mirror of
https://github.com/trustgraph-ai/trustgraph.git
synced 2026-05-01 03:16:23 +02:00
Add a docker-compose for just the stores (#13)
* - Added docker-compose-storage.yaml, just the infrastructure bits - Tidied storage invocation * Util, sits on chunker output and reports histogram of chunk sizes
This commit is contained in:
parent
b0fdb4f314
commit
0e2db095e3
12 changed files with 391 additions and 121 deletions
100
tests/report-chunk-sizes
Executable file
100
tests/report-chunk-sizes
Executable file
|
|
@ -0,0 +1,100 @@
|
|||
#!/usr/bin/env python3
|
||||
|
||||
"""
|
||||
Accepts entity/vector pairs and writes them to a Milvus store.
|
||||
"""
|
||||
|
||||
from trustgraph.schema import Chunk
|
||||
from trustgraph.schema import chunk_ingest_queue
|
||||
from trustgraph.log_level import LogLevel
|
||||
from trustgraph.base import Consumer
|
||||
from threading import Thread, Lock
|
||||
import time
|
||||
|
||||
module = "test-chunk-size"
|
||||
|
||||
default_input_queue = chunk_ingest_queue
|
||||
default_subscriber = module
|
||||
default_store_uri = 'http://localhost:19530'
|
||||
|
||||
class Processor(Consumer):
|
||||
|
||||
def __init__(self, **params):
|
||||
|
||||
input_queue = params.get("input_queue", default_input_queue)
|
||||
subscriber = params.get("subscriber", default_subscriber)
|
||||
width = params.get("width", 200)
|
||||
|
||||
super(Processor, self).__init__(
|
||||
**params | {
|
||||
"input_queue": input_queue,
|
||||
"subscriber": subscriber,
|
||||
"input_schema": Chunk,
|
||||
}
|
||||
)
|
||||
|
||||
self.sizes = {}
|
||||
self.width = width
|
||||
self.lock = Lock()
|
||||
|
||||
Thread(target=self.report).start()
|
||||
|
||||
def report(self):
|
||||
|
||||
while True:
|
||||
time.sleep(1)
|
||||
|
||||
print()
|
||||
|
||||
with self.lock:
|
||||
tot = 0
|
||||
for i in range(0, 20000, self.width):
|
||||
k = (i, i + self.width)
|
||||
if k in self.sizes:
|
||||
print(f"{i:5d} ..{i+self.width:5d}: {self.sizes[k]}")
|
||||
tot += self.sizes[k]
|
||||
print(f"{'Total':13s}: {tot}")
|
||||
|
||||
|
||||
|
||||
|
||||
def handle(self, msg):
|
||||
|
||||
v = msg.value()
|
||||
|
||||
chunk = v.chunk.decode("utf-8")
|
||||
|
||||
l = len(chunk)
|
||||
|
||||
|
||||
low = int(l / self.width) * self.width
|
||||
high = low + self.width
|
||||
key = (low, high)
|
||||
|
||||
with self.lock:
|
||||
|
||||
if key not in self.sizes:
|
||||
self.sizes[key] = 0
|
||||
|
||||
self.sizes[key] += 1
|
||||
|
||||
@staticmethod
|
||||
def add_args(parser):
|
||||
|
||||
Consumer.add_args(
|
||||
parser, default_input_queue, default_subscriber,
|
||||
)
|
||||
|
||||
parser.add_argument(
|
||||
'--width',
|
||||
type=int,
|
||||
default=200,
|
||||
help=f'Histogram width (default: 200)',
|
||||
)
|
||||
|
||||
def run():
|
||||
|
||||
Processor.start(module, __doc__)
|
||||
|
||||
run()
|
||||
|
||||
Loading…
Add table
Add a link
Reference in a new issue