* Fixes#3648
Code now builds without either MAEParser or coordgen integration enabled.
* add CI build with minimal dependencies
* remove doctests from the new CI build
they need the cairo support
* - replaced set with vector for SMILES-based R-group equivalence
- the first GreedyChunk is constituted by chunkSize+1 mols
- labeled R-groups may not be extracted when onlyMatchAtRGroups==false
- labeled geminal R-groups are incorrectly scored
- my attempt to introduce consistency in R-group labeling was buggy
- added a DEBUG pre-processor directive to the tests to make debugging easier
- added a unit test
- fixed unit test results which were inconsistent with the expected behavior
* changes in response to review
* works, but is not optimal
* swap back to using a reference
* convenience function and tests
* Update rdMolDraw2D.cpp
* another attempt
* trying to suppress the crashing problem
* trying to suppress the crashing problem2
* force a rebuild
* try to make this work
[skip ci]
* stupid typo
* change the SIP imports (suggestion from @ricrogz)
* Update linux_build.yml
* Speed up GraphMol/Chirality.cpp/iterateCIPRanks
The changes in this patch increases throughput of `iterateCIPRanks` by
about 61.8%. This patch consists of many changes that all combine to
contribute to the performance gains, and these are the major changes:
* Introduce Bond.getTwiceBondType to avoid floating point operations in
`floor(2*bond->getBondTypeAsDouble() + .1)`
* In RDGeneral/Ranking.h/rankVect, just store the last index (instead of
the last vector) and avoid creating a copy at the current index.
* In GraphMol/Chirality.cpp/iterateCIPRanks, reserve more capacity ahead
of time.
* In GraphMol/Chirality.cpp/iterateCIPRanks, avoid materializing
localEntry and then copying; instead, track the data to be added in a
`counts` vector and a `updatedNbrIdxs` vector.
In GraphMol/Chirality.cpp/iterateCIPRanks, we also make several
micro-optimizations that were individually measured to give a small
performance improvement.
* Chirality.cpp/iterateCIPRanks: avoid materializing allIndices vec
* misc cleanup
* add a wrapper for bulk nearest-neighbor calculations
A bit of minor refactoring
* header fix
* add tests for new neighbor functionality
* add the rest of the similarity metrics
* - added some missing const keywords
- added an addFingerprint overload to allow passing pointers
- added a test
* changes in response to review
* removed print
* added missing shared_ptr declaration
* added PatternNumBitsHolder serialization
* - merged with upstream changes and resolved conflicts
- got rid of PatternNumBitsHolder and leveraged the serialization version to get the PatternHolder to be backwards-compatible
* built substructLibV1.pkl with an older version of boost
* reverted serialization version to 1
only write numBits if != 2048 and only read numBits if it exists in the archive
* bogus commit just to trigger a rebuild
* Qt drawing works with freetype.
needs polish and testing
non-freetype drawing doesn't work still
* make the Qt rendering anti-aliased
* non-freetype fonts work in Qt
With the giant caveat that you need a QGuiApplication to use fonts. We try to detect that
* cleanup
* changes in response to review
* add qt support to CI builds
* fix VS build problem
* try resolving a linux link error
* Update linux_build.yml
* Update linux_build.yml
* another attempt to get the linux CI builds working
* getting closer?
* Update linux_build.yml
* disable the qt CI builds on windows
* backup
* further progress
* backup
* refactor handling of atom/bond annotations
move to a single annotation list to make it easier to add new annotation
types
* backup, doesn't actually work
* not unhappy with how the bracket rendering now works
still need to do bracket annotations
* backup
* commit first pass at brackets annotations
* add a bit more flexibility to how we handle annotations
Stll needs to be elaborated a bit, but the tests currently pass
* cleanup
* support comic mode with brackets
moves the handdrawnLine code to MolDraw2D_detail, which is a better place for it
* updates in response to review
* works
* auto download of comic-neue font
* support it in the python layer
* getting ready for a PR
* tone down the wobbly line effect a bit
* add end-point shifts
tone the wiggles down a bit
