data model

[DayalStrub]

• D is never actually used. columns, i.e. pd.Series, are used to create index
• R is used to determine documents to rerank, etc., thus the set of docno present in R is used as a way to subset the index, but D is not used in this way... say you have a column representing date and you want to seach a subset of the index...
• should D be thought of as core data and the index as an extension of it?