Entity Matching by Similarity Join
 
Loading...
Searching...
No Matches
get_true_pos.py File Reference

Namespaces

namespace  get_true_pos
 

Variables

 get_true_pos.gold = pd.read_csv("output/buffer/gold.csv")
 
 get_true_pos.ground_truth = set()
 
 get_true_pos.lid = int(row["id1"])
 
 get_true_pos.rid = int(row["id2"])
 
 get_true_pos.match_res = pd.read_csv("output/match_res/match_res.csv")
 
 get_true_pos.true_positive = pd.DataFrame(columns=match_res.columns)
 
 get_true_pos.false_negative = pd.DataFrame(columns=["_id", "ltable_id", "rtable_id", "ltable_title", "rtable_title"])
 
 get_true_pos.tableA = pd.read_csv("output/buffer/clean_A.csv")
 
 get_true_pos.tableB = pd.read_csv("output/buffer/clean_B.csv")
 
dict get_true_pos.map_A = {tableA.loc[ridx, "id"] : ridx for ridx in list(tableA.index)}
 
dict get_true_pos.map_B = {tableB.loc[ridx, "id"] : ridx for ridx in list(tableB.index)}
 
dict get_true_pos.lridx = map_A[tup[0]]
 
dict get_true_pos.rridx = map_B[tup[1]]
 
list get_true_pos.addlist = [idx, tup[0], tup[1], tableA.loc[lridx, "title"], tableB.loc[rridx, "title"]]
 
 get_true_pos.index