def query(rt, query_str, profiler):
idx = rt["index"]
docs = rt["documents"]
start = time.clock()
if profiler is None:
results, num_results = idx.query(query_str, BEST)
else:
if WARM_CACHE:
print "Warming the cache..."
idx.query(query_str, BEST)
start = time.clock()
results, num_results = profiler.runcall(idx.query, query_str, BEST)
elapsed = time.clock() - start
print "query:", query_str
print "# results:", len(results), "of", num_results, \
"in %.2f ms" % (elapsed * 1000)
tree = QueryParser(idx.lexicon).parseQuery(query_str)
qw = idx.index.query_weight(tree.terms())
for docid, score in results:
scaled = 100.0 * score / qw
print "docid %7d score %6d scaled %5.2f%%" % (docid, score, scaled)
if VERBOSE:
msg = docs[docid]
ctx = msg.text.split("\n", CONTEXT)
del ctx[-1]
print "-" * 60
print "message:"
for l in ctx:
print l
print "-" * 60
def query(rt, query_str, profiler):
idx = rt["index"]
docs = rt["documents"]
start = time.clock()
if profiler is None:
results, num_results = idx.query(query_str, BEST)
else:
if WARM_CACHE:
print "Warming the cache..."
idx.query(query_str, BEST)
start = time.clock()
results, num_results = profiler.runcall(idx.query, query_str, BEST)
elapsed = time.clock() - start
print "query:", query_str
print "# results:", len(results), "of", num_results, \
"in %.2f ms" % (elapsed * 1000)
tree = QueryParser(idx.lexicon).parseQuery(query_str)
qw = idx.index.query_weight(tree.terms())
for docid, score in results:
scaled = 100.0 * score / qw
print "docid %7d score %6d scaled %5.2f%%" % (docid, score, scaled)
if VERBOSE:
msg = docs[docid]
ctx = msg.text.split("\n", CONTEXT)
del ctx[-1]
print "-" * 60
print "message:"
for l in ctx:
print l
print "-" * 60
def _ranking_queries(self):
queries = ['eat', 'porridge', 'hot OR porridge',
'eat OR nine OR day OR old OR porridge']
wqs = [1.95, 1.10, 1.77, 3.55]
results = [[(6, 0.71)],
[(1, 0.61), (2, 0.58), (5, 0.71)],
[(1, 0.66), (2, 0.36), (4, 0.36), (5, 0.44)],
[(1, 0.19), (2, 0.18), (3, 0.63), (5, 0.22), (6, 0.39)]]
for i in range(len(queries)):
raw = queries[i]
q = QueryParser(self.lexicon).parseQuery(raw)
wq = self.index.query_weight(q.terms())
eq(wq, scaled_int(wqs[i]))
r, n = self.zc_index.query(raw)
self.assertEqual(len(r), len(results[i]))
# convert the results to a dict for each checking
d = {}
for doc, score in results[i]:
d[doc] = scaled_int(score)
for doc, score in r:
score = scaled_int(float(score / SCALE_FACTOR) / wq)
self.assertTrue(0 <= score <= SCALE_FACTOR)
eq(d[doc], score)
