Radomir Stevanovic

API to turn profiling on/off (collection of QPU/solver run stats) and expose the stats programmatically.

To support a negation of the feature-based solver filtering criterion, we should add a unary `not` operator that can prefix any (?) other operator in the keyword argument to `get_solvers`/`get_solver`/`DWaveSampler(solver=filters_dict)`:...

Model config/solvers/sample according to ping.

Right now `samples` key is an alias for `solutions` list. `solutions` is returned by SAPI, and we introduced `samples` as it sometimes makes more sense from user's perspective. Now that...

Implement a simple blocking submit-poll-retrieve function

1

Sometimes, the whole parallelization scheme we have in place is an overkill, and all a user wants is to blockingly submit a problem and retrieve the result. For this simple...

Initially, the only condition can be a time limit. On problem submit, max allotted wall clock time can be given as additional parameter, e.g.: ``` solver.sample_ising(h, J, timeout=100) ``` In...

Trim response to active qubits/variables only

1

Currently result/samples are expanded to full length of working qubits (2048) even when user has only a couple of active ones.

Abstract the logging layer, as currently implemented in `dwave-cloud-client` (with support for env modifications). Extract perhaps to ~`dwave-cloud-common` package, together with some other generic utilities, so it can be consistently...

With profiling ON, dump details about all calls to QPU to a profiling database. Provide a tool for analyzing. At very least, output summary stats.

Simplify and optimize, adjust for #166.