Feature/tidy kg load save (#198)

* Clean exit on ctrl-C
* More functionality in dump
* Dump some metadata
This commit is contained in:
cybermaggedon 2024-12-06 15:16:09 +00:00 committed by GitHub
parent 55c5c398b6
commit fd3db3c925
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
4 changed files with 177 additions and 37 deletions

View file

@ -16,11 +16,26 @@ import json
import sys
import argparse
import os
import signal
class Running:
def __init__(self): self.running = True
def get(self): return self.running
def stop(self): self.running = False
async def fetch_ge(running, queue, user, collection, url):
async def fetch_ge(queue, user, collection, url):
async with aiohttp.ClientSession() as session:
async with session.ws_connect(f"{url}stream/graph-embeddings") as ws:
async for msg in ws:
while running.get():
try:
msg = await asyncio.wait_for(ws.receive(), 1)
except:
continue
if msg.type == aiohttp.WSMsgType.TEXT:
data = msg.json()
@ -50,10 +65,19 @@ async def fetch_ge(queue, user, collection, url):
print("Error")
break
async def fetch_triples(queue, user, collection, url):
async def fetch_triples(running, queue, user, collection, url):
async with aiohttp.ClientSession() as session:
async with session.ws_connect(f"{url}stream/triples") as ws:
async for msg in ws:
while running.get():
try:
msg = await asyncio.wait_for(ws.receive(), 1)
except:
continue
if msg.type == aiohttp.WSMsgType.TEXT:
data = msg.json()
@ -85,27 +109,32 @@ async def fetch_triples(queue, user, collection, url):
ge_counts = 0
t_counts = 0
async def stats():
async def stats(running):
global t_counts
global ge_counts
while True:
await asyncio.sleep(5)
while running.get():
await asyncio.sleep(2)
print(
f"Graph embeddings: {ge_counts:10d} Triples: {t_counts:10d}"
)
async def output(queue, path, format):
async def output(running, queue, path, format):
global t_counts
global ge_counts
with open(path, "wb") as f:
while True:
while running.get():
msg = await queue.get()
try:
msg = await asyncio.wait_for(queue.get(), 0.5)
except TimeoutError:
continue
if format == "msgpack":
f.write(msgpack.packb(msg, use_bin_type=True))
@ -118,12 +147,15 @@ async def output(queue, path, format):
if msg[0] == "ge":
ge_counts += 1
async def run(**args):
print("Output file closed")
async def run(running, **args):
q = asyncio.Queue()
ge_task = asyncio.create_task(
fetch_ge(
running=running,
queue=q, user=args["user"], collection=args["collection"],
url=args["url"] + "api/v1/"
)
@ -131,26 +163,30 @@ async def run(**args):
triples_task = asyncio.create_task(
fetch_triples(
queue=q, user=args["user"], collection=args["collection"],
running=running, queue=q,
user=args["user"], collection=args["collection"],
url=args["url"] + "api/v1/"
)
)
output_task = asyncio.create_task(
output(
queue=q, path=args["output_file"], format=args["format"],
running=running, queue=q,
path=args["output_file"], format=args["format"],
)
)
stats_task = asyncio.create_task(stats())
stats_task = asyncio.create_task(stats(running))
await output_task
await triples_task
await ge_task
await stats_task
async def main():
print("Exiting")
async def main(running):
parser = argparse.ArgumentParser(
prog='tg-save-kg-core',
@ -193,7 +229,15 @@ async def main():
args = parser.parse_args()
await run(**vars(args))
await run(running, **vars(args))
asyncio.run(main())
running = Running()
def interrupt(sig, frame):
running.stop()
print('Interrupt')
signal.signal(signal.SIGINT, interrupt)
asyncio.run(main(running))