1.5.11 Release Notes
John May edited this page Sep 4, 2015
·
12 revisions
It's been a while since 1.5.10 (Dec '14) and although the bulk of this release consists of bug fixes, PMD coverage improvements, and clean up there are several notable additions.
- Bayesian modelling framework (see Clark et al. 2015).
- MMFF partial charges (see Blog Post).
Mmff mmff = new Mmff();
mmff.assignAtomTypes(mol); // check return
mmff.assignPartialCharges(mol);
- Data representation of pentuple and sextuple order bonds - note no format IO ATM.
- Faster pharmacophore though using modern search APIs.
- Extended canonical labelling to handle large (bio)molecules. Unique SMILES (all-atom) can now be generated for ~98% of SwissProt. The largest protein handled was CO1A1_BOVIN (1463 AAs) in 28.5 ms (see Blog Post).
- Faster InChI generation through disabling AuxInfo generation. This InChI part is normally discarded but can account for a moderate (>10%) portion of runtime.
- The ionisation potential tool has been deprecated. It is based on a machine learning method that used a bugged version of some descriptors. Now the descriptors are fixed the results are not correct.
-
DeAromatizationTool
andDeduceBondSystemTool
useKekulization
instead. -
MMFF94AtomTypeMatcher
useMmff
- AuxInfo is not longer generated for InChI unless no options are explicitly passed.
// wrong (okay under 1.5.10)
String auxinfo = InChIGeneratorFactory.getInChIGenerator(ac).getAuxInfo();
// right
String auxinfo = InChIGeneratorFactory.getInChIGenerator(ac, "").getAuxInfo();
- Several low-level methods have been upgraded from raw type lists, a cast may now be needed when invoking them.
ChiIndexUtils.evalValenceIndex
ChiIndexUtils.getUniqueBondSubgraphs
WeightedPathDescriptor. getPathWeights
MDMolecule.setResidues
MDMolecule.setChargeGroups
21,292 (+/-?) tests - not including slow tests
7 (-21) failures
0 (+0) errors
74 John May
68 Egon Willighagen
4 Tomas Pluskal
4 Krishna Dole
2 Zach Charlop-Powers
3 Yap Chun Wei
2 Alex Clark
1 Gilleain Torrance
1 Miguel Correa
39 Egon Willighagen
42 John May
- Bump version for release
8e11eb3
- Merge pull request #147 from cdk/java3d-dep
e2a307c
- Merge pull request #146 from egonw/fix/pmdWarnings
2b89dbd
- Creates fewer String objects: only concatenate when debugging is turned on
4bd50de
- Fixed a PMD warning: Variables that are final and static should be all capitals
1e6e999
- Close readers when we're done with them
220a863
- Removed some unused code
8f54e78
- Fixed the unit tests: fail() causes the method to exit, so that not all tests were run
cec4cb4
- Fixed a few 'unused variable' warnings
857b0c0
- Removed some dead code
c53eac9
- Cooked some lists (no longer 'raw type')
3398981
- Upgrade java3d dependency - we previously used 1.5.2 then switch to 1.3.1 but this had an API difference that users depended on.
40c1007
- Disable AuxInfo generation in InChIFactory by default. This can be as much as 20% of the compute time and is largely ignored. If the AuxInfo is wanted it can be generated by explicitly passing in an empty set of options.
dfd18ef
- Merge pull request #144 from egonw/fix/pmdWarnings
ace58d2
- Merge pull request #143 from egonw/feature/ignorePMDonSMSD
ee57194
- Removed unused imports
d47a308
- User indexOf(char) instead of indexOf(String).
a05ca52
- Fixed a few variable name warnings
ed84551
- Ignore PMD warnings/errors in the SMSD section
f22bc97
- More efficient use of StringBuilder.append()
d865450
- Don't use String concatenation to deal with null atoms
42ae450
- Don't use String concatenation to deal with null atoms
20c145a
- Use interfaces instead of implementations: IPseudoAtom
cd91c50
- Fixed a PMD warning: use append(char) instead of append(String) for single chars (and a few more append() split ups)
66f8f71
- Don't compare doubles as identical, as there is machine uncertainty in any non-integer value
af852f0
- Include copyright statements and some cleanup after review.
7acf190
- Deprecate implementations and ignore failing tests of old methods.
ef57f35
- Delegate existing MMFF partial charge methods to new implementation.
cce1021
- New MMFF partial charges implementation. Relocate tests from descriptor and rebase using SMILES (required due to InChI standardisations).
aea097e
- Correct citation id.
df40c1c
- Internal class for MMFF Parameter Loading - duplicates/replaces some functionality of the ForceFieldConfigurator/MMFF94BasedParameterSetReader.
