.
On 64-bit windows, I noticed the value of the QPainter pointer can exceed the
maximum unsigned long int value, which caused a python OverflowError.
Changing the type to size_t should prevent this.
* Most tests working
* All tests working
* Fixed tests after merge with master
* Create header and implementations for RCore
* Updated comments
* Removed old code
* DLL export for MolMatchFinalCheckFunctor
* Information line for failing Mac test
* Log replace core behaviour
* Ordering fix for OSX
* Possible fuzzer fix
* Removed debug output
* Fix unmatched user R group bug
* Code review changes
* Bug fix and ChemTransforms test
* Swap to using a data structure for default normalization parameters
* bring the default fragment data into the code too
* cleanup
* add reionizer parameters via data
change fragment parse failures to ValueErrorExceptions
* tautomer parameters in the code
* got a little over-enthusiastic in that last cleanup
* use boost::flyweight to cache normalization and charge data params
* a bit more cleanup
* support reading params from JSON
* fragments from JSON
single-call for fragment removal
* add a one-liner for the canonical tautomer
* quick refactor
* Fixes#4115
* complete the parents
* docs
* move the definitions to a namespace and make them const
* see if switching to c++14 fixes the CI compile problems with g++ 5.5
* somewhat uglier way of solving the initalizer list problem
* set all source code files to have native line endings
* normalized all source code line endings
Co-authored-by: Paolo Tosco <paolo.tosco@novartis.com>
* Avoid that MolStandardizer segfaults on empty mols
* changes in response to review
* added missing not operator
Co-authored-by: Paolo Tosco <paolo.tosco@novartis.com>
* start refactor query pickling
* further query pickling refactoring
* bit of modernization
* partial implementation of query writing
* serialize recursive queries
* new tests
* support QueryAtom and QueryBond with null queries
* more query cleanup
* tests pass
* switch from std::variant to boost::variant
* update clang version for CI
* explicit instantiation of templates
* get dll builds working on windows
* optionally skip h counts on parsing; more query tests
* change in response to review
* changes in response to review
* Support Chemical Markup Language, CML, for writing
Implement RDKit::MolToCMLBlock and RDKit::MolToCMLFile
http://www.xml-cml.org/
* a number of cleanups and fixes
1. stop writing the cml: namespace to the output. Tools like marvin can't read that and it's not nececssary
2. fix the H count
3. continue writing atom info for 2D confs
4. simplify the writing of atom parity
* make the conformer optional
write bond hash/wedge information
* update the expected python results
* Remove overloaded function
Co-authored-by: greg landrum <greg.landrum@gmail.com>
