pyJedAI Spatial Component

In this notebook we present the pyJedAI Geospatial Interlinking functionality.

!pip install pyjedai:0.1.4

^C

!python --version

Python 3.9.16

Libraries injection

import csv
import sys
import os

from pyjedai.datamodel import SpatialData
from pyjedai.spatial.filtering import StandardSpatialFiltering
from pyjedai.spatial.initialization import StandardSpatialInitialization
from pyjedai.spatial.verification import StandardSpatialVerification

Data reading and pyJedAI-formating

data_dir = '../data/spatial/'

maxInt = sys.maxsize
while True:
    try:
        csv.field_size_limit(maxInt)
        break
    except OverflowError:
        maxInt = int(maxInt/10)

path1 = os.path.join(data_dir,"regions_gr.csv")
path2 = os.path.join(data_dir,"wildlife_sanctuaries.csv")

delimiter1 = "\t"
delimiter2 = "\t"

# Read the CSV files
file1 = open(path1, newline='', encoding='utf-8')
file2 = open(path2, newline='', encoding='utf-8')

dataset_1 = csv.reader(file1, delimiter=delimiter1)
dataset_2 = csv.reader(file2, delimiter=delimiter1)

spatial_data = \
    SpatialData(source_reader=dataset_1,
                source_delimiter="\t",
                target_reader=dataset_2, target_delimiter="\t",
                skip_header=True)

Spatial Filtering

SSF = StandardSpatialFiltering()
spatial_index, theta_x, theta_y = SSF.process(spatial_data)

Spatial Initialization

SSI = StandardSpatialInitialization(budget=1000, wScheme="MBR")
priority_pairs = SSI.process(spatial_data, spatial_index, theta_x=theta_x, theta_y=theta_y)

Spatial Verification

SSV = StandardSpatialVerification(q_pairs=100)
SSV.process(spatial_data=spatial_data, priority_pairs=priority_pairs)
SSV.evaluate()

───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
Performance:
	Precision:      0.72% 
	Recall:         6.48%
	F1-score:       1.29%
	Progressive Geometry Recall:      2.33%
───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────

{'Precision:': 0.7192008879023307,
 'Recall:': 6.48,
 'F1-score:': 1.2947052947052946,
 'Progressivr geometry recall:': 2.3338068812430635}