variable solver data concept #231
Conversation
|
will look and answer through review, thanks for getting this going! |
Consolidate, drop buildSubFG_SPECIAL
save and load to gzip as default
add getVariablePPE and deprecated
fix fatal compilation errors
dont change directory for save tar
Codecov Report
@@ Coverage Diff @@
## 4Q19/poc/v0_6 #231 +/- ##
=================================================
- Coverage 68.78% 67.15% -1.64%
=================================================
Files 39 41 +2
Lines 2079 2195 +116
=================================================
+ Hits 1430 1474 +44
- Misses 649 721 +72
Continue to review full report at Codecov.
|
| label::Symbol | ||
| timestamp::DateTime | ||
| tags::Vector{Symbol} | ||
| estimateDict::Dict{Symbol, <:AbstractPointParametricEst} |
There was a problem hiding this comment.
latest v0.5.3 now has this as ppeDict
There was a problem hiding this comment.
see here:
There was a problem hiding this comment.
will likely fall out naturally when merging all this to master, just ask to keep an eye on it when we do please.
|
|
||
| Return the estimates for a variable. | ||
| """ | ||
| getEstimates(v::VariableDataLevel1) = v.estimateDict |
There was a problem hiding this comment.
This too has been deprecated, see master branch.
| # | ||
| # buildSubgraphFromLabels!, _copyIntoGraph!, getVariableIds | ||
| # """ | ||
| # function buildSubgraphFromLabels!_SPECIAL(dfg::G, |
| """ | ||
| function getVariableSolverData(dfg::AbstractDFG, variablekey::Symbol, solvekey::Symbol=:default) | ||
| error("not implemented") | ||
| return deepcopy(solverData(getVariable(dfg, variablekey), solvekey)) |
There was a problem hiding this comment.
should verb get not return the original memory object?
There was a problem hiding this comment.
Yes, I think you are correct.
|
|
||
| #var.solverDataDict[solvekey] = vnd | ||
|
|
||
| setSolverData(var, deepcopy(vnd), key) |
There was a problem hiding this comment.
see updated api in master to include an !
|
|
||
| #var.solverDataDict[solvekey] = deepcopy(vnd) | ||
| #for InMemoryDFGTypes, cloud would update here | ||
| setSolverData(var, deepcopy(vnd), key) |
There was a problem hiding this comment.
Also, why do a deep copy here? You think the user might accidentally point the code to the same vnd object and therefore become incorrect?
There was a problem hiding this comment.
I think we should maybe leave the ‘deepcopy’ responsibility to the user.
| error("VariableNodeData '$(solvekey)' already exists") | ||
| end | ||
|
|
||
| vnd = pop!(var.solverDataDict, solvekey) |
There was a problem hiding this comment.
This one seems right to me, will think on it more.
|
Closing because of worry about merges. I will cherry pick or manually copy the relevant code. |
draft to help with the discussion in #182 concept