02fc414 · searx · git.iko.soy

git.iko.soy/searx

[mod] dictzone, translated, currency_convert: use engine_type online_curency and online_dictionnary

commit02fc4147ce745325ff25146a8085a915a5d3cacd authorAlexandre Flament <alex@al-f.net> dateWed 16 Dec 2020, 16:20 (6 years ago) parent7ec8bc3

4 files changed+59−99

compare with master

unifiedsplit

docs/dev/engine_overview.rst+47−19 searx/engines/currency_convert.py+2−40 searx/engines/dictzone.py+6−18 searx/engines/translated.py+4−22

docs/dev/engine_overview.rst+47−19

@@ -37,15 +37,16 @@ settings. However, the standard way is the following:

3737engine file

3838-----------

3939

40−======================= =========== ===========================================

40+======================= =========== ========================================================

4141argument type information

42−======================= =========== ===========================================

42+======================= =========== ========================================================

4343categories list pages, in which the engine is working

4444paging boolean support multible pages

4545language_support boolean support language choosing

4646time_range_support boolean support search time range

47−engine_type str ``online`` by default, ``offline``

48−======================= =========== ===========================================

47+engine_type str ``online`` by default, other possibles values are

48+ ``offline``, ``online_dictionnary``, ``online_currency``

49+======================= =========== ========================================================

4950

5051.. _engine settings:

5152

@@ -111,22 +112,49 @@ passed arguments

111112These arguments can be used to construct the search query. Furthermore,

112113parameters with default value can be redefined for special purposes.

113114

115+If the ``engine_type`` is ``online```:

116+

117+====================== ============== ========================================================================

118+argument type default-value, information

119+====================== ============== ========================================================================

120+url str ``''``

121+method str ``'GET'``

122+headers set ``{}``

123+data set ``{}``

124+cookies set ``{}``

125+verify bool ``True``

126+headers.User-Agent str a random User-Agent

127+category str current category, like ``'general'``

128+safesearch int ``0``, between ``0`` and ``2`` (normal, moderate, strict)

129+time_range Optional[str] ``None``, can be ``day``, ``week``, ``month``, ``year``

130+pageno int current pagenumber

131+language str specific language code like ``'en_US'``, or ``'all'`` if unspecified

132+====================== ============== ========================================================================

133+

134+

135+If the ``engine_type`` is ``online_dictionnary```, in addition to the ``online`` arguments:

136+

114137====================== ============ ========================================================================

115138argument type default-value, information

116139====================== ============ ========================================================================

117−url string ``''``

118−method string ``'GET'``

119−headers set ``{}``

120−data set ``{}``

121−cookies set ``{}``

122−verify boolean ``True``

123−headers.User-Agent string a random User-Agent

124−category string current category, like ``'general'``

125−started datetime current date-time

126−pageno int current pagenumber

127−language string specific language code like ``'en_US'``, or ``'all'`` if unspecified

140+from_lang str specific language code like ``'en_US'``

141+to_lang str specific language code like ``'en_US'``

142+query str the text query without the languages

128143====================== ============ ========================================================================

129144

145+If the ``engine_type`` is ``online_currency```, in addition to the ``online`` arguments:

146+

147+====================== ============ ========================================================================

148+argument type default-value, information

149+====================== ============ ========================================================================

150+amount float the amount to convert

151+from str ISO 4217 code

152+to str ISO 4217 code

153+from_name str currency name

154+to_name str currency name

155+====================== ============ ========================================================================

156+

157+

130158parsed arguments

131159----------------

132160

@@ -137,12 +165,12 @@ request:

137165=================== =========== ==========================================================================

138166argument type information

139167=================== =========== ==========================================================================

140−url string requested url

141−method string HTTP request method

168+url str requested url

169+method str HTTP request method

142170headers set HTTP header information

143−data set HTTP data information (parsed if ``method != 'GET'``)

171+data set HTTP data information

144172cookies set HTTP cookies

145−verify boolean Performing SSL-Validity check

173+verify bool Performing SSL-Validity check

146174max_redirects int maximum redirects, hard limit

147175soft_max_redirects int maximum redirects, soft limit. Record an error but don't stop the engine

148176raise_for_httperror bool True by default: raise an exception if the HTTP code of response is >= 300

searx/engines/currency_convert.py+2−40

@@ -1,54 +1,16 @@

11import json

2−import re

3−import unicodedata

4−from searx.data import CURRENCIES # NOQA

4+engine_type = 'online_currency'

75categories = []

