Fix errors detected by Ruff (#3918)

### What problem does this PR solve?

Fix errors detected by Ruff

### Type of change

- [x] Refactoring
This commit is contained in:
Zhichang Yu
2024-12-08 14:21:12 +08:00
committed by GitHub
parent e267a026f3
commit 0d68a6cd1b
97 changed files with 2558 additions and 1976 deletions

View File

@ -33,14 +33,16 @@ def collect():
def main():
locations = collect()
if not locations:return
if not locations:
return
logging.info(f"TASKS: {len(locations)}")
for kb_id, loc in locations:
try:
if REDIS_CONN.is_alive():
try:
key = "{}/{}".format(kb_id, loc)
if REDIS_CONN.exist(key):continue
if REDIS_CONN.exist(key):
continue
file_bin = STORAGE_IMPL.get(kb_id, loc)
REDIS_CONN.transaction(key, file_bin, 12 * 60)
logging.info("CACHE: {}".format(loc))

View File

@ -23,18 +23,12 @@ import os
from api.utils.log_utils import initRootLogger
CONSUMER_NO = "0" if len(sys.argv) < 2 else sys.argv[1]
CONSUMER_NAME = "task_executor_" + CONSUMER_NO
LOG_LEVELS = os.environ.get("LOG_LEVELS", "")
initRootLogger(CONSUMER_NAME, LOG_LEVELS)
from datetime import datetime
import json
import os
import hashlib
import copy
import re
import sys
import time
import threading
from functools import partial
@ -63,6 +57,11 @@ from rag.utils import rmSpace, num_tokens_from_string
from rag.utils.redis_conn import REDIS_CONN, Payload
from rag.utils.storage_factory import STORAGE_IMPL
CONSUMER_NO = "0" if len(sys.argv) < 2 else sys.argv[1]
CONSUMER_NAME = "task_executor_" + CONSUMER_NO
LOG_LEVELS = os.environ.get("LOG_LEVELS", "")
initRootLogger(CONSUMER_NAME, LOG_LEVELS)
BATCH_SIZE = 64
FACTORY = {
@ -201,7 +200,8 @@ def build_chunks(task, progress_callback):
"doc_id": task["doc_id"],
"kb_id": str(task["kb_id"])
}
if task["pagerank"]: doc["pagerank_fea"] = int(task["pagerank"])
if task["pagerank"]:
doc["pagerank_fea"] = int(task["pagerank"])
el = 0
for ck in cks:
d = copy.deepcopy(doc)
@ -342,7 +342,8 @@ def run_raptor(row, chat_mdl, embd_mdl, callback=None):
"docnm_kwd": row["name"],
"title_tks": rag_tokenizer.tokenize(row["name"])
}
if row["pagerank"]: doc["pagerank_fea"] = int(row["pagerank"])
if row["pagerank"]:
doc["pagerank_fea"] = int(row["pagerank"])
res = []
tk_count = 0
for content, vctr in chunks[original_length:]: