resources.py

#!/usr/bin/env python
#coding=utf-8

'''
Definition of Resources class. Used inside collect.py, this stores
additional information of the training data such as the 
token-to-subgraph phrasetable, the list of named entities
etc.

@author: Marco Damonte (m.damonte@sms.ed.ac.uk)
@since: 03-10-16
'''

from collections import defaultdict
import cPickle as pickle
import operator
from subgraph import Subgraph
from node import Node
from variables import Variables

class Resources:

    @staticmethod
    def store_table(model_dir):

        table = {}
        freq = {}
        print "Storing tables.."
        print "Number of tokens:", len(Resources.phrasetable.keys())
        for i, token in enumerate(Resources.phrasetable):
            if i % 100 == 0:
                print "Token:", i
            sg = max(Resources.phrasetable[token].iteritems(), key=operator.itemgetter(1))[0]
            table[token] = sg

        pickle.dump(table, open(model_dir + "/phrasetable.p", "wb"))

    @staticmethod
    def init_table(model_dir, empty = True):
        Resources.phrasetable = defaultdict(lambda : defaultdict(int))

        if empty == False:
            Resources.phrasetable = pickle.load(open(model_dir + "/phrasetable.p", "rb"))
            Resources.organizations = defaultdict(list)
            for line in open(model_dir + "/organizations.txt"):
                fields = line.strip().split()
                if len(fields) > 1:
                    Resources.organizations[fields[0]] = fields[1]
        else:
            Resources.organizations = None
            Resources.seen_org = []
            Resources.forg = open(model_dir + "/organizations.txt","w")

        # Resources.verbalization_list = {}
        # for line in open("resources/verbalization-list-v1.06.txt"):
        #    line = line.strip().split()
        #    if line[0] == "VERBALIZE":
        #        var = Variables()
        #        nodes = []
        #        ntop = Node(None, var.nextVar(), line[3], False)
        #        nodes.append(ntop)
        #        relations = []
        #        fields = line[4:]
        #        for i in range(0,len(fields),2):
        #            if fields[i + 1] == "-":
        #                n = Node(None, '-', "", True)
        #            else:
        #                n = Node(None, var.nextVar(), fields[i + 1], False)
        #            nodes.append(n)
        #            relations.append((ntop,n,fields[i]))
        #        Resources.verbalization_list[line[1]] = Subgraph(nodes, relations)
        # for line in open("resources/have-org-role-91-roles-v1.06.txt"):
        #    line = line.strip().split()
        #    if line[0] == "USE-HAVE-ORG-ROLE-91-ARG2":
        #        var = Variables()
        #        ntop = Node(None, var.nextVar(), "have-org-role-91", False)
        #        node = Node(None, var.nextVar(), line[1], False)
        #        Resources.verbalization_list[line[1]] = Subgraph([ntop, node], [(ntop, node, ":ARG2")])

        # for line in open("resources/have-rel-role-91-roles-v1.06.txt"):
        #    if "#" in line:
        #        line = line.split("#")[0]
        #    line = line.strip().split()
        #    if len(line) > 0 and line[0] == "USE-HAVE-REL-ROLE-91-ARG2":
        #        var = Variables()
        #        if len(line) >= 3 and line[2] == ":standard":
        #            ntop = Node(None, var.nextVar(), "have-rel-role-91", False)
        #            node = Node(None, var.nextVar(), line[3], False)
        #            Resources.verbalization_list[line[1]] = Subgraph([ntop, node], [(ntop, node, ":ARG2")])
        #            Resources.verbalization_list[line[3]] = Subgraph([ntop, node], [(ntop, node, ":ARG2")])
        #        else:
        #            ntop = Node(None, var.nextVar(), "have-rel-role-91", False)
        #            node = Node(None, var.nextVar(), line[1], False)
        #            Resources.verbalization_list[line[1]] = Subgraph([ntop, node], [(ntop, node, ":ARG2")])