wwwjdic: New plugin

omnipresence/plugins/wwwjdic/__init__.py

@@ -0,0 +1,87 @@
+# -*- test-case-name: omnipresence.plugins.wwwjdic.test_wwwjdic
+"""Event plugins for searching WWWJDIC."""
+
+
+import re
+import urllib
+
+from bs4 import BeautifulSoup
+from twisted.internet.defer import inlineCallbacks, returnValue
+from twisted.web.client import readBody
+try:
+    from waapuro import romanize
+except ImportError:
+    romanize = None
+
+from ...plugin import EventPlugin, UserVisibleError
+from ...web.http import default_agent
+
+
+#: A regex for identifying pronunciations in a JDIC entry, if present.
+PRONUNCIATIONS_RE = re.compile(ur'\[([^\]]+)\]')
+
+#: A regex for identifying markings at the end of a kana pronunciation.
+MARKINGS_RE = re.compile(ur'(?:\([^)]+\))+$')
+
+
+class Default(EventPlugin):
+    def __init__(self):
+        self.agent = default_agent
+        self.romanize = romanize
+
+    @inlineCallbacks
+    def on_command(self, msg):
+        if not msg.content:
+            raise UserVisibleError('Please specify a search query.')
+        q = urllib.quote_plus(msg.content)
+        response = yield self.agent.request('GET',
+            'http://www.edrdg.org/cgi-bin/wwwjdic/wwwjdic?1ZUJ{}'.format(q))
+        content = yield readBody(response)
+        soup = BeautifulSoup(content)
+        if not soup.pre:
+            raise UserVisibleError('No results found for \x02{}\x02.'
+                                   .format(msg.content))
+        raw = soup.pre.string.strip().splitlines()
+        results = []
+        for i, result in enumerate(raw):
+            if not result.strip():
+                continue
+            # Find the kana pronunciations and add their romanizations.
+            if self.romanize:
+                match = PRONUNCIATIONS_RE.search(result)
+                if match is None:
+                    pronunciations = result.split(None, 1)[0]
+                    start = 0
+                    end = len(pronunciations)
+                else:
+                    pronunciations = match.group(1)
+                    start = match.start(1)
+                    end = match.end(1)
+                pronunciations = pronunciations.split(u';')
+                with_romanizations = []
+                for pronunciation in pronunciations:
+                    match = MARKINGS_RE.search(pronunciation)
+                    if match is not None:
+                        pronunciation = pronunciation[:match.start()]
+                    with_romanizations.append(
+                        pronunciation +
+                        u' (' + self.romanize(pronunciation) + u')' +
+                        (u'' if match is None else u' ' + match.group(0)))
+                result = (result[:start] +
+                          u'; '.join(with_romanizations) +
+                          result[end:])
+            # Strip off the trailing slash for the last gloss, then
+            # replace the first slash with nothing and the remaining
+            # ones with semicolons, in an approximation of the Web
+            # interface.
+            result = result[:-1].strip()
+            result = result.replace(u'/', u'', 1)
+            result = result.replace(u'/', u'; ')
+            results.append(result)
+        returnValue(results)
+
+    def on_cmdhelp(self, msg):
+        return collapse("""\
+            \x1Fquery\x1F - Look up a Japanese word or phrase in Jim
+            Breen's WWWJDIC <http://wwwjdic.org/>.
+            """)

omnipresence/plugins/wwwjdic/test_wwwjdic.py

@@ -0,0 +1,37 @@
+# -*- coding: utf-8
+"""Unit tests for the wwwjdic event plugin."""
+# pylint: disable=missing-docstring,too-few-public-methods
+
+
+from ...message import collapse
+from ...test.helpers import AbstractCassetteTestCase
+
+from . import Default
+
+
+class WWWJDICTestCase(AbstractCassetteTestCase):
+    command_class = Default
+
+    @staticmethod
+    def romanize(string):
+        return 'mogiroomazi'
+
+    def setUp(self):
+        super(WWWJDICTestCase, self).setUp()
+        self.command.romanize = WWWJDICTestCase.romanize
+
+    @AbstractCassetteTestCase.use_cassette('wwwjdic/no-results')
+    def test_no_results(self):
+        return self.assert_error(
+            'slartibartfast',
+            'No results found for \x02slartibartfast\x02.')
+
+    @AbstractCassetteTestCase.use_cassette('wwwjdic/some-results')
+    def test_some_results(self):
+        return self.assert_reply('amanogawa', map(collapse, [
+            u"""天の川(P);天の河(P)
+                [あまのがわ (mogiroomazi) (P); あまのかわ (mogiroomazi)]
+                (n) Milky Way; (P)""",
+            u"""天の川銀河
+                [あまのがわぎんが (mogiroomazi); あまのかわぎんが (mogiroomazi)]
+                (n) Milky Way Galaxy"""]))

