def internal_directed_cites_matrix( w ) :
"""
returns a directed citations graph; this only involves papers listed in the data, and is a DIRECTED graph adgacency matrix
such that cell (row,col) means "node row points to node col"
"""
row_dois = w.dois()
result = zeros( (len(row_dois), len(row_dois)), dtype=int8 )
for paper in w.reader() :
if paper['DI'] != '' :
for doi in cited_dois(paper) :
if row_dois.count( doi ) :
result[ row_dois.index(paper['DI']), row_dois.index(doi)] = 1
return (result, row_dois, row_dois )
def cites_matrix( w ) :
"""
Returns an occurrence matrix; the rows represent the papers in the corpus; the columns the cited papers.
The entries are the intersections where a corpus paper cites a column paper
"""
column_dois = w.set_cited_dois()
row_dois = w.dois()
result = zeros( (len(row_dois), len(column_dois)), dtype = int8 )
for paper in w.reader() :
if paper['DI'] != '' :
for doi in cited_dois( paper ) :
result[row_dois.index( paper['DI'] )][column_dois.index( doi )] = 1
return ( result, row_dois, column_dois )
def internal_directed_authorcite_matrix( w ) :
"""
returns a directed citations graph; this only involves authors
such that cell (row,col) means "node row cites node col"
"""
authordict = authors_by_doi( w )
authors = list(w.set_authors())
result = zeros( (len(authors), len(authors)), dtype=int32 )
for paper in w.reader() :
for node_author in authorlist_from_authorfield( paper['AU'] ) :
for doi in cited_dois(paper) :
if authordict.has_key(doi) :
for cited_author in authordict[doi] :
result[ authors.index( node_author ), authors.index( cited_author )] += 1
return (result, authors, authors)
