1 # This file is part of python-rwhoisd
3 # Copyright (C) 2003, David E. Blacka
5 # $Id: QueryProcessor.py,v 1.3 2003/04/28 16:44:56 davidb Exp $
7 # This program is free software; you can redistribute it and/or modify
8 # it under the terms of the GNU General Public License as published by
9 # the Free Software Foundation; either version 2 of the License, or
10 # (at your option) any later version.
12 # This program is distributed in the hope that it will be useful, but
13 # WITHOUT ANY WARRANTY; without even the implied warranty of
14 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 # General Public License for more details.
17 # You should have received a copy of the GNU General Public License
18 # along with this program; if not, write to the Free Software
19 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
23 import Cidr, Rwhois, QueryParser
27 def __init__(self, db):
30 def _filter_obj_term(self, obj, term):
31 """Given a rwhoisobject and a query term (a 3 element tuple:
32 attr, operator, value), determine if the object satisfies the
33 term. Returns True if the object matches the term, False if
36 attr, op, searchval = term
39 # filter by named attribute
41 vals = obj.get_attr(attr)
45 res = match_values(searchval, vals)
46 if op == "!=": return not res
48 # filter by general term
50 for val in obj.values():
51 if match_value(searchval, val):
55 def _filter_obj(self, obj, terms):
56 """Given a rwhoisobject and a list of query terms (i.e., a
57 whole AND clause), return True if the object satisfies the
61 if not self._filter_obj_term(obj, term): return False
64 def _filter_results(self, reslist, terms):
65 """Given list of result objects (not simply the ids returned
66 from the search) and a list of query terms (i.e., a query
67 clause), remove elements that do not satisfy the terms.
68 Returns a list of objects that satisfy the filters."""
70 if not terms: return reslist
71 return [ x for x in reslist if self._filter_obj(x, terms) ]
73 def process_query_clause(self, clause, max=0):
74 """Process a query clause (a grouping of terms ANDed
75 together). This is where the indexed searches actually get
76 done. The technique used here is to search on one index and
77 use the rest of the clause to filter the results. Returns a
80 # the technique is to do an index search on the first (or
81 # maybe best) indexed term (bare terms are always considered
82 # indexed), and filter those results with the remaining terms.
84 # Note: this could be better if we found the "optimal" query
85 # term. One approach may be to create a cost function and
86 # search for the minimum cost term.
88 # Note: another approach might be to actually do indexed
89 # searches on all applicable terms (bare or using an indexed
90 # attribute) and find the intersection of the results.
95 orig_clause = clause[:]
97 # find the first searchable term:
98 for term, i in zip(clause, xrange(sys.maxint)):
99 attr, op, value = term
100 if op == "!=": continue
101 if not attr or self.db.is_indexed_attr(attr):
105 raise Rwhois.RwhoisError, (351, "No indexed terms in query clause")
107 # remove the search term from the clause, what remains is the
111 # if we have an attribute name, search on that.
113 res = self.db.search_attr(st[0], st[2], max)
115 if Cidr.valid_cidr(st[2].strip("*")):
116 res = self.db.search_cidr(st[2], max)
118 res = self.db.search_normal(st[2], max)
120 objs = self._filter_results(self.db.fetch_objects(res.list()), clause)
122 queryres = QueryResult(objs)
125 refs = self.process_referral_search(orig_clause)
126 queryres.add_referrals(refs)
130 def _is_in_autharea(self, value):
131 """Returns True if value could be considered to be contained
132 within an authority area. That is, is a subnet of a
133 network-type authority area or a subdomain of a domainname
134 type authority area."""
136 aas = self.db.get_authareas()
138 if isinstance(value, Cidr.Cidr):
140 cv = Cidr.valid_cidr(aa)
141 if cv and cv.is_supernet(value):
145 if is_domainname(aa) and is_subdomain(aa, value):
149 def _referral_search_cidr(self, cv, value):
150 """Return the IndexResult of a referral search for value, or
151 None if the value doesn't qualify for a Cidr referral
154 if not cv: return None
155 if not self._is_in_autharea(cv): return None
156 return self.db.search_referral(value)
158 def _referral_search_domain(self, value):
159 """Return the IndexResult of a referral search for value, or
160 None if the value doesn't qualify for a domain referral
163 if not is_domainname(value): return None
164 if not self._is_in_autharea(value): return None
168 res = self.db.search_referral(dn)
170 dn = reduce_domain(dn)
173 def _referral_search_term(self, value):
174 """Return the IndexResult of a referral search for value, or
175 None if the value didn't qualify for a referral search."""
177 cv = Cidr.valid_cidr(value)
179 return self._referral_search_cidr(cv, value)
180 elif is_domainname(value):
181 return self._referral_search_domain(value)
184 def process_referral_search(self, clause):
185 """Given a query clause, attempt to search for referrals
186 associated with the terms. Return a list of referral strings
187 that matched terms in the clause (if any). The only terms
188 that actually get searched are the ones that look
189 'heirarchical'. For now, the attribute part of the term is
190 essentially ignored, so a search for something like
191 'name=127.0.0.1' might concievably generate a referral, when
192 perhaps it shouldn't."""
194 # first check to see if the search is explictly for a referral
196 if (term[0] == "class-name" and term[1] == "="
197 and term[2] == "referral") or term[0] == "referred-auth-area":
198 # in which case, we return nothing
203 # look for heirarchical-looking terms.
