simjoin_entitymatching.blocker.graph.TripartiteGraph Class Reference

Public Member Functions
	__init__ (self)
	generate_signature (self, temp_rule, cur_tree_node, cur_rule_node)
	extract_rules (self, tree, node, temp_rule)
	build_graph (self, rf, feature_names, if_report=False, default_feature_names_dir="")
	update_range2 (self, idx, upper_bound)
	special_case_update_range2 (self, idx)
	special_case_sort_ranges2 (self, idx, sim_str)
	sort_ranges2 (self)
	update_range_rule_node (self)
	visualize_graph (self)
	report_ranges (self, path)
	graph_stat (self)

Public Attributes
int	INF = 100000
int	num_tree = 0
int	num_rule = 0
int	num_feature = 0
list	tree_node = []
list	rule_node = []
list	feature_node = []
list	rule_list = []
	rule_signature = defaultdict(int)
	rule_signature_trees = defaultdict(set)
list	feature_list = []
list	feature_range = []
	feature_div = np.full((100000,), 0, dtype=int)
	feature_appeared = np.full((100000,), -1, dtype=int)
	trigraph = nx.Graph()
list	bigraList1 = []
list	bigraList2 = []
list	num_tree = [[], [], []]
int	feature_node = 0)

Detailed Description

    Graph used for feature selection

Constructor & Destructor Documentation

__init__()

simjoin_entitymatching.blocker.graph.TripartiteGraph.__init__

(

self

)

Member Function Documentation

build_graph()

simjoin_entitymatching.blocker.graph.TripartiteGraph.build_graph	(		self,
			rf,
			feature_names,
			if_report = False,
			default_feature_names_dir = "" )

extract_rules()

simjoin_entitymatching.blocker.graph.TripartiteGraph.extract_rules	(		self,
			tree,
			node,
			temp_rule )

        Args:
            temp_rule: [[], [], []] feature, threshold, sign (0: <, 1: >)

generate_signature()

simjoin_entitymatching.blocker.graph.TripartiteGraph.generate_signature	(		self,
			temp_rule,
			cur_tree_node,
			cur_rule_node )

        return values: 
            the first bool: whether add tree node; 
        the second bool: whether add rule node

graph_stat()

simjoin_entitymatching.blocker.graph.TripartiteGraph.graph_stat

(

self

)

report_ranges()

simjoin_entitymatching.blocker.graph.TripartiteGraph.report_ranges	(		self,
			path )

sort_ranges2()

simjoin_entitymatching.blocker.graph.TripartiteGraph.sort_ranges2

(

self

)

        Comparing wirh sort_ranges:
            1. Drop features like 'jaro' & 'needleman'
            2. Sort p(c) in a more resonable way
            3. Do not consider interval for a feature range

special_case_sort_ranges2()

simjoin_entitymatching.blocker.graph.TripartiteGraph.special_case_sort_ranges2	(		self,
			idx,
			sim_str )

        When we use some sim funcs out of supported funcs.
        update_range2 does not support them.
            1. monge_elkan: mel
            2. needleman_wunsch: nmw
            3. smith_waterman: sw
            4. jaro: jar
            5. jaro_winkler: jwn
            6. rel_diff: rdf
            7. lev_dist: lev_dist

special_case_update_range2()

simjoin_entitymatching.blocker.graph.TripartiteGraph.special_case_update_range2	(		self,
			idx )

        For 'lev' and 'rdf', the value should be as small as possible

update_range2()

simjoin_entitymatching.blocker.graph.TripartiteGraph.update_range2	(		self,
			idx,
			upper_bound )

        Sort & update the range dict taking feature into consideration

update_range_rule_node()

simjoin_entitymatching.blocker.graph.TripartiteGraph.update_range_rule_node

(

self

)

        After sorting, we need to add rule node in tighter range to the looser one.
            e.g., (1, 0.9): [rule_id1] & (1, 0.8): [rule_id2]
            it's obviously that if (1, 0.8) cannot be satisfied, rule_id1 cannot be satisfied
            use set to avoid duplicate

visualize_graph()

simjoin_entitymatching.blocker.graph.TripartiteGraph.visualize_graph

(

self

)

Member Data Documentation

bigraList1

simjoin_entitymatching.blocker.graph.TripartiteGraph.bigraList1 = []

bigraList2

simjoin_entitymatching.blocker.graph.TripartiteGraph.bigraList2 = []

feature_appeared

simjoin_entitymatching.blocker.graph.TripartiteGraph.feature_appeared = np.full((100000,), -1, dtype=int)

feature_div

simjoin_entitymatching.blocker.graph.TripartiteGraph.feature_div = np.full((100000,), 0, dtype=int)

feature_list

simjoin_entitymatching.blocker.graph.TripartiteGraph.feature_list = []

feature_node [1/2]

list simjoin_entitymatching.blocker.graph.TripartiteGraph.feature_node = []

feature_node [2/2]

int simjoin_entitymatching.blocker.graph.TripartiteGraph.feature_node = 0)

feature_range

list simjoin_entitymatching.blocker.graph.TripartiteGraph.feature_range = []

INF

int simjoin_entitymatching.blocker.graph.TripartiteGraph.INF = 100000

num_feature

simjoin_entitymatching.blocker.graph.TripartiteGraph.num_feature = 0

num_rule

int simjoin_entitymatching.blocker.graph.TripartiteGraph.num_rule = 0

num_tree [1/2]

int simjoin_entitymatching.blocker.graph.TripartiteGraph.num_tree = 0

num_tree [2/2]

list simjoin_entitymatching.blocker.graph.TripartiteGraph.num_tree = [[], [], []]

rule_list

list simjoin_entitymatching.blocker.graph.TripartiteGraph.rule_list = []

rule_node

list simjoin_entitymatching.blocker.graph.TripartiteGraph.rule_node = []

rule_signature

simjoin_entitymatching.blocker.graph.TripartiteGraph.rule_signature = defaultdict(int)

rule_signature_trees

simjoin_entitymatching.blocker.graph.TripartiteGraph.rule_signature_trees = defaultdict(set)

tree_node

list simjoin_entitymatching.blocker.graph.TripartiteGraph.tree_node = []

trigraph

simjoin_entitymatching.blocker.graph.TripartiteGraph.trigraph = nx.Graph()

---

The documentation for this class was generated from the following file:

• simjoin_entitymatching/blocker/graph.py