Entities keyworded DataspectsSearch[edit source]
|BusinessService "Run your organizational Find-And-Learn network!"|
|DataspectsSearch Terms||Mentioned In|
Story[edit | edit source]
- A MongoDB declares ResourceSilo and Resource collections.
- Feeding: resources are fed to Resource collections by being modeled according to their SourceDataModel obtained when extracting Resources from ResourceSilos.
- See https://search.dataspects.com/Specifies::SourceDataModel for SourceDataModels and corresponding MongoDocs that are fed to a MongoDB.
- Indexing: resources are compiled into one or multiple ES documents (SearchDataModels), which represent individual ES hits resulting from a query.
- For the purpose of minimizing signal discordance, ES documents may substantially alter and enrich source data models.
- ES documents are stored to the general index and search context indeces featuring pertinent settings and mappings again to minimize signal discordance stemming from SourceDataModels.
- IndexSettings: implement custom analyzers.
- IndexMappings: implement custom fields specific to the domain and search contexts.
- Searching: the initial query is run against the GeneralIndex and results are immediately returned.
- In parallel the initial query is run against SearchContextIndices and based on some metric (e.g. distinctly high scores) one or more SearchContext(s) are suggested as possibly representing the corresponding search user's truly intended/expected search context.
- The user can switch SearchContexts at will. In this case, the corresponding SearchContext index will get boosted.
- Search is performed on direct fields and nested fields (annotations) by inner hits.
- Since a general or SearchContext search runs several queries it is valuable to know which queries matched by using named queries/filters.
- Documents are returned by query and post_filter phases.