Alexander Condello
Alexander Condello
Currently the C++ quadratic models are all indexed `[0, num_variables)` always. That makes working with multiple QMs tricky. For instance, say that you had two bqms, `x0 + x1` and...
It would be nice to add some additional configuration options to the [Formatter](https://github.com/dwavesystems/dimod/blob/36810275d8e2305c0a93495d5b96d5c3a5d6ae05/dimod/serialization/format.py#L209). Some ideas: * Toggle full field names - right now they default to a max length of...
It would be helpful to have a Exact and/or Random DQM sampler for testing. Note that the exact version would be extremely limited.
It would be useful to merge several BQM/DQMs. At the moment one can do ``` bqm.update(other_bqm) ``` to merge two BQM objects. We should add a `DQM.update` as well. Further,...
#788 removes `AdjArrayBQM` and `AdjMapBQM`. We should test that BQMs of those types that were saved to files are loaded as `AdjVectorBQM` and the correct warning is raised.
Right now `bqm = dimod.BQM(vartype='SPIN')` works but `bqm = dimod.BQM(offset=5, vartype='SPIN')` does not. The arguments are (largely) determined by position.
Something like ``` def dqm_to_bqm(dqm, strength): case_starts, linear, quadratic, _ = dqm.to_numpy_vectors() bqm = dimod.AdjVectorBQM.from_numpy_vectors(linear, quadratic, 0.0, dimod.BINARY) for i in range(len(case_starts)-1): cases = range(case_starts[i], case_starts[i+1]) bqm.update(dimod.generators.combinations(cases, 1, strength=strength)) if...
See https://github.com/dwavesystems/dwave-ocean-sdk/pull/21 Also see #298
``` sampleset.recalculate_energy(bqm) ``` should change the sampleset's energy to that as calculated with the given bqm.
pandas is [now at version 1.1.4](https://pypi.org/project/pandas/1.1.4/) but we are still [testing with 0.25.3](https://github.com/dwavesystems/dimod/blob/0c4f0c7a146cf5dfb7d6249230e65e89bac693d0/tests/requirements.txt#L6). We should test with the new version or deprecate support.