#!/usr/bin/env python

# works2matrix.py - given a directory of works and sources, output a matrix of bibliometrics

# Eric Lease Morgan <emorgan@nd.edu>
# (c) University of Notre Dame; distributed under a GNU Public License

# January 24, 2024 - first investigations


# configure
PATTERN = '*.json'
WORK    = 'https://doi.org/'
SOURCE  = 'https://openalex.org/'
COLUMNS = [ 'doi', 'title', 'cited', 'journal', 'hindex', 'i10index', 'abstract' ]

# require
from   json     import loads
from   pathlib  import Path
import requests
import sys
import pandas

# get input
if len( sys.argv ) != 3 : sys.exit( 'Usage: ' + sys.argv[ 0 ] + " <directory> <directory>" )
works   = sys.argv[ 1 ]
sources = sys.argv[ 2 ]

# initialize
works   = Path( works )
sources = Path( sources )

# process each source; create an index of sources ("Cool!")
index = {}
for source in sources.glob( PATTERN ) :

	# re-initialize
	identifier = None
	hindex     = None
	i10index   = None

	# open the given work and parse the easy stuff
	with open( source ) as handle : record = loads( handle.read() )
	id       = str( record[ 'id' ].replace( SOURCE, '' ) )
	hindex   = str( record[ 'summary_stats' ][ 'h_index' ] )
	i10index = str( record[ 'summary_stats' ][ 'i10_index' ] )

	# update
	index[ id ] = { 'hindex':hindex, 'i10index': i10index }

# get and process each work; create a matrix (list) of records
matrix = []
for work in works.glob( PATTERN ) :
		
	# re-initialize
	doi          = None
	title        = None
	citedby      = None
	journalTitle = None
	source       = None
	hindex       = None
	i10index     = None
	abstract     = None
	
	# open the given work and parse the easy stuff
	with open( work ) as handle : record = loads( handle.read() )
	doi     = record[ 'doi' ].replace( WORK, '' )
	title   = record[ 'title' ].title()
	citedby = str( record[ 'cited_by_count' ] )
	
	if record[ 'abstract_inverted_index' ] : abstract = ' '.join( list( ( record[ 'abstract_inverted_index' ]).keys() ) )
	
	# try to parse the harder stuff
	try :
	
		journal = record[ 'primary_location' ][ 'source' ][ 'display_name' ]
		source  = record[ 'primary_location' ][ 'source' ][ 'id' ].replace( SOURCE, '' )
		
	# source not found (probably)
	except : pass
	
	# check for source
	if source :
	
		# look up impact factors
		hindex   = index[ source ][ 'hindex' ]
		i10index = index[ source ][ 'i10index' ]
		
	# debug
	sys.stderr.write( '       doi: ' + doi             + '\n' )
	sys.stderr.write( '     title: ' + title           + '\n' )
	sys.stderr.write( '  cited by: ' + citedby         + '\n' )
	sys.stderr.write( '   journal: ' + str( journal )  + '\n' )
	sys.stderr.write( '    hindex: ' + str( hindex )   + '\n' )
	sys.stderr.write( '  i10index: ' + str( i10index ) + '\n' )
	sys.stderr.write( '\n' )
	
	# update
	matrix.append( [ doi, title, citedby, journal, hindex, i10index, abstract ] )

# create a dataframe, output, and done
matrix = pandas.DataFrame( matrix, columns=COLUMNS )
print( matrix.to_csv( index=False ) )
exit()