Replace artlebedev typograph with mdash

This commit is contained in:
2015-12-26 14:59:13 +03:00
parent b235152453
commit 82ad98e302
3 changed files with 97 additions and 136 deletions

View File

@ -1,111 +0,0 @@
"""
remotetypograf.py
python-implementation of ArtLebedevStudio.RemoteTypograf class (web-service client)
Copyright (c) Art. Lebedev Studio | http://www.artlebedev.ru/
Typograf homepage: http://typograf.artlebedev.ru/
Web-service address: http://typograf.artlebedev.ru/webservices/typograf.asmx
WSDL-description: http://typograf.artlebedev.ru/webservices/typograf.asmx?WSDL
Default charset: UTF-8
Python version
Author: Sergey Lavrinenko (s.lavrinenko@gmail.com)
Version: 1.0 (2007-05-18) based on script by Andrew Shitov (ash@design.ru)
Example:
from RemoteTypograf import RemoteTypograf
rt = RemoteTypograf()
# rt = RemoteTypograf('windows-1251')
print rt.processText ('"Вы все еще кое-как верстаете в "Ворде"? - Тогда мы идем к вам!"');
"""
import socket
class RemoteTypograf:
_entityType = 4
_useBr = 1
_useP = 1
_maxNobr = 3
_encoding = 'UTF-8'
def __init__(self, encoding='UTF-8'):
self._encoding = encoding
def htmlEntities(self):
self._entityType = 1
def xmlEntities(self):
self._entityType = 2
def mixedEntities(self):
self._entityType = 4
def noEntities(self):
self._entityType = 3
def br(self, value):
self._useBr = 1 if value else 0
def p(self, value):
self._useP = 1 if value else 0
def nobr(self, value):
self._maxNobr = int(value) if value else 0
def processText(self, text):
text = text.replace('&', '&')
text = text.replace('<', '&lt;')
text = text.replace ('>', '&gt;')
SOAPBody = '<?xml version="1.0" encoding="UTF-8"?>\n'
SOAPBody += '<soap:Envelope xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:soap="http://schemas.xmlsoap.org/soap/envelope/">\n'
SOAPBody += '<soap:Body>\r\n'
SOAPBody += ' <ProcessText xmlns="http://typograf.artlebedev.ru/webservices/">\n'
SOAPBody += ' <text>%s</text>\n' % text
SOAPBody += ' <entityType>%s</entityType>\n' % self._entityType
SOAPBody += ' <useBr>%s</useBr>\n' % self._useBr
SOAPBody += ' <useP>%s</useP>\n' % self._useP
SOAPBody += ' <maxNobr>%s</maxNobr>\n' % self._maxNobr
SOAPBody += ' </ProcessText>\n'
SOAPBody += ' </soap:Body>\n'
SOAPBody += '</soap:Envelope>\n'
host = 'typograf.artlebedev.ru';
SOAPRequest = 'POST /webservices/typograf.asmx HTTP/1.1\n'
SOAPRequest += 'Host: typograf.artlebedev.ru\n'
SOAPRequest += 'Content-Type: text/xml\n'
SOAPRequest += 'Content-Length: %d\n' % len(bytearray(SOAPBody, 'utf-8'))
SOAPRequest += 'SOAPAction: "http://typograf.artlebedev.ru/webservices/ProcessText"\n\n'
SOAPRequest += SOAPBody
remoteTypograf = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
remoteTypograf.connect((host, 80))
remoteTypograf.sendall(bytearray(SOAPRequest, 'utf-8'))
typografResponse = bytearray()
while 1:
buf = remoteTypograf.recv(8192)
if len(buf) == 0: break
typografResponse += buf
remoteTypograf.close()
typografResponse = typografResponse.decode(encoding='utf-8')
startsAt = typografResponse.find('<ProcessTextResult>') + 19
endsAt = typografResponse.find('</ProcessTextResult>')
typografResponse = typografResponse[startsAt:endsAt]
typografResponse = typografResponse.replace('&amp;', '&' )
typografResponse = typografResponse.replace('&lt;', '<')
typografResponse = typografResponse.replace ('&gt;', '>')
return typografResponse

65
lib/mdash.py Normal file
View File

@ -0,0 +1,65 @@
import json
import urllib.request
__author__ = 'Anton Vakhrushev'
__email__ = 'anwinged@ya.ru'
MDASH_URL = 'http://mdash.ru/api.v1.php'
class TypographError(RuntimeError):
pass
class Typograph(object):
def __init__(self, rules = None):
self.rules = self.__prepare_rules(rules)
def process(self, text):
data = self.__prepare_data(text)
answer = self.__request(data)
answer = json.loads(answer)
self.__ensure_answer_is_correct(answer)
return answer['result']
def __prepare_data(self, text):
data = {}
data.update(self.rules)
data.update({'text': text})
return data
def __request(self, data):
data = urllib.parse.urlencode(data)
data = data.encode('utf-8')
request = urllib.request.Request(MDASH_URL, data)
responce = urllib.request.urlopen(request)
return responce.read().decode('utf-8')
def __ensure_answer_is_correct(self, answer):
if answer.get('status') == 'error':
raise TypographError()
def __prepare_rules(self, params):
def convert(v):
return 'on' if v == 'on' or v == True else 'off'
result = {}
params = params or {}
for k in params:
result[k] = convert(params[k])
return result
if __name__ == '__main__':
params = {
'Text.paragraphs': False,
'Text.breakline': 'off',
}
t = Typograph(params)
print(t.process('"Вы все еще кое-как верстаете в "Ворде"? - Тогда мы идем к вам!"'))