Conversion into Gremlin format csv

Python Lambda function is the fastest way I can think of. In pseudo code I would have it as:

FUNCTION lambda_handler(event, context):
    INPUT: 
        - event: S3 event triggered by object creation in a bucket
        - context: Lambda execution context
        
    1. Extract the bucket name and key from the S3 event.
    2. Download the CSV file from the specified bucket and key to the Lambda execution environment.
    3. Process the CSV file to identify node properties and edges.
    4. Write the node properties to a CSV file named "nodes.csv" and edges to a CSV file named "edges.csv".
    5. Upload the "nodes.csv" and "edges.csv" files to the same S3 bucket.
    
END FUNCTION

Uf you are happy with pseudo definitions above, you can have two separate csv files with the following (Check the bucket name):

import csv
import os
import boto3

s3 = boto3.client('s3')

def lambda_handler(event, context):
    # Get the bucket and key from the S3 event
    bucket = event['Records'][0]['s3']['bucket']['name']
    key = event['Records'][0]['s3']['object']['key']
    
    # Download the CSV file from S3 to the Lambda execution environment
    download_path = '/tmp/input.csv'
    s3.download_file(bucket, key, download_path)
    
    # Process the CSV file
    node_properties, edges = process_csv_file(download_path)
    
    # Write nodes and edges to CSV files
    write_nodes(node_properties)
    write_edges(edges)
    
    # Upload the output files to S3
    upload_file_to_s3('nodes.csv', bucket)
    upload_file_to_s3('edges.csv', bucket)

def process_csv_file(input_file):
    node_properties = set()
    edges = []
    with open(input_file, 'r', newline='') as csvfile:
        reader = csv.DictReader(csvfile)
        for row in reader:
            for key, value in row.items():
                if key != 'source' and key != 'target' and key != 'relationship':
                    node_properties.add((key, value))
                if 'source' in row and 'target' in row and 'relationship' in row:
                    edges.append((row['source'], row['target'], row['relationship']))
    return node_properties, edges

def write_nodes(node_properties):
    with open('/tmp/nodes.csv', 'w', newline='') as csvfile:
        writer = csv.writer(csvfile)
        writer.writerow(['property', 'value'])
        for property, value in node_properties:
            writer.writerow([property, value])

def write_edges(edges):
    with open('/tmp/edges.csv', 'w', newline='') as csvfile:
        writer = csv.writer(csvfile)
        writer.writerow(['source', 'target', 'relationship'])
        for edge in edges:
            writer.writerow(edge)

def upload_file_to_s3(file_name, bucket_name):
    s3.upload_file('/tmp/' + file_name, bucket_name, file_name)

Then you can load them in Neptune with (Check the buckets name):

neptune-loader load --iam-role-arn arn:aws:iam::123456789012:role/NeptuneBulkLoadRole --url https://neptune-cluster-url --files s3://mybucket/nodes.csv, s3://mybucket/edges.csv