Camera Trap Images

A dataset containing one thousand images of Elephas maxims was modelled. Before transforming them into RDF graphs, image metadata was extracted and saved as CSV files. The RDF dataset has a unique path that points to the location of the image on a protected cloud server. The figure below depicts the dataset entities and the output of the semantic modelling. Data are accessable on Github (https://github.com/Naeima/Forest-Observatory-Ontology/tree/main/Camera Trap images metadata).

!pip install rdflib
     
from google.colab import files
import pandas as pd
from rdflib import Graph, Namespace, URIRef, Literal
from rdflib.namespace import RDF, XSD
import csv

# Prompt for file upload
uploaded = files.upload()

# Read the uploaded CSV file into a pandas DataFrame
for filename in uploaded.keys():
    print('User uploaded file "{name}" with length {length} bytes'.format(
        name=filename, length=len(uploaded[filename])))
    df = pd.read_csv(filename)
    print(df.head())
     
# Define namespaces
FOO = Namespace("https://w3id.org/def/foo#")
SOSA = Namespace("http://www.w3.org/ns/sosa/")
GEO = Namespace("http://www.w3.org/2003/01/geo/wgs84_pos#")

# Create a new graph
g = Graph()

# Bind namespaces
g.bind("foo", FOO)
g.bind("sosa", SOSA)
g.bind("geo", GEO)

# Path to your CSV file
csv_file_path = filename  # Adjusted to dynamically handle the uploaded CSV file

# Read and process data from the image CSV
with open(csv_file_path, newline='') as csvfile:
    reader = csv.DictReader(csvfile, delimiter=',')

    # Print headers for debugging
    print("Headers:", reader.fieldnames)

    for row in reader:
        image_id = row.get('Image_ID', '').strip()  # Assuming there's an 'Image_ID' field
        timestamp = row.get('Timestamp', '').strip()  # Assuming there's a 'Timestamp' field
        camera_id = row.get('Camera_ID', '').strip()  # Assuming there's a 'Camera_ID' field
        location = row.get('Location', '').strip()  # Assuming there's a 'Location' field
        species = row.get('Species', '').strip()  # Assuming there's a 'Species' field

        # Create an observation URI for the image
        observation = URIRef(f"https://w3id.org/def/foo#image_{image_id}")
        g.add((observation, RDF.type, FOO.imageObservation))
        g.add((observation, FOO.imageID, Literal(image_id, datatype=XSD.string)))
        g.add((observation, FOO.timestamp, Literal(timestamp, datatype=XSD.dateTime)))
        g.add((observation, FOO.cameraID, Literal(camera_id, datatype=XSD.string)))
        g.add((observation, FOO.location, Literal(location, datatype=XSD.string)))
        if species:
            g.add((observation, FOO.observes, Literal(species, datatype=XSD.string)))

# Serialize the graph to a file
output_file = "image_metadata_knowledge_graph.ttl"
g.serialize(destination=output_file, format="turtle")

print(f"Knowledge graph has been serialized to {output_file}")     

# Download the RDF Turtle file
from google.colab import files
files.download(output_file)
     
```python
# adding serialized data to stardog 

# conn_details = {
#   'endpoint': 'http://localhost:5820',
#   'username': 'admin',
#   'password': 'admin'
# }
# with stardog.Admin(**conn_details) as admin:
#     images = admin.new_database('images')

# conn = stardog.Connection('images', **conn_details)

# conn.begin()

# conn.add(
#     stardog.content.File('images.rdf', stardog.content_types.TURTLE),
# )

# conn.commit()