86fcbe1
- Map symbolic types HOCS and HON down to integer type 21. These appear in the validation suite and labelled as new in the parameter files but do not appear in the papers.
ac69e27
- Internal API additions allow us to pass back which bonds are aromatic by MMFF's model. Ideally we would seperate out the aromaticity to a optional preprocessing step but at the moment the aromaticity needs to be assigned after the initial symbolic types.
9e6a6e4
- Representation independant type assignment for nitro groups.
5d825f6
- Load prime numbers from an external file.
dc7f453
- In lieu of planned improvments to canonical labelling, extend the primes table to allow handling of much large molecules.
e037988
- Added Miguel Correa, of a recent patch
4072b0d
- Default name is null.
dfc4d22
- Bad SMILES are marked with an empty molecule and property. This mirrors behaviour of the IteratingSMILESReader.
a45f745
- Removed code to enable debug output
9bec3fc
- It's 2015.
8afd272
- Report the build status in the README
1bfa2d1
- Enable travis-ci building
b7b89f3
- Detect correct molfile format in IteratingSDFReader.
976ad3e
- Added the literature reference
0efa7e1
- Just count the number of oxygens directly from the containers
d26551b
- Backported the first descriptor from NanoJava: chem formula oxygen count
4436bde
- Revert "Backported the first descriptor from NanoJava: chem formula oxygen count"
dbdb7ed
- Revert "Added JavaDoc"
3a9dea3
- Revert "Just count the number of oxygens directly from the containers"
8b36b5e
- Revert "Added the literature reference"
6e21795
- Added the literature reference
a84db31
- Just count the number of oxygens directly from the containers
bddda96
- Added JavaDoc
68331e1
- Backported the first descriptor from NanoJava: chem formula oxygen count
c92f855
- Updated to fit in.
cfcd236
- Copied in ISubstance from NanoJava
835e0fc
- Merge branch 'master' of github.com:cdk/cdk
5b169e9
- Update README.md
58a4e74
- Base object property handling tweaks.
753adfe
- SMARTS matching for smallest ring size atom defaults to any ring atom.
fc07d45
- Allowing setting of stereochemistry.
ba945d3
- Expose lower level API allowing efficient matching of different parameters (e.g. charge and symettry class etc).
1b022f0
- Provide access to the sub-expressions in the newer logical atom implementations.
877427c
- Handle quadruple bonds in SMARTS matches.
c963d7b
- Merge pull request #134 from cdd/readme-fixes
1ba0333
- Merge pull request #133 from cdd/java-version
be3a339
- Java 1.6 does not appear to be supported anymore: update readme.
dc6d42c
- Readme typo fixes and consistent command examples.
f5de611
- Merge pull request #132 from cdd/bundle-cdk-model
19c14f8
- Include cdk-model when packaging a single JAR.
e159604
- Not all InChIs have connections
c097bb0
- Unit test for InChIs with no connectivity layer
17ae8f2
- Closes readers when we're done with them
96b4896
- Merge pull request #130 from miguelcorrea/master
b1a7e4d
- Fixed several grammar and spelling errors in the Javadocs
090b355
- Added a note about this not being the primary source of this file
912ba0f
- Use the newer cdk-build-utils version with cite support (inbook and techreport) and DOI outlinks.
feb5fd8
- Added email and url for developer Egon
f98e38f
- Use new Mappings API - the current pcore API doesn't really fit well since the mappings is lazy and it is redone multiple times. However the new calls are now decoupled and only required calling of matches() instead of other methods in order. Dirrect return of Mappings instance (ala. Pattern.matchAll) would perhaps be better - also interfaces (List instead List) would be an improvement but would break existing usages.
a0c9dc7
- Slightly faster tests using silent and avoiding redundant setup.
bf8dbbd
- There appears to be some equality logic in the PCoreAtom so the contains check on IAtomContainer might be needed. However it was still doing reference equality, behaviour is now working as expected.
58cccb6
- Extract new pcore atom creation to method.
b1c2c43
- contains does an identity check and is therefore redunant - unless some extra logic was intended here?
cb28cde
- Use newer SMARTS/Substructure API calls.
1d76286
- Minor tidy up/variable renaming.
b0c9468
- YourKit info.
fafd8b3
- Access new instance variables in test.
3b021dd
- Move DeAromatizationTool to legacy and deprecate.
1438a84
- Improve the dearomatization of aromatic rings by extending it to handle three fused 6-membered rings, fused 5-membered ring with two 6-membered rings, two fused 6-membered rings, 5-membered ring fused with 6-membered ring systems. Also generalize the dearomatization of 5-membered ring and 6-membered ring systems.
6a9ff75
- Remove static variables so that VFlibTurboHandler can be used in a threaded environment.
f0c1512
- Remove static keywords from variables source and target so that this class can be used in a threaded environment.
15d8205
- Correct flag for single or double.
13c5303
- Lazy initialize substructure templates, also logging tool is suspiciously expensive.
f5b6ad7
- Cheaper SDG creation through reuse of the same identity template library.
4660f71
- Additional hint in LengthOverBreadthDescriptor to configure istopes.
de3ce25
- SMARTS stereo matching requires additional checks to ensure we don't accidentally attempt to compare a tetrahedral centre to a double-bond (was causing a class cast). The typical case is when this occurs is when bond orders are not matched.
b12d2ca
- Correct handling of explicit hydrogens in model builder 3D.
60d85be
- Exclude slf4j-log4j from iordf.
c64509b
- Synchronise and look atom type matcher.
f7b3388
- Fixed setting the electron counts for QUINTUPLE and SEXTUPLE
9a4638e
- Added testing for QUINTUPLE and SEXTUPLE electron counts
ce411d7
- Added the missing JavaDoc for QUINTUPLE and SEXTUPLE
4153827
- Ensure that V2000 fail in STRICT mode
f78a9c7
- Resolve issue with properties in old MDL format - bug 1356
ef3ec5d
- Updated the AUTHOR list
40746fd
- Handle minus character in formula parsing [O]2- vs [O2]–. Hyphen is still handled correctly.
f63fce8
- Resolve NPE in circular fingerprint - SF bug:1357.
989fd7d
- Tests taking more than a second tagged as slow (non-unit) tests.
f698b3d
- fingerprint.model.Bayesian & test materials added
a2ba77e
- Added literature reference
62f9ddd
- A new module for the cdd bayes modelling code.
5ee6b1c
- Mark unspecified double bond steroechemistry when labelling bonds of a depiction.
bed9d12
- tyrosine and tryptophan were switched.
28e1215
- Merge pull request #113 from zachcp/patch-2
b228ea8
- add the github site to the readme
03c2353
- Removed no longer existing dependencies
98b9d77
- Removed TestClass and TestMethod
7e1f347
- Thirteenth batch of removal of TestClass/-Method use
a3d6cdc
- Twelveth batch of removal of TestClass/-Method use
b7a348d
- Eleventh batch of removal of TestClass/-Method use
1c0410b
- Tenth batch of removal of TestClass/-Method use
15f0098
- Nineth batch of removal of TestClass/-Method use
ffa5858
- Eight batch of removal of TestClass/-Method use
00bd54c
- Seventh batch of removal of TestClass/-Method use
9424578
- Sixth batch of removal of TestClass/-Method use
1951f68
- Fifth batch of removal of TestClass/-Method use
fb51e91
- Fourth batch of removal of TestClass/-Method use
1f3633c
- Third batch of removal of TestClass/-Method use
415f7ce
- Second batch of removal of TestClass/-Method use
ac85c9f
- First batch of removal of TestClass/-Method use
5ad585b
- Describing Maven Plugin Usage
5f7e99b
- Replacing CDKException in MolecularFormulaGenerator constructor with IllegalArgumentException.
fa05c86
- Added two additional tests for molecular formula generator. Fixed a problem that was revealed by these tests.
f05e04d
- Remove now defunct modulesuites, tests are now grouped by module automatically due separate source tree.
a70db08
- Demonstrate correct parsing of SMARTS in bug 909.
e977970
- Demonstrate bug 844 has been resolved.
41adfbc
- Ensure aromaticity applies a consistent representation.
2ffb503
- Deprecating DeduceBondSystemTool.
9bc384c
- Resolve off by one error in bug 489 by ensuring substring doesn't fail.
09ef737
- Demonstrate bug 328 is resolved.
a25a2da
- Bug report number is incorrect
185c53d
- Bug reported fixed and closed
426800b
- All three bugs are reported fixed and reports are closed
3966474
- Bug #545 (previously #1793446) is fixed in master
896a9bd
- Reported fixed: https://sourceforge.net/p/cdk/bugs/463/
75e9450
- This bug was fixed. See https://sourceforge.net/p/cdk/bugs/581/
f02abd7
- The class seems tested now
8af6be5
- PMD variable name warning fixes
9805f3e
- Bug #549 (previously #1802998) is fixed: diborane parses fine in 1.5
d93a4ef
- Move ionpot classes to the legacy module.
d0f7f98
- Deprecate the IonizationPotentialTool and ignore the fail tests. Mark tests as slow - most of these are taking more than a second.
6ab531a
- Updates to README.
7432e34
- Ignore deficiency tests in the old MassToFormulaTool.
4084574
- Minor code cleanup
2d784cc
- Added 2 additional tests for MassToFormulaTool and MolecularFormulaGenerator
8110419
- Use the hasItems matcher instead of is to match a list, avoiding ordering problems
d4215c0
- Resolve failing unit test from vecmath bug. Looks to be a bug in the 64-bit (double) rotation, using the 32-bit (float) transformation gives the expected and correct value.
5d22594
- Bumping version, open for changes.
a2ebdef