Safe Haskell | None |
---|
- newtype Inverted occ = Inverted {
- unInverted :: Parts occ
- type Parts occ = Map Context (Part occ)
- type Part occ = PrefixTree occ
- type Inverted0 = Inverted Occ0
- type InvertedCompressed = Inverted OccCompressed
- type InvertedSerialized = Inverted OccSerialized
- type InvertedCSerialized = Inverted OccCSerialized
- type InvertedOSerialized = Inverted OccOSerialized
- class ComprOccurrences s where
- toOccurrences :: s -> Occurrences
- fromOccurrences :: Occurrences -> s
- emptyInverted0 :: Inverted0
- emptyInvertedCompressed :: InvertedCompressed
- emptyInvertedSerialized :: InvertedSerialized
- emptyInvertedCSerialized :: InvertedCSerialized
- emptyInvertedOSerialized :: InvertedOSerialized
- class Sizeof a where
- sizeofAttrsInverted :: Sizeof i => Inverted i -> Int64
- mapOcc :: ComprOccurrences s => (Occurrences -> Occurrences) -> s -> s
- zipOcc :: ComprOccurrences s => (Occurrences -> Occurrences -> Occurrences) -> s -> s -> s
- emptyOcc :: ComprOccurrences s => s
- theOcc :: ComprOccurrences s => s -> Occurrences
- nullOcc :: ComprOccurrences s => s -> Bool
- unionOcc :: ComprOccurrences s => Occurrences -> s -> s
- diffOcc :: ComprOccurrences s => Occurrences -> s -> s
- insertPosOcc :: ComprOccurrences s => DocId -> Position -> s -> s
- deletePosOcc :: ComprOccurrences s => DocId -> Position -> s -> s
- updateDocIdOcc :: ComprOccurrences s => (DocId -> DocId) -> s -> s
- deleteDocIds :: ComprOccurrences s => Occurrences -> s -> s
- removeDocIdsInverted :: ComprOccurrences i => Occurrences -> Inverted i -> Inverted i
Documentation
The index consists of a table which maps documents to ids and a number of index parts.
Inverted | |
|
type Parts occ = Map Context (Part occ)Source
The index parts are identified by a name, which should denote the context of the words.
type Part occ = PrefixTree occSource
The index part is the real inverted index. Words are mapped to their occurrences. The part is implemented as a prefix tree
type Inverted0 = Inverted Occ0Source
The pure inverted index implemented as a prefix tree without any space optimizations. This may be taken as a reference for space and time measurements for the other index structures
type InvertedCompressed = Inverted OccCompressedSource
The inverted index with simple-9 encoding of the occurence sets
type InvertedSerialized = Inverted OccSerializedSource
The inverted index with serialized occurence maps with simple-9 encoded sets
type InvertedCSerialized = Inverted OccCSerializedSource
The inverted index with serialized occurence maps with simple-9 encoded sets and with the serialized bytestrings compressed with bzip2
type InvertedOSerialized = Inverted OccOSerializedSource
The pure inverted index with serialized occurence maps and with the serialized bytestrings compressed with bzip2, no simple-9 encoding. This is the most space efficient index of the 5 variants, even a few percent smaller then InvertedCSerialized, and a few percent faster in lookup
class ComprOccurrences s whereSource
toOccurrences :: s -> OccurrencesSource
fromOccurrences :: Occurrences -> sSource
ComprOccurrences OccOSerialized | |
ComprOccurrences OccCSerialized | |
ComprOccurrences OccSerialized | |
ComprOccurrences OccCompressed | |
ComprOccurrences Occ0 |
sizeofAttrsInverted :: Sizeof i => Inverted i -> Int64Source
mapOcc :: ComprOccurrences s => (Occurrences -> Occurrences) -> s -> sSource
zipOcc :: ComprOccurrences s => (Occurrences -> Occurrences -> Occurrences) -> s -> s -> sSource
emptyOcc :: ComprOccurrences s => sSource
theOcc :: ComprOccurrences s => s -> OccurrencesSource
nullOcc :: ComprOccurrences s => s -> BoolSource
unionOcc :: ComprOccurrences s => Occurrences -> s -> sSource
diffOcc :: ComprOccurrences s => Occurrences -> s -> sSource
insertPosOcc :: ComprOccurrences s => DocId -> Position -> s -> sSource
deletePosOcc :: ComprOccurrences s => DocId -> Position -> s -> sSource
updateDocIdOcc :: ComprOccurrences s => (DocId -> DocId) -> s -> sSource
deleteDocIds :: ComprOccurrences s => Occurrences -> s -> sSource
removeDocIdsInverted :: ComprOccurrences i => Occurrences -> Inverted i -> Inverted iSource
Remove DocIds from index