Bump StatsBase and StatsModels requirements

The release of StatsBase v0.19.2 resulted in this package failing to precompile because StatsBase and StatsModels both exported coefnames. StatsModels v0.2.0 resolved this problem. The main purpose of this release is to require these versions of StatsBase and StatsModels.

The dof and nobs methods have been updates and tests added.

Some of the documentation files have been updated by re-weaving the .jmd sources. Those that depend on the Gadfly package cannot be updated until a new version of Gadfly is released.

Require julia v0.6.1

In 0.6.0 the colblocks member of the BlockedSparse type could not be created because a reshaped contiguous view of a vector was not recognized as a StridedMatrix.

Use DataFrames 0.11, CategoricalArrays, StatsModels

Update the code to use v0.11.1 or later of DataFrames, CategoricalArrays, and StatsModels packages.

Use BlockedSparse type to reduce allocation

• Reduce memory allocation and garbage collection in models with multiple vector-valued random effects
• Add tests and test coverage
• Add benchmarks

Documentation updates

Extended documentation - more to come.

Switch to BlockArrays for penalized least squares

• Use BlockArrays types for the A and L members of LinearMixedModel
• Create ScalarFactorReTerm and VectorFactorReTerm types
• Create UniformBlockDiagonal for diagonal blocks associated with VectorFactorReTerm
• Add more benchmarks
• Add tests
• Clean up code to reduce allocation

Fix loglikelihood calculation with weights

• Fixes #89
• Adds more benchmarks
• Introduces and uses model_response(mf::ModelFrame, d::Distribution) to convert a binary PooledDataArray response to a 0/1 floating point vector. This should be done in DataFrames and a pull request will be made for this.
• pass the random number generator to the simulate! method
• methods for A_rdiv_Bc!, etc. that are now in Base are commented with if VERSION < ...
• allow a contrasts specification in lmm and glmm

Fold wttrms and Λ into trms

• reformulate the LinearMixedModel type by incorporating the wttrms and Λ members into the trms member.
• create AbstractTerm with subtypes MatrixTerm and FactorReTerm
• add some benchmarks using the BenchmarkTools package
• remove some of the instances of method definitions for functions from Base with signatures of Base classes only
  • BLAS-like in-place linear algebra with scalar multipliers are now called e.g. αβA_mul_Bc!
  • the remaining problematic methods are operations with Diagonal for which I plan to create a PR on the julia repository after consulting with Tony and Andreas

Lower Cholesky formulation

Travis failures are timeouts on julia-0.6.0-pre. Once the dust settles on the julia new release I will check for bottlenecks.

Last release before v0.8.0

Incorporate a couple of commits on the master branch prior to major changes from merging the LowerCholesky branch.