Extracting Data Records from Query Result Pages Based on Visual Features - Advances in Databases - page 150

Algorithm 1. Grouping Data Units
Input: a set of query terms T , a data section block B
Output: a set of data records R
1: Set R , a set of leaf nodes N l , a set of starting leaf nodes N s , a set of data unit
groups G ,asetofleafnodegroups G l , and a set of horizontally expanded data
unit groups G all to {}
2: Add every text node in B to N l
3: for every leaf node n l ∈ N l do
4: if n l contains a query term t ∈ T then
5: Add n l to N s
6: Remove n l from N l
7: for every starting leaf node n s ∈ N s do
8: Set a data unit group g to {n s }

n l ∈ N l do
10: if n l is horizontally aligned with n s then
11: Add n l to g
12: Remove n l from N l
13: Add g to G
14: repeat
15: Remove a leaf node n l from N l
16: Set a leaf node group g l = {n l }
17: for each leaf node n l ∈ N l do
18: if n l is horizontally aligned with n l then
19: Add n l to g l
20: Remove n l from N l
21: Add g l to G l
22: until N l = {}
23: repeat
24: Remove a data unit group g from G
25: for each data unit group g ∈ G do
26: if g is horizontally aligned with g then
27: Set g to g ∪ g
28: Remove g from G
29: Add g to G
30: until G = {}
31: repeat
32:

9: for every leaf node

Remove a horizontally expanded data unit group g from G

33: for each horizontally expanded data unit group g ∈ G do

34: if g is vertically adjacent to g then

35: Set g to g ∪ g
36: Remove g from G
37: for each leaf node group g l ∈ G l do
38: if g l is vertically adjacent to g then
39: Set g to g ∪ g l
40: Remove g l from G l
41: Add g to R
42: until G = {}
43: Return R