trustgraph/trustgraph-parquet/scripts/load-triples
cybermaggedon 9ebfe0d94a
Maint/knowledge load collections (#132)
* Add user/collection support to knowledge loaders

* Fix timeouts
2024-10-31 16:28:03 +00:00

180 lines
4.2 KiB
Python
Executable file

#!/usr/bin/env python3
"""
Loads Graph embeddings into TrustGraph processing.
"""
import pulsar
from pulsar.schema import JsonSchema
from trustgraph.schema import Triples, Triple, Value, Metadata
from trustgraph.schema import triples_store_queue
import argparse
import os
import time
import pyarrow as pa
import pyarrow.parquet as pq
from trustgraph.log_level import LogLevel
class Loader:
def __init__(
self,
pulsar_host,
output_queue,
log_level,
file,
user,
collection,
):
self.client = pulsar.Client(
pulsar_host,
logger=pulsar.ConsoleLogger(log_level.to_pulsar())
)
self.producer = self.client.create_producer(
topic=output_queue,
schema=JsonSchema(Triples),
chunking_enabled=True,
)
self.file = file
self.user = user
self.collection = collection
def run(self):
try:
path = self.file
print("Reading file...")
table = pq.read_table(path)
print("Loaded.")
names = set(table.column_names)
if "s" not in names:
print("No 's' column")
if "p" not in names:
print("No 'p' column")
if "o" not in names:
print("No 'o' column")
sc = table.column("s")
pc = table.column("p")
oc = table.column("o")
for s, p, o in zip(sc, pc, oc):
r = Triples(
metadata=Metadata(
metadata=[],
user=self.user,
collection=self.collection,
),
triples=[
Triple(
s=Value(
value=s.as_py(), is_uri=True
),
p=Value(
value=p.as_py(), is_uri=True
),
o=Value(
value=o.as_py(),
is_uri=o.as_py().startswith("https:")
)
)
]
)
self.producer.send(r)
except Exception as e:
print(e, flush=True)
def __del__(self):
self.client.close()
def main():
parser = argparse.ArgumentParser(
prog='loader',
description=__doc__,
)
default_pulsar_host = os.getenv("PULSAR_HOST", 'pulsar://localhost:6650')
default_output_queue = triples_store_queue
default_user = 'trustgraph'
default_collection = 'default'
parser.add_argument(
'-p', '--pulsar-host',
default=default_pulsar_host,
help=f'Pulsar host (default: {default_pulsar_host})',
)
parser.add_argument(
'-o', '--output-queue',
default=default_output_queue,
help=f'Output queue (default: {default_output_queue})'
)
parser.add_argument(
'-u', '--user',
default=default_user,
help=f'User ID (default: {default_user})'
)
parser.add_argument(
'-c', '--collection',
default=default_collection,
help=f'Collection ID (default: {default_collection})'
)
parser.add_argument(
'-l', '--log-level',
type=LogLevel,
default=LogLevel.ERROR,
choices=list(LogLevel),
help=f'Output queue (default: info)'
)
parser.add_argument(
'-f', '--file',
required=True,
help=f'File to load'
)
args = parser.parse_args()
while True:
try:
p = Loader(
pulsar_host=args.pulsar_host,
output_queue=args.output_queue,
log_level=args.log_level,
file=args.file,
user=args.user,
collection=args.collection,
)
p.run()
print("File loaded.")
break
except Exception as e:
print("Exception:", e, flush=True)
print("Will retry...", flush=True)
time.sleep(10)
main()