def batchQuery(outf,r,d,ref_db,queries,coord_file,matrix_file,targetNames, embedOnly=False ):
query_cdb="query.cdb"
puzzle_file="puzzle"
queries_cdb="queries.cdb"
createPuzzleFile(r,d,coord_file,puzzle_file)
parsing_time = 0
refine_time = 0
coord_time,solver = time_function(CoordinateSolver,puzzle_file)
if not embedOnly:
lsh_time,lshsearcher = time_function(LSHSearcher,matrix_file,lsh_param)
else:
lsh_time=0
parsing_time += time_function(createQueryCdb,queries,queries_cdb)[0]
queryNames = [name.strip() for name in file(queries_cdb+".names")]
debug("targetNames: "+str(targetNames))
debug("queryNames: "+str(queryNames))
dist_time,subp = time_function(Popen, [DB2DB_DISTANCE,queries_cdb,ref_db ],stdout=PIPE)
queryIndex=0
for line in subp.stdout:
queryIndex += 1
t,solverResult = time_function(solver.solve,line)
coord_time += t
debug("solverResult: "+solverResult)
if embedOnly:
print solverResult
continue
t,lshResult = time_function(lshsearcher.search,solverResult)
lsh_time += t
lshResult = lshResult.strip()
info("lshResult %d: %s" % (queryIndex-1,lshResult))
f=file("query.iddb","w")
f.write(str(queryIndex)+"\n")
f.close()
check_call([DB_SUBSET,queries_cdb,"query.iddb",query_cdb])
t,distances = time_function(refine,QueryFile(query_cdb),lshResult)
refine_time+=t
for candidate_index,dist in distances:
outf.write("%s\t%s\t%s\n" %
(queryNames[queryIndex-1],targetNames[candidate_index-1],dist))
sys.stderr.write('timing: parsing=%s embedding=%s lsh=%s refine=%s \n' %
(parsing_time, dist_time + coord_time, lsh_time, refine_time))
def indexed_search(matrix_file,
coord_query_file,
output="indexed.gz",
evaluation_out="indexed.performance"):
"""perform indexed search"""
info("running indexed search")
# create subset from TEST_QUERIES
# select TEST_QUERIES from matrix file (matrix_file)
# find neighbors using lshServer
# refine results
# compare with gen_chemial_search_results
def doTest():
from gzip import GzipFile as zfile
test_queries_indicies = [
int(value) - 1 for value in file(TEST_QUERIES)
]
lshsearcher = LSHSearcher(matrix_file, lsh_param)
out = zfile(output, "w")
line_num = 0
for query_coords in file(coord_query_file):
info("query_coords( " + str(test_queries_indicies[line_num]) +
"): " + query_coords)
candidates = lshsearcher.search(query_coords)
query_iddb = file("iquery.iddb", "w")
query_iddb.write("%d\n" % test_queries_indicies[line_num])
query_iddb.close()
results = refine(QueryIDDB("iquery.iddb"), candidates)
os.unlink("iquery.iddb")
out.write(" ".join(["%d:%f" % pair for pair in results]) + "\n")
line_num += 1
out.close()
test_time = time_function(doTest)[0]
#clean up stuff left by LSHServer
if os.path.isfile("coord.in"): os.unlink("coord.in")
info("total time: %.1f" % test_time)
cmd = "echo %s > index.search.timing" % test_time
os_run(cmd)
gen_chemical_search_results()
cmd = "%s %s %s > %s" % (INDEXED_SEARCH_EVALUATOR, CHEMICAL_SEARCH_RESULTS,
output, evaluation_out)
os_run(cmd)
return output
def indexed_search(matrix_file, coord_query_file,output="indexed.gz",
evaluation_out="indexed.performance"):
"""perform indexed search"""
info("running indexed search")
# create subset from TEST_QUERIES
# select TEST_QUERIES from matrix file (matrix_file)
# find neighbors using lshServer
# refine results
# compare with gen_chemial_search_results
def doTest():
from gzip import GzipFile as zfile
test_queries_indicies = [int(value)-1 for value in file(TEST_QUERIES)]
lshsearcher = LSHSearcher(matrix_file,lsh_param)
out = zfile(output,"w")
line_num=0
for query_coords in file(coord_query_file):
info("query_coords( "+str(test_queries_indicies[line_num])+"): "+query_coords)
candidates = lshsearcher.search(query_coords)
query_iddb=file("iquery.iddb","w")
query_iddb.write("%d\n" % test_queries_indicies[line_num])
query_iddb.close()
results = refine(QueryIDDB("iquery.iddb"),candidates)
os.unlink("iquery.iddb")
out.write(" ".join(["%d:%f" % pair for pair in results])+"\n")
line_num+=1
out.close()
test_time = time_function(doTest)[0]
#clean up stuff left by LSHServer
if os.path.isfile("coord.in"): os.unlink("coord.in")
info( "total time: %.1f" % test_time )
cmd = "echo %s > index.search.timing" % test_time
os_run(cmd)
gen_chemical_search_results()
cmd = "%s %s %s > %s" % (INDEXED_SEARCH_EVALUATOR,
CHEMICAL_SEARCH_RESULTS, output, evaluation_out)
os_run(cmd)
return output
def batchQuery(outf,
r,
d,
ref_db,
queries,
coord_file,
matrix_file,
targetNames,
embedOnly=False):
query_cdb = "query.cdb"
puzzle_file = "puzzle"
queries_cdb = "queries.cdb"
createPuzzleFile(r, d, coord_file, puzzle_file)
parsing_time = 0
refine_time = 0
coord_time, solver = time_function(CoordinateSolver, puzzle_file)
if not embedOnly:
lsh_time, lshsearcher = time_function(LSHSearcher, matrix_file,
lsh_param)
else:
lsh_time = 0
parsing_time += time_function(createQueryCdb, queries, queries_cdb)[0]
queryNames = [name.strip() for name in file(queries_cdb + ".names")]
debug("targetNames: " + str(targetNames))
debug("queryNames: " + str(queryNames))
dist_time, subp = time_function(Popen,
[DB2DB_DISTANCE, queries_cdb, ref_db],
stdout=PIPE)
queryIndex = 0
for line in subp.stdout:
queryIndex += 1
t, solverResult = time_function(solver.solve, line)
coord_time += t
debug("solverResult: " + solverResult)
if embedOnly:
print solverResult
continue
t, lshResult = time_function(lshsearcher.search, solverResult)
lsh_time += t
lshResult = lshResult.strip()
info("lshResult %d: %s" % (queryIndex - 1, lshResult))
f = file("query.iddb", "w")
f.write(str(queryIndex) + "\n")
f.close()
check_call([DB_SUBSET, queries_cdb, "query.iddb", query_cdb])
t, distances = time_function(refine, QueryFile(query_cdb), lshResult)
refine_time += t
for candidate_index, dist in distances:
outf.write("%s\t%s\t%s\n" %
(queryNames[queryIndex - 1],
targetNames[candidate_index - 1], dist))
sys.stderr.write(
'timing: parsing=%s embedding=%s lsh=%s refine=%s \n' %
(parsing_time, dist_time + coord_time, lsh_time, refine_time))