You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
158 lines
6.6 KiB
158 lines
6.6 KiB
# coding:utf8
|
|
import os, sys
|
|
import io
|
|
from jsonpath_ng import jsonpath, parse
|
|
import uuid
|
|
sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding='utf8')
|
|
cur_dir = os.path.dirname(os.path.abspath(__file__)) or os.getcwd()
|
|
par_dir = os.path.abspath(os.path.join(cur_dir, os.path.pardir))
|
|
sys.path.append(cur_dir)
|
|
sys.path.append(par_dir)
|
|
import json
|
|
from django.http import HttpResponse
|
|
from text_analysis.tools import to_kafka
|
|
from django.views.decorators.csrf import csrf_exempt
|
|
from log_util.set_logger import set_logger
|
|
|
|
logging = set_logger('logs/results.log')
|
|
import traceback
|
|
import queue
|
|
import requests
|
|
from text_analysis.tools.tool import parse_data
|
|
from text_analysis.chroma1 import LangChainChroma
|
|
import time
|
|
from kazoo.client import KazooClient
|
|
from kazoo.protocol.states import EventType
|
|
import queue
|
|
task_queue = queue.PriorityQueue()
|
|
stop_dict={}
|
|
from text_analysis.read_config import load_config
|
|
config=load_config()
|
|
|
|
|
|
@csrf_exempt
|
|
def createChroma(request):
|
|
if request.method == 'POST':
|
|
try:
|
|
raw_data = json.loads(request.body)
|
|
if "trace" in raw_data.keys() and raw_data["trace"]==True:
|
|
task_queue.put((-1, time.time(),raw_data))
|
|
else:
|
|
task_queue.put((1, time.time(),raw_data))
|
|
return HttpResponse(json.dumps({"code": 1, "msg": "请求正常！"}, ensure_ascii=False))
|
|
except:
|
|
logging.error(traceback.format_exc())
|
|
return HttpResponse(json.dumps({"code": 0, "msg": "请求json格式不正确！"}, ensure_ascii=False))
|
|
else:
|
|
return HttpResponse(json.dumps({"code": 0, "msg": "请求方式错误，改为post请求"}, ensure_ascii=False))
|
|
|
|
|
|
def upload():
|
|
while True:
|
|
try:
|
|
if task_queue.qsize()>0:
|
|
p,t,raw_data = task_queue.get(timeout=1)
|
|
logging.info("当前任务队列长度{}".format(task_queue.qsize()+1))
|
|
output=raw_data["output"]
|
|
res_tmp={key: "" for key in output}
|
|
if "id" in res_tmp.keys():
|
|
res_tmp["id"]=str(uuid.uuid4())
|
|
res_tmp["isLast"]=1
|
|
task_id=raw_data["scenes_id"]
|
|
task_version=raw_data["version"]
|
|
logging.info("任务数据为：{}".format(raw_data))
|
|
logging.info("当前version信息为：{}".format(stop_dict))
|
|
if task_id in stop_dict.keys() and task_version!=stop_dict[task_id]["version"]:
|
|
logging.info("已暂停任务，数据过滤掉")
|
|
continue
|
|
# chunkSize=parse_data(raw_data,raw_data["input"]["chunkSize"])
|
|
if ':$[' not in raw_data["input"]["content"]:
|
|
content=raw_data["input"]["content"]
|
|
else:
|
|
content=parse_data(raw_data,raw_data["input"]["content"])
|
|
if ':$[' not in raw_data["input"]["fieldName"]:
|
|
fieldName=raw_data["input"]["fieldName"]
|
|
else:
|
|
fieldName=parse_data(raw_data,raw_data["input"]["fieldName"])
|
|
if ':$[' not in raw_data["input"]["dataId"]:
|
|
dataId=raw_data["input"]["dataId"]
|
|
else:
|
|
dataId=parse_data(raw_data,raw_data["input"]["dataId"])
|
|
# dataId=raw_data["dataId"]
|
|
if content and fieldName and dataId:
|
|
vector_db=LangChainChroma(fieldName)
|
|
docs=vector_db.text_splitter.split_text(content)
|
|
res,db_count=vector_db.add_documents(docs,dataId)
|
|
vector_db.db_close()
|
|
logging.info('当前数据划分{}个块。数据库{}共有{}个块'.format(len(res), fieldName,db_count))
|
|
# res=LC.addChroma(content,fieldName,logging,chunkSize)
|
|
res_tmp['resultsID']=res
|
|
raw_data["result"] = {"successCode": "", "errorLog": "", "results": ""}
|
|
if res:
|
|
res_tmp["status"]=1
|
|
raw_data["result"]["successCode"] = "1"
|
|
raw_data["result"]["status"] = 1
|
|
raw_data["result"]["message"] = "成功"
|
|
else:
|
|
res_tmp["status"]=3
|
|
raw_data["result"]["successCode"] = "0"
|
|
raw_data["result"]["status"] = 2
|
|
raw_data["result"]["message"] = "异常"
|
|
else:
|
|
res_tmp["status"] = 3
|
|
raw_data["result"]["successCode"] = "0"
|
|
raw_data["result"]["errorLog"] = "请检查content/fieldName/dataId，要求非空"
|
|
raw_data["result"]["status"] = 2
|
|
raw_data["result"]["message"] = "请检查content/fieldName/dataId，要求非空"
|
|
|
|
res_tmp_json = json.dumps(res_tmp, ensure_ascii=False)
|
|
raw_data["result"]["results"]=res_tmp_json
|
|
logging.info("结果数据为：{}".format(raw_data))
|
|
to_kafka.send_kafka(raw_data, logging)
|
|
else:
|
|
# 暂无任务，进入休眠
|
|
time.sleep(10)
|
|
except:
|
|
raw_data["result"]={}
|
|
raw_data["result"]["successCode"] = "0"
|
|
raw_data["result"]["errorLog"] = traceback.format_exc()
|
|
res_tmp["status"] = 3
|
|
raw_data["result"]["status"] = 2
|
|
raw_data["result"]["message"] = "异常"
|
|
raw_data["result"]["results"] = json.dumps(res_tmp, ensure_ascii=False)
|
|
logging.error(traceback.format_exc())
|
|
to_kafka.send_kafka(raw_data, logging)
|
|
|
|
|
|
def zk_monitoring():
|
|
try:
|
|
#线上环境
|
|
zk = KazooClient(hosts=config['zookeeper']['zkhost'])
|
|
#测试环境
|
|
# zk = KazooClient(hosts='172.16.12.55:2181,172.16.12.56:2181,172.16.12.57:2181')
|
|
zk.start()
|
|
# 设置监听器
|
|
@zk.DataWatch("/analyze")
|
|
def watch_node(data, stat, event):
|
|
if event is not None and event.type == EventType.CHANGED:
|
|
data, stat = zk.get("/analyze")
|
|
logging.info("执行删除操作：{}".format(data))
|
|
d = json.loads(data)
|
|
id = d["scenes_id"]
|
|
stop_dict[id] = {}
|
|
stop_dict[id]["version"] = d["version"]
|
|
stop_dict[id]["operation"] = d["operation"]
|
|
# 保持程序运行以监听节点变化
|
|
try:
|
|
while True:
|
|
time.sleep(1)
|
|
except:
|
|
logging.info("Stopping...")
|
|
# 关闭连接
|
|
zk.stop()
|
|
zk.close()
|
|
except:
|
|
logging.error(traceback.format_exc())
|
|
|
|
|
|
|