v0.25.0

IncrementalInference v0.25.0

Diff since v0.24.6

Closed issues:

• Replace all instances of getManifolds with getManifold (#1233)
• on-manifold covariance estimation (#1280)
• Basic on-manifold example with Sphere(2) and one prior fail (#1293)
• ccw.params::Vector --> .params::NamedTuple (#1321)
• calcFactorResidualTemporary doc and function does not agree (#1332)
• cov function for manififolds (#1336)
• update calcVariableDistanceExpectedFractional to manifolds (#1355)
• Upgrade to RealCircleGroup (#1365)
• Remove tuple requirement from getSample (#1371)
• _solveLambdaNumeric for AbstractManifoldMinimize does not work for deconv (#1374)
• solveTree! should return just one value, tree (#1379)
• Check differential factors use correct approxDeconv result (#1386)

Merged pull requests:

• WIP on manifold optimization and factors (#1292) (@Affie)
• fix null hypo partials (WIP) (#1295) (@dehann)
• towards passthrough 'sampling', (#1297) (@dehann)
• varType in _solveLambdaNumeric and translation and broken SE2 tests (#1299) (@Affie)
• FND w/ typeof(ccw) (#1300) (@dehann)
• bugfix for ProductRepr and include SE(2) test (#1301) (@Affie)
• fix nullhypo mask bug in evalPotentialSpecific (#1302) (@dehann)
• ccw points Any to fix mixed manifolds bug (#1303) (@Affie)
• assertion manifold point type, try calcZDim (#1304) (@dehann)
• fix null and partial tests (#1307) (@dehann)
• restore testPartialNH.jl (#1308) (@dehann)
• improve API, towards infoPerCoord (#1309) (@dehann)
• Use getManifold on factors (#1310) (@Affie)
• Update SE2 manifold test with test_broken (#1311) (@Affie)
• update deprecation in API from AMP v0.4.5 (#1312) (@dehann)
• add _evalFactorTemporary, new test, and fixes (#1313) (@dehann)
• add a small comment (#1314) (@dehann)
• fast-forward dfg integrationt testing (#1315) (@dehann)
• refactor calcFactorResidualTemporary, add _slack (#1316) (@dehann)
• upgrade to FactorGradientsCached! (#1317) (@dehann)
• move MKD serialization downstream to IIF (#1318) (@dehann)
• fix solveTree SE2 (#1319) (@dehann)
• Towards infoPerCoord fix for partials (#1320) (@dehann)
• test SE2 hex (#1322) (@Affie)
• Manifold factors change to return coordinates (#1323) (@Affie)
• Fix approxDeconv for manifold (#1324) (@Affie)
• fix tests on joint relatives (#1325) (@dehann)
• add calcHelix_T and testing (#1326) (@dehann)
• conv belief partials on AbstractPrior (#1327) (@dehann)
• Adding getFactorOperationalMemoryType for NoSolverParams (#1328) (@GearsAD)
• wip toward ipc via perturbation on relative (#1329) (@dehann)
• fast-forward dfg integration testing (#1330) (@dehann)
• towards partial IPC on relative via gradients (#1331) (@dehann)
• swap arg order calcFactorRe.. and _evalFactorTemp... (#1333) (@dehann)
• better organize factor ops memory as entities (#1334) (@dehann)
• mv EvalFactor ApproxConv GenericMarginal (#1335) (@dehann)
• add AbstractManifoldMinimize to factor gradients (#1337) (@Affie)
• CompatHelper: bump compat for "Manifolds" to "0.6" (#1338) (@github-actions[bot])
• Fix identity to identity_element for Manifolds 0.6 (#1339) (@Affie)
• file cleanup and standarization, towards propagateBelief (#1340) (@dehann)
• fix propagateBelief N, more testing, passThru fix (#1342) (@dehann)
• consolidate solveFactorParametric, and deprs (#1343) (@dehann)
• some cleanup (#1344) (@dehann)
• Standardize to TranslationGroup (#1345) (@dehann)
• Mutating function in _solveLambdaNumeric cost (#1346) (@Affie)
• fix getManifolds -> getManifold (#1347) (@Affie)
• fix solveFactorParametric as coordinates only (#1349) (@dehann)
• more cleanup, rm getManifolds (#1350) (@dehann)
• mv approxConvCircular, depr printSummaryGraph, approxConvCircular (#1351) (@dehann)
• comments for future work (#1352) (@dehann)
• change to hyporecipe and better names (#1354) (@dehann)
• fix calcVariableDistanceExpectedFractional (#1356) (@Affie)
• refactor hyporecipe (#1358) (@dehann)
• sampleTangent and samplePoint (#1359) (@Affie)
• Merge sampleTangent and samplePoint to master (#1360) (@Affie)
• Remove N-number of samples from getSample and change to Vector{T} (#1361) (@Affie)
• ccw.params is now a NamedTuple (#1362) (@dehann)
• fast-forward dfg integration test branch (#1364) (@dehann)
• update API getPointPartial! (#1366) (@dehann)
• update heatmap sampler for new API tangents (#1367) (@dehann)
• Try getManifold fall back to getManifoldPartial to get the manifold used in partial priors (#1368) (@Affie)
• Remove getSample tuple (#1373) (@Affie)
• remove ccw.specialzDim (#1375) (@dehann)
• consolidate to ccw.partialDims, not cpt_.p (#1376) (@dehann)
• depr CSMOccuranceUtils and bug fix (#1377) (@dehann)
• move MsgLikelihoods services (#1378) (@dehann)
• Fix and test manifold approxDeconv (#1380) (@Affie)
• solveTree! only returns tree (#1381) (@dehann)
• fast-forward DFG integration testing (#1384) (@dehann)
• differential factors from deconv tangents to points (#1387) (@dehann)
• simplify localProduct (#1388) (@dehann)
• Update to RealCircularGroup (#1389) (@Affie)
• isapprox on ProductRepr (#1390) (@dehann)
• fix bug in _checkVarByRef (#1392) (@dehann)
• better IPC on ARel partial, (#1395) (@dehann)
• bump Julia compat to 1.5 (#1396) (@dehann)
• fsat-forward dfg integration testing IIF v0.25.0 (#1397) (@dehann)
• restore test #1305 (#1398) (@dehann)