def on_post(self, req, resp):
"""Handles POST requests"""
resp.set_header('Access-Control-Allow-Origin', 'http://localhost:8081')
resp.set_header('Access-Control-Allow-Methods', '*')
resp.set_header('Access-Control-Allow-Headers', '*')
resp.set_header('Access-Control-Allow-Credentials', 'true')
resp.set_header("Cache-Control", "no-cache")
data = req.stream.read(req.content_length)
jsondata = json.loads(data)
clean_title = shortenlines(jsondata.title)
clean_content = cleanall(jsondata.content)
resp.media = self.bert_classification(clean_title, clean_content)
def on_get(self, req, resp):
logger.info("...")
resp.set_header('Access-Control-Allow-Origin', 'http://localhost:8081')
resp.set_header('Access-Control-Allow-Methods', '*')
resp.set_header('Access-Control-Allow-Headers', '*')
resp.set_header('Access-Control-Allow-Credentials', 'true')
title = req.get_param('1', True)
content = req.get_param('2', True)
clean_title = shortenlines(title)
clean_content = cleanall(content)
resp.media = self.bert_classification(clean_title, clean_content)
logger.info("###")
def on_post(self, req, resp):
"""Handles POST requests"""
resp.set_header('Access-Control-Allow-Origin', '*')
resp.set_header('Access-Control-Allow-Methods', '*')
resp.set_header('Access-Control-Allow-Headers', '*')
resp.set_header('Access-Control-Allow-Credentials', 'true')
resp.set_header("Cache-Control", "no-cache")
data = req.stream.read(req.content_length)
start = time.process_time_ns()
jsondata = json.loads(data)
clean_title = shortenlines(jsondata['1'])
clean_content = cleanall(jsondata['2'])
resp.media = self.bert_classification(clean_title, clean_content)
logger.info("tot:{}ns".format(time.process_time_ns() - start))
logger.info("###")
import json
import pandas
import urllib3
from classmerge import match
from dataclean import cleanall
df = pandas.read_csv("data/valid1.csv")
http = urllib3.PoolManager()
correct = 0
for index, row in df.iterrows():
label = row[0]
title = row[1].replace(".doc", "").replace(".docx", "")
content = cleanall(row[2])
url = "http://192.168.0.161:58080/z?1={}&2={}".format(title, content)
print(url)
if len(url) > 9999:
url = url[:9999]
result = http.request('GET', url)
result = json.loads(result.data)
print(label, result['answer'][0])
df.at[index, 'type1'] = result['answer'][0]
df.at[index, 'title'] = title
df.at[index, 'content'] = content
if match(result['answer'][0], label):
correct += 1
df.to_csv("data/valid1-filled.csv", index=False)
print('ACCURACY:{}%'.format(correct * 100.0 / len(df)))
import json
import pandas
import urllib3
from classmerge import class_case
from dataclean import cleanall
df = pandas.read_csv("data/test.csv")
http = urllib3.PoolManager()
correct = 0
for index, row in df.iterrows():
label = row[0]
title = str(row[1]) #.replace(".doc","").replace(".docx","")
content = cleanall(str(row[2]))
url = "http://192.168.0.161:58081/z?1={}&2={}".format(title, content)
print(url)
if len(url) > 9999:
url = url[:9999]
result = http.request('GET', url)
result = json.loads(result.data)
print(class_case[label - 1], result['answer'][0])
# df.at[index, 'category'] = result['answer'][0]
# df.at[index, 'title'] = title
# df.at[index, 'fulltext'] = content
if class_case.index(result['answer'][0]) == label - 1:
correct += 1
#df.to_csv("eval/test-bert.csv", index=False)
print('ACCURACY:{}%'.format(correct * 100.0 / len(df)))
