Given this structure the way the tree is held in memory will have to be changed to a map like structure, where each entry can be directly accessed to build the expansion of any termcode (node).
for Example:
309143001 would expand to: 309143001, 396526009 and 396527000
note that each "tree map" is per context and system.
The reason for this is to save space when transferring the information for the tree.
=> given multiple contexts and systems which flare has to support simultaniously it will have to create a map of tree maps,
where first the correct tree map is selected by context + system and version from the termcode
and then the code of the termcode is used to expand it within the selected tree map
note that this issue is similar to flare - as the same tree structure is used for resolving code hierachies. https://github.com/medizininformatik-initiative/flare/issues/191
The current mapping tree is based on a tree structure where each tree contains all subtrees, in essence making it a mono hierarchical structure.
As we are now expanding to more terminologies future terminologies will also include mulit-hierarchical terminologies like SNOMED.
This is why we will change the tree structure to the following:
Given this structure the way the tree is held in memory will have to be changed to a map like structure, where each entry can be directly accessed to build the expansion of any termcode (node).
for Example:
309143001 would expand to: 309143001, 396526009 and 396527000
note that each "tree map" is per context and system. The reason for this is to save space when transferring the information for the tree.
=> given multiple contexts and systems which flare has to support simultaniously it will have to create a map of tree maps,
where first the correct tree map is selected by context + system and version from the termcode and then the code of the termcode is used to expand it within the selected tree map