86url = 'https://duckduckgo.com/js/spice/currency/1/{0}/{1}'

97weight = 100

108

11−parser_re = re.compile('.*?(\\d+(?:\\.\\d+)?) ([^.0-9]+) (?:in|to) ([^.0-9]+)', re.I)

129https_support = True

1310

1411

15−def normalize_name(name):

16− name = name.lower().replace('-', ' ').rstrip('s')

17− name = re.sub(' +', ' ', name)

18− return unicodedata.normalize('NFKD', name).lower()

19−

20−

21−def name_to_iso4217(name):

22− global CURRENCIES

23−

24− name = normalize_name(name)

25− currency = CURRENCIES['names'].get(name, [name])

26− return currency[0]

27−

28−

29−def iso4217_to_name(iso4217, language):

30− global CURRENCIES

31−

32− return CURRENCIES['iso4217'].get(iso4217, {}).get(language, iso4217)

33−

34−

3512def request(query, params):

36− m = parser_re.match(query)

37− if not m:

38− # wrong query

39− return params

40− amount, from_currency, to_currency = m.groups()

41− amount = float(amount)

42− from_currency = name_to_iso4217(from_currency.strip())

43− to_currency = name_to_iso4217(to_currency.strip())

44−

45− params['url'] = url.format(from_currency, to_currency)

46− params['amount'] = amount

47− params['from'] = from_currency

48− params['to'] = to_currency

49− params['from_name'] = iso4217_to_name(from_currency, 'en')

50− params['to_name'] = iso4217_to_name(to_currency, 'en')

51−

13+ params['url'] = url.format(params['from'], params['to'])

5214 return params

5315

5416

searx/engines/dictzone.py+6−18

@@ -9,36 +9,24 @@

99 @parse url, title, content

1010"""

1111

12−import re

1312from urllib.parse import urljoin

1413from lxml import html

15−from searx.utils import is_valid_lang, eval_xpath

14+from searx.utils import eval_xpath

1615

16+

17+engine_type = 'online_dictionnary'

1718categories = ['general']

1819url = 'https://dictzone.com/{from_lang}-{to_lang}-dictionary/{query}'

1920weight = 100

2021

21−parser_re = re.compile('.*?([a-z]+)-([a-z]+) ([^ ]+)$', re.I)

2222results_xpath = './/table[@id="r"]/tr'

2323https_support = True

2424

2525

2626def request(query, params):

27− m = parser_re.match(query)

28− if not m:

29− return params

30−

31− from_lang, to_lang, query = m.groups()

32−

33− from_lang = is_valid_lang(from_lang)

34− to_lang = is_valid_lang(to_lang)

35−

36− if not from_lang or not to_lang:

37− return params

38−

39− params['url'] = url.format(from_lang=from_lang[2],

40− to_lang=to_lang[2],

41− query=query)

27+ params['url'] = url.format(from_lang=params['from_lang'][2],

28+ to_lang=params['to_lang'][2],

29+ query=params['query'])

4230

4331 return params

4432

searx/engines/translated.py+4−22

@@ -8,44 +8,26 @@

88 @stable yes

99 @parse url, title, content

1010"""

11−import re

12−from searx.utils import is_valid_lang

1311

12+engine_type = 'online_dictionnary'

1413categories = ['general']

1514url = 'https://api.mymemory.translated.net/get?q={query}&langpair={from_lang}|{to_lang}{key}'

1615web_url = 'https://mymemory.translated.net/en/{from_lang}/{to_lang}/{query}'

1716weight = 100

1817https_support = True

1918

20−parser_re = re.compile('.*?([a-z]+)-([a-z]+) (.{2,})$', re.I)

2119api_key = ''

2220

2321

2422def request(query, params):

25− m = parser_re.match(query)

26− if not m:

27− return params

28−

29− from_lang, to_lang, query = m.groups()

30−

31− from_lang = is_valid_lang(from_lang)

32− to_lang = is_valid_lang(to_lang)

33−

34− if not from_lang or not to_lang:

35− return params

36−

3723 if api_key:

3824 key_form = '&key=' + api_key

3925 else:

4026 key_form = ''

41− params['url'] = url.format(from_lang=from_lang[1],

42− to_lang=to_lang[1],

43− query=query,

27+ params['url'] = url.format(from_lang=params['from_lang'][1],

28+ to_lang=params['to_lang'][1],

29+ query=params['query'],

4430 key=key_form)

45− params['query'] = query

46− params['from_lang'] = from_lang

47− params['to_lang'] = to_lang

48−

4931 return params

5032

5133