def format_value(self, entry):
return entry.pop()
def get_similar(self, word, data):
if not data:
try:
data = self.whisperer.findSimilarByWord(word,
max=self.max_similar+10)
data = data[0]['value']
except:
# error, the word was not found
data = self.get_suggestions(word)
data.append('<2nd lookup, %s not found>' % word)
return data
semes = data.split(' ')
data = self.whisperer.getSimilarBySem(semes, 10, True)
if len(data) < 2:
return []
records = []
for info in data:
if info['key'] != word:
records.append(self.prepare_entry([info]))
return records[0:self.max_similar]
#test(INDEX, 'scalar', [('particle', None)], cls=TaggingWhisper)
#test(INDEX, 'scalar', [('boson', '0012v _rel194 _rel195')], cls=TaggingWhisper)
get_worker, application = app(INDEX, MAX_AUTOCOMPLETE, MAX_SIMILAR, TaggingWhisper)
#!python -u
import sys
import os
import dumean
_d = os.path.abspath(os.path.join(os.path.dirname(dumean.__file__), '../../..'))
if _d not in sys.path:
sys.path.append(_d)
MAX_AUTOCOMPLETE = 10
MAX_SIMILAR = 10
INDEX = os.path.join(_d, 'indices/hepnet')
from whisper import app, test, Whisperer
class KeywordWhisper(Whisperer):
def format_value(self, entry):
if len(entry) > 1:
entry = 'keyword:(%s)' % ' OR '.join(entry)
else:
entry = 'keyword:%s' % entry.pop()
return entry
#test(INDEX, 'scalarz', [('particle', None)])
get_worker, application = app(INDEX, MAX_AUTOCOMPLETE, MAX_SIMILAR, KeywordWhisper)
seen[r['key']] = 1
out.append({'label': r['key'], 'value': r['key'], 'data': r['value']})
if len(out) < self.max_autocomplete:
others = Whisperer.get_suggestions(self, word)
i = 0
while len(out) < self.max_autocomplete and i < len(others):
x = others[i]
if x['label'] not in seen:
out.append(x)
i += 1
return out
def format_value(self, entry):
return entry.pop()
#test(INDEX, 'xA', [('xA9Pv YRLVs', None)], cls=RandomWhisper)
get_worker, application = app(INDEX, MAX_AUTOCOMPLETE, MAX_SIMILAR, RandomWhisper)
def run_wsgi():
port = 8080
from wsgiref.util import setup_testing_defaults
from wsgiref.simple_server import make_server
httpd = make_server('', port, application)
print "Serving on port %s..." % port
httpd.serve_forever()
if __name__ == "__main__":
run_wsgi()
def get_similar(self, word, data):
# for authors, we don't have any information so we'll search
# for similar names
if not data or (len(data) > 1 and not data[0]):
return self.get_suggestions(word)
recs = []
for n in data[0].split("|"):
recs.append({"label": n, "value": n})
return recs
def prepare_entry(self, info):
"""Info Entry is actually array of possible records
becaue every word may have many meanings"""
score, entry, data = 0.0, set(), []
key = "%s (%d)" % (info[0]["key"], len(info))
for x in info:
if " " in x["key"]:
entry.add('"%s"' % x["key"])
else:
entry.add(x["key"])
data.append(x["value"])
score += x["score"]
return {"label": key, "value": self.format_value(entry), "data": self.format_data(data)}
# test(INDEX, 'Ellis', [('Ellis', ['Ellis J', 'Ellis J.R'])],
# cls=AuthorWhisper)
get_worker, application = app(INDEX, MAX_AUTOCOMPLETE, MAX_SIMILAR, AuthorWhisper)