Check-in [85d939b85a]
Many hyperlinks are disabled.
Use anonymous login
to enable hyperlinks.
Overview
Comment: | Add cli(), translate_python() providers, basic deepl_api(), and non-working deepl_web() implementation. |
---|---|
Downloads: | Tarball | ZIP archive | SQL archive |
Timelines: | family | ancestors | descendants | both | trunk |
Files: | files | file ages | folders |
SHA1: |
85d939b85a873d09b131fb2308ca81e5 |
User & Date: | mario 2020-05-24 19:01:38 |
Context
2020-05-24
| ||
19:04 | Rename pagetranslate_opts to settings. Adapt registry leave, automate property name detection, safeguard against unknown control types (e.g. framebox), prepare transition to registry read/write (instead of config file). Simplify argparse() and self.params update. check-in: 23054650f3 user: mario tags: trunk | |
19:01 | Add cli(), translate_python() providers, basic deepl_api(), and non-working deepl_web() implementation. check-in: 85d939b85a user: mario tags: trunk | |
18:59 | Introduce more options (microsoft, mymemory, cli) check-in: 4232826ef2 user: mario tags: trunk | |
Changes
Changes to pythonpath/Makefile.
1 2 3 4 5 6 7 8 9 | # The pythonpath/ directory can be packaged alongside, to inject new python # packages into LibreOffices` python bundle. Which really just makes sense # for the Windows package, because distro-packaged Office setups utilize the # system python path. all: pip install requests -t ./ --upgrade | > > | 1 2 3 4 5 6 7 8 9 10 11 | # The pythonpath/ directory can be packaged alongside, to inject new python # packages into LibreOffices` python bundle. Which really just makes sense # for the Windows package, because distro-packaged Office setups utilize the # system python path. all: pip install requests -t ./ --upgrade clean: find . -iname '*py[co]*' -exec rm -r {} \; |
Changes to pythonpath/translationbackends.py.
1 2 3 4 5 6 | # encoding: utf-8 # api: pagetranslate # type: classes # category: language # title: via_* translation backends # description: Implements the alternative services (google, deepl, ...) | | | > < < < < | | | | | | > > > > > > > > > | | | | | | > | > > > > > | | > > > > > | > > > > > > > > > > > > > > > > > > | > > > > > > | > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > | > > | > > > > > > > > > > > > > | > > > > > > > > > > > > > > > > > > > > > | > | > > | > > > > > > > > | > > > > > > > > > > > > > > > > > > > > > > > > > > > > > | | | > > < | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 | # encoding: utf-8 # api: pagetranslate # type: classes # category: language # title: via_* translation backends # description: Implements the alternative services (google, deepl, ...) # version: 1.3 # state: beta # depends: python:requests (>= 2.5) # config: - # # Different online service backends and http interfaces are now coalesced here. # # modules import re, json, time import os, subprocess, shlex import urllib from urllib.parse import urlencode, quote, quote_plus from httprequests import http log = None # regex rx_gtrans = re.compile('class="t0">(.+?)</div>', re.S) rx_splitpara = re.compile("(.{1,1895\.}|.{1,1900}\s|.*$)", re.S) rx_empty = re.compile("^[\s\d,.:;§():-]+$") rx_letters = re.compile("\w\w+", re.UNICODE) rx_breakln = re.compile("\s?/\s?#\s?§\s?/\s?") # Google Translate (default backend) # # · calls mobile page http://translate.google.com/m?hl=en&sl=auto&q=TRANSLATE # · iterates over each 1900 characters # class google: def __init__(self, params={}): self.params = params # config+argparse # request text translation from google def askgoogle(self, text, dst_lang="en", src_lang='auto'): # fetch translation page url = "http://translate.google.com/m?hl=%s&sl=%s&q=%s" % ( dst_lang, src_lang, quote_plus(text) ) html = http.get(url).content.decode("utf-8") # extract content from text <div> m = rx_gtrans.search(html) if m: text = m.group(1) text = text.replace("'", "'").replace("&", "&").replace("<", "<").replace(">", ">").replace(""", '"') #@todo: https://stackoverflow.com/questions/2087370/decode-html-entities-in-python-string else: log.warning("NO TRANSLATION RESULT EXTRACTED: " + html) log.debug("ORIG TEXT: " + repr(text)) return text # iterate over text segments (1900 char limit) def translate(self, text, lang="auto"): if lang == "auto": lang = self.params["lang"] #log.debug("translate %d chars" % len(text)) if len(text) < 2: log.debug("skipping/len<2") return text elif rx_empty.match(text): log.debug("skipping/empty") return text elif not rx_letters.search(text): log.debug("skipping/noletters") return text elif len(text) >= 1900: log.debug("spliterate/1900+") return " ".join(self.askgoogle(segment, lang) for segment in rx_splitpara.findall(text)) else: return self.askgoogle(text, lang) # translate w/ preserving paragraph breaks (meant for table cell content) def linebreakwise(self, text, lang="auto"): if self.params["crlf"] != "quick": # split on linebreaks and translate each individually text = "\n\n".join(self.translate(text, lang) for text in text.split("\n\n")) else: # use temporary placeholder `/#§/` text = self.translate(text.replace("\n\n", "/#§/")) text = re.sub(rx_breakln, "\n\n", text) return text # DeepL online translator uses some kind of json-rpc # # · haven't quite extracted all necesssary bits (origin of id unclear) # · will yield HTTP 429 Too many requests, # so probably not useful for multi-paragraph translation anyway # # class deepl_web(google): # < https://www2.deepl.com/jsonrpc # cookies: LMTBID: GUID... # referer: https://www.deepl.com/translator # body: # > result.translations[0].beams[0].postprocessed_sentence def __init__(self, params): self.params = params self.id = 702005000 self.lang = params["lang"].upper() r = http.get("https://www.deepl.com/translator") # should fetch us the cookie / No, it doesn't def rpc(self, text): return json.dumps({ "jsonrpc" : "2.0", "method" : "LMT_handle_jobs", "id" : self.id, "params" : { "lang" : { "target_lang" : self.lang, "user_preferred_langs" : [ self.lang, "EN" ], "source_lang_user_selected" : "auto" }, "timestamp" : int(time.time()*1000), "priority" : -1, "commonJobParams" : {}, "jobs" : [ { "raw_en_context_after" : [], "raw_en_context_before" : [], "kind" : "default", "preferred_num_beams" : 4, "raw_en_sentence" : text, "quality" : "fast" } ] } }) def translate(self, text): # skip empty paragraph/table snippets if len(text) < 2 or rx_empty.match(text) or not rx_letters.search(text): return text # request r = http.post( "https://www2.deepl.com/jsonrpc", data=self.rpc(text), headers={"Referer": "https://www.deepl.com/translator", "Content-Type": "text/plain"} ) if r.status_code != 200: log.error(repr(r.content)) return text #return r, r.content # decode r = r.json() if r.get("id"): self.id = r["id"] + 1 if r.get("result"): return r["result"]["translations"][0]["beams"][0]["postprocessed_sentence"] else: return text # DeepL API costs money # # Not sure if anyone will use this really. Unless the _web version allows testing, # nobody's gonna shell out money for a subscription - even if it surpassed GoogleT. # Likely makes sense for commercial users however. And the API is quite simple, so # that's why it's here. # # ENTIRELY UNTESTED # class deepl_api(deepl_web): def __init__(self, params): self.params = params def translate(self, text, preserve=0): # skip empty paragraph/table snippets if len(text) < 2 or rx_empty.match(text) or not rx_letters.search(text): return text # https://www.deepl.com/docs-api/translating-text/request/ r = http.get( "https://api.deepl.com/v2/translate", params={ "auth_key": self.params["api_key"], "text": text, "target_lang": self.params["lang"], "split_sentences": "1", "preserve_formatting": str(preserve) #"tag_handling": "xml" } ) if r.status_code == 200: r = r.json().get("translations") if r: return r[0]["text"] else: log.error(r.text) return text def linebreakwise(self, text): return self.translate(text, preserve=1) # Translate-python # requires `pip install translate` # # · provides "microsoft" backend (requires OAuth secret in api_key) # # · or "mymemory" (with email in `api_key` instead) # # https://translate-python.readthedocs.io/en/latest/ # class translate_python(google): def __init__(self, params={}): self.params = params # config+argparse #self.error = pagetranslate.MessageBox Translator = None try: from translate import Translator except: raise Exception("Run `pip install translate` to use this module.") # interestingly this backend function might just work as is. if params.get("mymemory"): self.translate = Translator( provider="mymemory", to_lang=params["lang"], email=params["api_key"] ) else: self.translate = Translator( provider="microsoft", to_lang=params["lang"], secret_access_key=params["api_key"] ) # though .linebreakwise has no equivalent, not sure if necessaary, # or if formatting/linebreaks are preserved anyway # (or: we might just use the default google. implementation) self.linebreakwise = self.translate translate = None #linebreakwise = None # Because, why not? # Invokes a commandline tool for translating texts. # # → with e.g. `translate-cli -t {}` in "api_key" # class cli(google): def __init__(self, params): self.params = params self.cmd = params["api_key"] # pipe text through external program def translate(self, text): if rx_empty.match(text) or not rx_letters.search(text): return text cmd = [s.format(text) for s in shlex.split(self.cmd)] proc = subprocess.run(cmd, stdout=subprocess.PIPE) return proc.stdout.decode("utf-8") # maps a pagetranslate.t.* object (in main module), # according to config dict {"goog":1, "deepl":0} def assign_service(params): if params.get("deepl_web"): return deepl_web(params) elif params.get("deepl_api"): return deepl_api(params) elif params.get("translate_python") or params.get("microsoft") or params.get("mymemory"): return translate_python(params) elif params.get("cli"): return cli(params) else: return google(params) |