X-Git-Url: https://git.mdrn.pl/wolnelektury.git/blobdiff_plain/357027375ff8867f42ca34bcbfb5a78b5b185fc3..907aba88951334676a259dccc8f7edd8934d5e00:/src/search/custom.py?ds=sidebyside
diff --git a/src/search/custom.py b/src/search/custom.py
index b3b704d0b..933715719 100644
--- a/src/search/custom.py
+++ b/src/search/custom.py
@@ -1,116 +1,86 @@
-# -*- coding: utf-8 -*-
# This file is part of Wolnelektury, licensed under GNU Affero GPLv3 or later.
# Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
#
-from sunburnt import sunburnt
-from lxml import etree
-import urllib
+import re
+from urllib.parse import urlencode
import warnings
-from sunburnt import search
-import copy
from httplib2 import socket
-import re
+from lxml import etree
+from scorched import connection, exc, search
-class TermVectorOptions(search.Options):
- def __init__(self, schema, original=None):
- self.schema = schema
- if original is None:
- self.fields = set()
- self.positions = False
- else:
- self.fields = copy.copy(original.fields)
- self.positions = copy.copy(original.positions)
-
- def update(self, positions=False, fields=None):
- if fields is None:
- fields = []
- if isinstance(fields, basestring):
- fields = [fields]
- self.schema.check_fields(fields, {"stored": True})
- self.fields.update(fields)
- self.positions = positions
-
- def options(self):
- opts = {}
- if self.positions or self.fields:
- opts['tv'] = 'true'
- if self.positions:
- opts['tv.positions'] = 'true'
- if self.fields:
- opts['tv.fl'] = ','.join(sorted(self.fields))
- return opts
-
-
-class CustomSolrConnection(sunburnt.SolrConnection):
+class CustomSolrConnection(connection.SolrConnection):
def __init__(self, *args, **kw):
super(CustomSolrConnection, self).__init__(*args, **kw)
self.analysis_url = self.url + "analysis/field/"
def analyze(self, params):
- qs = urllib.urlencode(params)
+ qs = urlencode(params)
url = "%s?%s" % (self.analysis_url, qs)
if len(url) > self.max_length_get_url:
- warnings.warn("Long query URL encountered - POSTing instead of "
- "GETting. This query will not be cached at the HTTP layer")
+ warnings.warn("Long query URL encountered - POSTing instead of GETting. "
+ "This query will not be cached at the HTTP layer")
url = self.analysis_url
kwargs = dict(
method="POST",
- body=qs,
+ data=qs,
headers={"Content-Type": "application/x-www-form-urlencoded"},
)
else:
kwargs = dict(method="GET")
- r, c = self.request(url, **kwargs)
- if r.status != 200:
- raise sunburnt.SolrError(r, c)
- return c
-
-
-# monkey patching sunburnt SolrSearch
-search.SolrSearch.option_modules += ('term_vectorer',)
+ response = self.request(url=url, **kwargs)
+ if response.status_code != 200:
+ raise exc.SolrError(response)
+ return response.content
-def __term_vector(self, positions=False, fields=None):
- newself = self.clone()
- newself.term_vectorer.update(positions, fields)
- return newself
-setattr(search.SolrSearch, 'term_vector', __term_vector)
-
-
-def __patched__init_common_modules(self):
- __original__init_common_modules(self)
- self.term_vectorer = TermVectorOptions(self.schema)
-__original__init_common_modules = search.SolrSearch._init_common_modules
-setattr(search.SolrSearch, '_init_common_modules', __patched__init_common_modules)
-
-
-class CustomSolrInterface(sunburnt.SolrInterface):
+class CustomSolrInterface(connection.SolrInterface):
# just copied from parent and SolrConnection -> CustomSolrConnection
- def __init__(self, url, schemadoc=None, http_connection=None, mode='', retry_timeout=-1, max_length_get_url=sunburnt.MAX_LENGTH_GET_URL):
- self.conn = CustomSolrConnection(url, http_connection, retry_timeout, max_length_get_url)
- self.schemadoc = schemadoc
- if 'w' not in mode:
- self.writeable = False
- elif 'r' not in mode:
- self.readable = False
- try:
- self.init_schema()
- except socket.error, e:
- raise socket.error, "Cannot connect to Solr server, and search indexing is enabled (%s)" % str(e)
+ def __init__(self, url, http_connection=None, mode='',
+ retry_timeout=-1, max_length_get_url=connection.MAX_LENGTH_GET_URL,
+ search_timeout=()):
+ """
+ :param url: url to Solr
+ :type url: str
+ :param http_connection: optional -- already existing connection
+ :type http_connection: requests connection
+ :param mode: optional -- mode (readable, writable) Solr
+ :type mode: str
+ :param retry_timeout: optional -- timeout until retry
+ :type retry_timeout: int
+ :param max_length_get_url: optional -- max length until switch to post
+ :type max_length_get_url: int
+ :param search_timeout: (optional) How long to wait for the server to
+ send data before giving up, as a float, or a
+ (connect timeout, read timeout) tuple.
