Algebra.Graph.IO.Datasets.LINQS.Citeseer

Citeseer document classification dataset, from :

Qing Lu, and Lise Getoor. "Link-based classification." ICML, 2003.

https://linqs.soe.ucsc.edu/data

1. Download the dataset

stash

Arguments

:: FilePath

directory where the data files will be saved

-> IO ()

Download, parse, serialize and save the dataset to local storage

2. Reconstruct the citation graph

citeseerGraph

Arguments

:: FilePath

directory where the data files were saved

-> IO (Graph ContentRow)

Reconstruct the citation graph

NB : relies on the user having stashed the dataset to local disk first.

citeseerGraphEdges

Arguments

:: (MonadResource m, MonadThrow m)

=> FilePath

directory of data files

-> Map String (Seq Int16, DocClass)

content data

-> ConduitT i (Maybe (Graph ContentRow)) m ()

Stream out the edges of the citation graph, in which the nodes are decorated with the document metadata.

The full citation graph can be reconstructed by folding over this stream and overlaying the graph edges as they arrive.

This way the graph can be partitioned in training , test and validation subsets at the usage site

restoreContent

Arguments

:: FilePath

directory where the data files are saved

-> IO (Map String (Seq Int16, DocClass))

Load the graph node data from local storage

Types

data ContentRow

Dataset row of the .content file

The .content file contains descriptions of the papers in the following format:

<paper_id> <word_attributes> <class_label>

The first entry in each line contains the unique string ID of the paper followed by binary values indicating whether each word in the vocabulary is present (indicated by 1) or absent (indicated by 0) in the paper (vocabulary : 3703 unique words). Finally, the last entry in the line contains the class label of the paper.

Constructors

CRow

Fields

• crId :: String

  identifier

• crFeatures :: Seq Int16

  features, in sparse format (without the zeros)

• crClass :: DocClass

  document class label

Instances

Eq ContentRow

Defined in Algebra.Graph.IO.Datasets.LINQS.Citeseer

Methods

(==) :: ContentRow -> ContentRow -> Bool

(/=) :: ContentRow -> ContentRow -> Bool

Ord ContentRow

Defined in Algebra.Graph.IO.Datasets.LINQS.Citeseer

Methods

compare :: ContentRow -> ContentRow -> Ordering

(<) :: ContentRow -> ContentRow -> Bool

(<=) :: ContentRow -> ContentRow -> Bool

(>) :: ContentRow -> ContentRow -> Bool

(>=) :: ContentRow -> ContentRow -> Bool

max :: ContentRow -> ContentRow -> ContentRow

min :: ContentRow -> ContentRow -> ContentRow

Show ContentRow

Defined in Algebra.Graph.IO.Datasets.LINQS.Citeseer

Methods

showsPrec :: Int -> ContentRow -> ShowS

show :: ContentRow -> String

showList :: [ContentRow] -> ShowS

Generic ContentRow

Defined in Algebra.Graph.IO.Datasets.LINQS.Citeseer

Associated Types

type Rep ContentRow :: Type -> Type

Methods

from :: ContentRow -> Rep ContentRow x

to :: Rep ContentRow x -> ContentRow

Binary ContentRow

Defined in Algebra.Graph.IO.Datasets.LINQS.Citeseer

Methods

put :: ContentRow -> Put

get :: Get ContentRow

putList :: [ContentRow] -> Put

type Rep ContentRow

Defined in Algebra.Graph.IO.Datasets.LINQS.Citeseer

type Rep ContentRow = D1 ('MetaData "ContentRow" "Algebra.Graph.IO.Datasets.LINQS.Citeseer" "algebraic-graphs-io-0.2-HM3hsaOKtKl5JJVlqGeywp" 'False) (C1 ('MetaCons "CRow" 'PrefixI 'True) (S1 ('MetaSel ('Just "crId") 'NoSourceUnpackedness 'NoSourceStrictness 'DecidedLazy) (Rec0 String) :*: (S1 ('MetaSel ('Just "crFeatures") 'NoSourceUnpackedness 'NoSourceStrictness 'DecidedLazy) (Rec0 (Seq Int16)) :*: S1 ('MetaSel ('Just "crClass") 'NoSourceUnpackedness 'NoSourceStrictness 'DecidedLazy) (Rec0 DocClass))))

data DocClass

document classes of the Citeseer dataset

Constructors

Agents

AI

DB

IR

ML

HCI

Instances

Enum DocClass

Defined in Algebra.Graph.IO.Datasets.LINQS.Citeseer

Methods

succ :: DocClass -> DocClass

pred :: DocClass -> DocClass

toEnum :: Int -> DocClass

fromEnum :: DocClass -> Int

enumFrom :: DocClass -> [DocClass]

enumFromThen :: DocClass -> DocClass -> [DocClass]

enumFromTo :: DocClass -> DocClass -> [DocClass]

enumFromThenTo :: DocClass -> DocClass -> DocClass -> [DocClass]

Eq DocClass

Defined in Algebra.Graph.IO.Datasets.LINQS.Citeseer

Methods

(==) :: DocClass -> DocClass -> Bool

(/=) :: DocClass -> DocClass -> Bool

Ord DocClass

Defined in Algebra.Graph.IO.Datasets.LINQS.Citeseer

Methods

compare :: DocClass -> DocClass -> Ordering

(<) :: DocClass -> DocClass -> Bool

(<=) :: DocClass -> DocClass -> Bool

(>) :: DocClass -> DocClass -> Bool

(>=) :: DocClass -> DocClass -> Bool

max :: DocClass -> DocClass -> DocClass

min :: DocClass -> DocClass -> DocClass

Show DocClass

Defined in Algebra.Graph.IO.Datasets.LINQS.Citeseer

Methods

showsPrec :: Int -> DocClass -> ShowS

show :: DocClass -> String

showList :: [DocClass] -> ShowS

Generic DocClass

Defined in Algebra.Graph.IO.Datasets.LINQS.Citeseer

Associated Types

type Rep DocClass :: Type -> Type

Methods

from :: DocClass -> Rep DocClass x

to :: Rep DocClass x -> DocClass

Binary DocClass

Defined in Algebra.Graph.IO.Datasets.LINQS.Citeseer

Methods

put :: DocClass -> Put

get :: Get DocClass

putList :: [DocClass] -> Put

type Rep DocClass

Defined in Algebra.Graph.IO.Datasets.LINQS.Citeseer

type Rep DocClass = D1 ('MetaData "DocClass" "Algebra.Graph.IO.Datasets.LINQS.Citeseer" "algebraic-graphs-io-0.2-HM3hsaOKtKl5JJVlqGeywp" 'False) ((C1 ('MetaCons "Agents" 'PrefixI 'False) (U1 :: Type -> Type) :+: (C1 ('MetaCons "AI" 'PrefixI 'False) (U1 :: Type -> Type) :+: C1 ('MetaCons "DB" 'PrefixI 'False) (U1 :: Type -> Type))) :+: (C1 ('MetaCons "IR" 'PrefixI 'False) (U1 :: Type -> Type) :+: (C1 ('MetaCons "ML" 'PrefixI 'False) (U1 :: Type -> Type) :+: C1 ('MetaCons "HCI" 'PrefixI 'False) (U1 :: Type -> Type))))