204 for attr, op, value in clause:
205 if op == "!=": continue
206 res = self._referral_search_term(value)
207 if not res or not res.list():
210 ref_objs = self.db.fetch_objects(res.list())
211 ref_strs = [x for y in ref_objs for x in y.get_attr("referral")]
212 referrals.extend(ref_strs)
217 def process_full_query(self, query, max=0):
218 """Given a parsed query object, process it by unioning the
219 results of the various ORed together clauses"""
221 # shortcut for the very common single clause case:
222 if len(query.clauses) == 1:
223 res = self.process_query_clause(query.clauses[0], max)
226 # otherwise, union the results from all the causes
228 for clause in query.clauses:
229 res.extend(self.process_query_clause(clause), max)
230 if max and len(res) >= max:
236 def process_query(self, session, queryline):
237 """Given a session config and a query line, parse the query,
238 perform any searches, return any referrals."""
240 if not session.queryparser:
241 session.queryparser = QueryParser.get_parser()
245 query = QueryParser.parse(session.queryparser, queryline)
246 except Rwhois.RwhoisError, x:
247 session.wfile.write(Rwhois.error_message(x))
253 print "process_query: max =", max
255 query_result = self.process_full_query(query, max)
257 objects = query_result.objects()
258 referrals = query_result.referrals()
260 if not objects and not referrals:
261 session.wfile.write(Rwhois.error_message(230))
262 # session.wfile.write("\r\n")
265 limit_exceeded = False
266 if session.limit and len(objects) > session.limit:
267 del objects[session.limit:]
268 limit_exceeded = True
271 session.wfile.write(obj.to_wire_str())
272 session.wfile.write("\r\n")
276 session.wfile.write("\r\n")
277 session.wfile.write("\r\n".join(referrals))
278 session.wfile.write("\r\n")
281 session.wfile.write(Rwhois.error_message(330))
283 session.wfile.write(Rwhois.ok())
287 def __init__(self, objs=[], referrals=[]):
289 self.ids = [ x.getid() for x in objs ]
290 self._dict = dict(zip(self.ids, self.ids))
291 self.refs = referrals
293 def extend(self, list):
294 if isinstance(list, type(self)):
295 list = list.objects()
296 new_objs = [ x for x in list if not self._dict.has_key(x.getid()) ]
297 new_ids = [ x.getid() for x in new_objs ]
298 self.data.extend(new_objs)
299 self.ids.extend(new_ids)
300 self._dict.update(dict(zip(new_ids, new_ids)))
302 def add_referrals(self, referrals):
303 self.refs.extend(referrals)
314 def truncate(self, n=0):
315 to_del = self.ids[n:]
316 for i in to_del: del self._dict[i]
317 self.ids = self.ids[:n]
318 self.data = self.data[:n]
321 def match_value(searchval, val):
322 """Determine if a search value matches a data value. If both
323 matching terms are valid CIDR objects, then they are matched
324 according the CIDR wildcard rules (i.e., a single trailing * is a
325 supernet search, ** is a subnet search). If the search value is
326 not wildcarded, then they are just tested for numeric equality.
327 Otherwise, the terms are compared using string semantics
328 (substring, prefix, suffix, and exact match."""
330 if match_cidr(searchval, val): return True
332 # normalize the values for comparison.
333 searchval = searchval.lower()
337 if searchval.startswith("*") and searchval.endswith("*"):
338 sv = searchval.strip("*");
339 if val.find(sv) >= 0:
344 elif searchval.startswith("*"):
345 sv = searchval.lstrip("*")
346 return val.endswith(sv)
348 elif searchval.endswith("*"):
349 sv = searchval.rstrip("*")
350 return val.startswith(sv)
351 # the exact match case
353 return searchval == val
355 def match_values(searchval, val_list):
358 if match_value(searchval, val): return True
361 def match_cidr(searchval, val):
362 """If both terms are valid CIDR values (minus any trailing
363 wildcards of the search value), compare according the CIDR
364 wildcard rules: subnet, supernet, and exact match. If both terms
365 are not CIDR address, return False."""
368 sv = Cidr.valid_cidr(searchval.rstrip("*"))
369 rv = Cidr.valid_cidr(val)
371 if not sv or not rv: return False
373 if (searchval.endswith("**")):
374 return rv.is_subnet(sv)
375 elif (searchval.endswith("*")):
376 return rv.is_supernet(sv)
381 # this forms a pretty basic heuristic to see of a value looks like a
383 domain_regex = re.compile("[a-z0-9-]+\.[a-z0-9-.]+", re.I)
385 def is_domainname(value):
386 if domain_regex.match(value):
390 def is_subdomain(domain, subdomain):
391 domain = domain.lower();
392 subdomain = subdomain.lower();
394 dlist = domain.split('.')
395 sdlist = subdomain.split('.')
397 if len(dlist) > len(sdlist): return False
398 if len(dlist) == len(sdlist): return domain == subdomain
403 return dlist == sdlist[:len(dlist)]
405 def reduce_domain(domain):
406 dlist = domain.split('.')
408 return '.'.join(dlist)
410 def is_heirarchical(value):
411 if cidr.valid_cidr(value): return True
412 if is_domainname(value): return True
415 if __name__ == '__main__':
417 import MemDB, Session
421 print "loading schema:", sys.argv[1]
422 db.init_schema(sys.argv[1])
423 for data_file in sys.argv[2:]:
424 print "loading data file:", data_file
425 db.load_data(data_file)
429 processor = QueryProcessor(db)
431 session = Session.Context()
432 session.wfile = sys.stdout
435 line = sys.stdin.readline().strip();
437 if line.startswith("#"): continue
439 print "parsing: '%s'" % line
440 processor.process_query(session, line)
441 session.wfile.write("\r\n");
442 session.wfile.flush()