Research for Low-Level Storage Layout for DataFacade Implementation

[wangyoucao577]

DataFacadeProvider, ContinuousBlobAllocator, Storage

More information please go to here: #56

[CodeBear801]

https://github.com/Telenav/open-source-spec/blob/master/osrm/doc/osrm_facade.md#initialization

[CodeBear801]

facade layer is the one designed for service, so all data in all kind of model will be initialized by default.
This layer mainly designed for visiting data not updating data.

In customizer.cpp, it loads needed data to update cost metric, this step need loads and writes data, so not directly use current facade layer.

[CodeBear801]

In Customizer::Run, It will load data via

    auto graph = LoadAndUpdateEdgeExpandedGraph(
        config, mlp, node_weights, node_durations, node_distances, connectivity_checksum);

    partitioner::files::readCells(config.GetPath(".osrm.cells"), storage);

    extractor::files::readNodeData(config.GetPath(".osrm.ebg_nodes"), node_data);

    extractor::files::readProfileProperties(config.GetPath(".osrm.properties"), properties);

Then will recalculate metric by

    auto metrics = customizeFilteredMetrics(graph, storage, CellCustomizer{mlp}, filter);

And then dump result

   files::writeCellMetrics(config.GetPath(".osrm.cell_metrics"), metric_exclude_classes);
   MultiLevelEdgeBasedGraph shaved_graph{std::move(graph),
                                          std::move(node_weights),
                                          std::move(node_durations),
                                          std::move(node_distances)};
   customizer::files::writeGraph(
        config.GetPath(".osrm.mldgr"), shaved_graph, connectivity_checksum);