+ :type search_timeout: float or tuple
+ """
+
+ self.conn = CustomSolrConnection(
+ url, http_connection, mode, retry_timeout, max_length_get_url)
+ self.schema = self.init_schema()
+ self._datefields = self._extract_datefields(self.schema)
+
def _analyze(self, **kwargs):
- if not self.readable:
+ if not self.conn.readable:
raise TypeError("This Solr instance is only for writing")
args = {
'analysis_showmatch': True
}
- if 'field' in kwargs: args['analysis_fieldname'] = kwargs['field']
- if 'text' in kwargs: args['analysis_fieldvalue'] = kwargs['text']
- if 'q' in kwargs: args['q'] = kwargs['q']
- if 'query' in kwargs: args['q'] = kwargs['q']
-
- params = map(lambda (k, v): (k.replace('_', '.'), v), sunburnt.params_from_dict(**args))
+ if 'field' in kwargs:
+ args['analysis_fieldname'] = kwargs['field']
+ if 'text' in kwargs:
+ args['analysis_fieldvalue'] = kwargs['text']
+ if 'q' in kwargs:
+ args['q'] = kwargs['q']
+ if 'query' in kwargs:
+ args['q'] = kwargs['q']
+
+ params = [
+ (k.replace('_', '.'), v)
+ for (k, v) in search.params_from_dict(**args)
+ ]
content = self.conn.analyze(params)
doc = etree.fromstring(content)
@@ -126,16 +96,15 @@ class CustomSolrInterface(sunburnt.SolrInterface):
matches.add((start, end))
if matches:
- return self.substring(kwargs['text'], matches,
- margins=kwargs.get('margins', 30),
- mark=kwargs.get('mark', ("", "")))
+ return self.substring(
+ kwargs['text'], matches, margins=kwargs.get('margins', 30), mark=kwargs.get('mark', ("", "")))
else:
return None
def analyze(self, **kwargs):
doc = self._analyze(**kwargs)
terms = doc.xpath("//lst[@name='index']/arr[last()]/lst/str[1]")
- terms = map(lambda n: unicode(n.text), terms)
+ terms = map(lambda n: str(n.text), terms)
return terms
def expand_margins(self, text, start, end):
@@ -155,37 +124,31 @@ class CustomSolrInterface(sunburnt.SolrInterface):
break
end += 1
- return (start, end)
+ return start, end
def substring(self, text, matches, margins=30, mark=("", "")):
- start = None
- end = None
totlen = len(text)
- matches_margins = map(lambda (s, e):
- ((s, e),
- (max(0, s - margins), min(totlen, e + margins))),
- matches)
- matches_margins = map(lambda (m, (s, e)):
- (m, self.expand_margins(text, s, e)),
- matches_margins)
-
- # lets start with first match
+ matches_margins = [
+ ((s, e), self.expand_margins(text, max(0, s - margins), min(totlen, e + margins))) for s, e in matches]
+
+ # lets start with first match
(start, end) = matches_margins[0][1]
- matches = [matches_margins[0][0]]
+ new_matches = [matches_margins[0][0]]
for (m, (s, e)) in matches_margins[1:]:
if end < s or start > e:
continue
start = min(start, s)
end = max(end, e)
- matches.append(m)
+ new_matches.append(m)
snip = text[start:end]
- matches.sort(lambda a, b: cmp(b[0], a[0]))
+ new_matches.sort(key=lambda a: -a[0])
- for (s, e) in matches:
- off = - start
+ for (s, e) in new_matches:
+ off = -start
snip = snip[:e + off] + mark[1] + snip[e + off:]
snip = snip[:s + off] + mark[0] + snip[s + off:]
+ snip = re.sub('%s[ \t\n]+%s' % (mark[1], mark[0]), " ", snip)
return snip