As a user of crossmap I need to be able to do crossmap from several sources

GlobalNamesArchitecture / gnresolver

MIT License

0 stars 0 forks source link

As a user of crossmap I need to be able to do crossmap from several sources #70

Closed dimus closed 7 years ago

dimus commented 7 years ago

We need our system to support more than one source for crossmapping. When people do crossmap -- they enter

resolution data source (Open tree, or Encyclopedia of life)
original data source (which has local ids)
target data source
one or several local IDs from original data source (probably we can make limit pretty high 50k?)

output is IDs from target data source

As a first step we do not support jumping from one resolution source to another, it is a future user story.

dimus commented 7 years ago

cross_maps table should have the following fields:

data_source_id | name_string_id | taxon_id | cm_data_source_id | cm_local_id

(data_source_id, name_string_id, taxon_id) form a composite primary key and are identical to the key from name_string_indices. The order in the key is important. Number of composite keys for open_tree_taxonomy should be the same in composite_maps and in name_string_indices.

we can try to make a "search" key as a composite key as well -- (cm_local_id, cm_data_source_id, data_source_id)

alexander-myltsev commented 7 years ago

Hm, can it be -- two local ids for the same data_source_id, name_string_id, taxon_id:

179 00026087-718e-5616-97f2-9e3780ff3f7c    578296  11  29
179 00026087-718e-5616-97f2-9e3780ff3f7c    578296  11  302

dimus commented 7 years ago

Hmm, looks like unresolved homonyms to me -- they did not know which of the homonyms to pick. Interesting to see what do we have in their raw data