omnipresence/test/fixtures/cassettes/wwwjdic/no-results.json

@@ -0,0 +1 @@
+{"recorded_with": "Stenographer 0.1-dev", "http_interactions": [{"response": {"status": {"message": "OK", "code": 200}, "http_version": "1.1", "body": {"base64_string": "H4sIAAAAAAAAAy2Oy07DMBBF9/mKqfdhisoCgWOptY0ISh8LV1aXUWoUq6kT7CmQv6dR2Mzo6NzRXL5Qe2lOBw3vZlvB4bipSgksR7QriaiMmsXTw/IRTKxD8uT7UHeIesdExic7Lb1Wgm+1WUNLNOTu6+a/Cyb7QC5QbsbBMWhmKhi5X8KWrt0rNG0dk6PiaN7yZya4KU2lhbX2Q5XyBWwfz6B8Grp65DjLjOP8brNXpzsNItv1cK2paV2CHxcdfPa3cL7PCNT6BBc3LqYYx/8TnGv/AYzDF4r+AAAA", "encoding": "utf-8"}, "headers": {"Content-Encoding": ["gzip"], "Access-Control-Allow-Origin": ["*"], "Date": ["Mon, 10 Aug 2015 18:21:36 GMT"], "Vary": ["Accept-Encoding"], "Server": ["Apache/2.2.29 (Unix) mod_ssl/2.2.29 OpenSSL/1.0.1f DAV/2 mod_wsgi/4.2.8 Python/2.6.4 PHP/5.4.9 mod_perl/2.0.4 Perl/v5.12.2"], "Access-Control-Allow-Credentials": ["false"], "Content-Type": ["text/html; charset=UTF-8"]}}, "recorded_at": "Mon, 10 Aug 2015 18:21:36 -0000", "request": {"method": "GET", "uri": "http://www.edrdg.org/cgi-bin/wwwjdic/wwwjdic?1ZUJslartibartfast", "body": {"string": "", "encoding": "utf-8"}, "headers": {"Accept-Encoding": ["gzip"], "User-Agent": ["Omnipresence/3.0alpha1 (+bot; https://github.com/kxz/omnipresence)"]}}}]}

omnipresence/test/fixtures/cassettes/wwwjdic/some-results.json

@@ -0,0 +1 @@
+{"recorded_with": "Stenographer 0.1-dev", "http_interactions": [{"recorded_at": "Mon, 10 Aug 2015 18:22:46 -0000", "request": {"uri": "http://www.edrdg.org/cgi-bin/wwwjdic/wwwjdic?1ZUJamanogawa", "method": "GET", "body": {"encoding": "utf-8", "string": ""}, "headers": {"User-Agent": ["Omnipresence/3.0alpha1 (+bot; https://github.com/kxz/omnipresence)"], "Accept-Encoding": ["gzip"]}}, "response": {"headers": {"Access-Control-Allow-Credentials": ["false"], "Content-Type": ["text/html; charset=UTF-8"], "Access-Control-Allow-Origin": ["*"], "Date": ["Mon, 10 Aug 2015 18:22:46 GMT"], "Content-Encoding": ["gzip"], "Server": ["Apache/2.2.29 (Unix) mod_ssl/2.2.29 OpenSSL/1.0.1f DAV/2 mod_wsgi/4.2.8 Python/2.6.4 PHP/5.4.9 mod_perl/2.0.4 Perl/v5.12.2"], "Vary": ["Accept-Encoding"]}, "body": {"base64_string": "H4sIAAAAAAAAA7NRdPF3DokMcFXwCPH1UQgIdfLxdFZQ0tXXDzd21td3CXGBSJjoGRgqhBQl5hVnlmTm5yXm6Ou7+inZcdmAZEGUq6OLnY2va4ijQkZJSYFuamFpZpmtknN+XklqXoluSGVBqpJCMoRnq1SSWlGin1GSm2OtkJyRWFScWmIbGuKma6FkZxPiGeLjahceHu7l4ulspRCeX5Si4JJZXJCTWGmjD5HkstGHWOfk7xIJ5BWAcFGqHdfTJSsfN657un2uRoCmNYTzbNNmIEch+nFj0+PGfUCBx409j5v6QQqQhLqBQrEK+hp5mgq+mTnZlQrhiZX6QDX6CCNfdjUAzcIw6HFj3+OmyUA2unFwCXRzFdwTcxIrKvWBvgA72kYf6g19SFACADGwpz6SAQAA", "encoding": "utf-8"}, "status": {"message": "OK", "code": 200}, "http_version": "1.1"}}]}