1 # This file is part of python-rwhoisd
3 # Copyright (C) 2003, David E. Blacka
5 # This program is free software; you can redistribute it and/or modify
6 # it under the terms of the GNU General Public License as published by
7 # the Free Software Foundation; either version 2 of the License, or
8 # (at your option) any later version.
10 # This program is distributed in the hope that it will be useful, but
11 # WITHOUT ANY WARRANTY; without even the implied warranty of
12 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 # General Public License for more details.
15 # You should have received a copy of the GNU General Public License
16 # along with this program; if not, write to the Free Software
17 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
21 import Cidr, Rwhois, QueryParser
25 def __init__(self, db):
28 def _filter_obj_term(self, obj, term):
29 """Given a rwhoisobject and a query term (a 3 element tuple:
30 attr, operator, value), determine if the object satisfies the
31 term. Returns True if the object matches the term, False if
34 attr, op, searchval = term
37 # filter by named attribute
39 vals = obj.get_attr(attr)
43 res = match_values(searchval, vals)
44 if op == "!=": return not res
46 # filter by general term
48 for val in obj.values():
49 if match_value(searchval, val):
53 def _filter_obj(self, obj, terms):
54 """Given a rwhoisobject and a list of query terms (i.e., a
55 whole AND clause), return True if the object satisfies the
59 if not self._filter_obj_term(obj, term): return False
62 def _filter_results(self, reslist, terms):
63 """Given list of result objects (not simply the ids returned
64 from the search) and a list of query terms (i.e., a query
65 clause), remove elements that do not satisfy the terms.
66 Returns a list of objects that satisfy the filters."""
68 if not terms: return reslist
69 return [ x for x in reslist if self._filter_obj(x, terms) ]
71 def process_query_clause(self, clause, max=0):
72 """Process a query clause (a grouping of terms ANDed
73 together). This is where the indexed searches actually get
74 done. The technique used here is to search on one index and
75 use the rest of the clause to filter the results. Returns a
78 # the technique is to do an index search on the first (or
79 # maybe best) indexed term (bare terms are always considered
80 # indexed), and filter those results with the remaining terms.
82 # Note: this could be better if we found the "optimal" query
83 # term. One approach may be to create a cost function and
84 # search for the minimum cost term.
86 # Note: another approach might be to actually do indexed
87 # searches on all applicable terms (bare or using an indexed
88 # attribute) and find the intersection of the results.
93 orig_clause = clause[:]
95 # find the first searchable term:
96 for term, i in zip(clause, xrange(sys.maxint)):
97 attr, op, value = term
98 if op == "!=": continue
99 if not attr or self.db.is_indexed_attr(attr):
103 raise Rwhois.RwhoisError, (351, "No indexed terms in query clause")
105 # remove the search term from the clause, what remains is the
109 # if we have an attribute name, search on that.
111 res = self.db.search_attr(st[0], st[2], max)
113 if Cidr.valid_cidr(st[2].strip("*")):
114 res = self.db.search_cidr(st[2], max)
116 res = self.db.search_normal(st[2], max)
118 objs = self._filter_results(self.db.fetch_objects(res.list()), clause)
120 queryres = QueryResult(objs)
123 refs = self.process_referral_search(orig_clause)
124 queryres.add_referrals(refs)
128 def _is_in_autharea(self, value):
129 """Returns True if value could be considered to be contained
130 within an authority area. That is, is a subnet of a
131 network-type authority area or a subdomain of a domainname
132 type authority area."""
134 aas = self.db.get_authareas()
136 if isinstance(value, Cidr.Cidr):
138 cv = Cidr.valid_cidr(aa)
139 if cv and cv.is_supernet(value):
143 if is_domainname(aa) and is_subdomain(aa, value):
147 def _referral_search_cidr(self, cv, value):
148 """Return the IndexResult of a referral search for value, or
149 None if the value doesn't qualify for a Cidr referral
152 if not cv: return None
153 if not self._is_in_autharea(cv): return None
154 return self.db.search_referral(value)
156 def _referral_search_domain(self, value):
157 """Return the IndexResult of a referral search for value, or
158 None if the value doesn't qualify for a domain referral
161 if not is_domainname(value): return None
162 if not self._is_in_autharea(value): return None
166 res = self.db.search_referral(dn)
168 dn = reduce_domain(dn)
171 def _referral_search_term(self, value):
172 """Return the IndexResult of a referral search for value, or
173 None if the value didn't qualify for a referral search."""
175 cv = Cidr.valid_cidr(value)
177 return self._referral_search_cidr(cv, value)
178 elif is_domainname(value):
179 return self._referral_search_domain(value)
182 def process_referral_search(self, clause):
183 """Given a query clause, attempt to search for referrals
184 associated with the terms. Return a list of referral strings
185 that matched terms in the clause (if any). The only terms
186 that actually get searched are the ones that look
187 'heirarchical'. For now, the attribute part of the term is
188 essentially ignored, so a search for something like
189 'name=127.0.0.1' might concievably generate a referral, when
190 perhaps it shouldn't."""
192 # first check to see if the search is explictly for a referral
194 if (term[0] == "class-name" and term[1] == "="
195 and term[2] == "referral") or term[0] == "referred-auth-area":
196 # in which case, we return nothing
201 # look for heirarchical-looking terms.
202 for attr, op, value in clause:
203 if op == "!=": continue
204 res = self._referral_search_term(value)
205 if not res or not res.list():
208 ref_objs = self.db.fetch_objects(res.list())
209 ref_strs = [x for y in ref_objs for x in y.get_attr("referral")]
210 referrals.extend(ref_strs)
215 def process_full_query(self, query, max=0):
216 """Given a parsed query object, process it by unioning the
217 results of the various ORed together clauses"""
219 # shortcut for the very common single clause case:
220 if len(query.clauses) == 1:
221 res = self.process_query_clause(query.clauses[0], max)
224 # otherwise, union the results from all the causes
226 for clause in query.clauses:
227 res.extend(self.process_query_clause(clause))
228 if max and len(res) >= max:
234 def process_query(self, session, queryline):
235 """Given a session config and a query line, parse the query,
236 perform any searches, return any referrals."""
238 if not session.queryparser:
239 session.queryparser = QueryParser.get_parser()
243 query = QueryParser.parse(session.queryparser, queryline)
244 except Rwhois.RwhoisError, x:
245 session.wfile.write(Rwhois.error_message(x))
252 query_result = self.process_full_query(query, max)
253 except Rwhois.RwhoisError, x:
254 session.wfile.write(Rwhois.error_message(x))
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()