def query_lowlevel(db, param_dict, sort_order):
vars = []
clauses = []
for column, values in param_dict.items():
norm_path = db.path + "/frequencies/normalized_" + column + "_frequencies"
for v in values:
parsed = parse_query(v)
if db.locals['debug']:
print >> sys.stderr, "METADATA_TOKENS:", parsed
grouped = group_terms(parsed)
if db.locals['debug']:
print >> sys.stderr, "METADATA_SYNTAX GROUPED:", grouped
expanded = expand_grouped_query(grouped, norm_path)
if db.locals['debug']:
print >> sys.stderr, "METADATA_SYNTAX EXPANDED:", expanded
sql_clause = make_grouped_sql_clause(expanded, column)
if db.locals['debug']:
print >> sys.stderr, "SQL_SYNTAX:", sql_clause
clauses.append(sql_clause)
if not sort_order:
sort_order = ["rowid"]
if clauses:
query = "SELECT philo_id FROM toms WHERE " + " AND ".join("(%s)" % c for c in clauses) + " order by %s;" % ", ".join(sort_order)
else:
query = "SELECT philo_id FROM toms order by %s;" % ", ".join(sort_order)
if db.locals['debug']:
print >> sys.stderr, "INNER QUERY: ", "%s %% %s" % (query, vars), sort_order
results = db.dbh.execute(query, vars)
return results
def query_lowlevel(db,param_dict):
vars = []
clauses = []
# if column = _philo_id I can do a special query here
for column,values in param_dict.items():
norm_path = db.path+"/frequencies/normalized_" + column + "_frequencies"
for v in values:
parsed = parse_query(v)
print >> sys.stderr, "METADATA_TOKENS:", parsed
grouped = group_terms(parsed)
print >> sys.stderr, "METADATA_SYNTAX:", grouped
expanded = expand_grouped_query(grouped,norm_path)
print >> sys.stderr, "METADATA_SYNTAX:", expanded
sql_clause = make_grouped_sql_clause(expanded,column)
print >> sys.stderr, "SQL_SYNTAX:", sql_clause
clauses.append(sql_clause)
# clause,some_vars = make_clause(column,parsed,norm_path)
# print >> sys.stderr, "METADATA_QUERY:",clause,some_vars
# clauses.append(clause)
# vars += some_vars
if clauses:
query = "SELECT philo_id FROM toms WHERE " + " AND ".join("(%s)" % c for c in clauses) + " order by rowid;"
else:
query = "SELECT philo_id FROM toms order by rowid;"
# vars = [v.decode("utf-8") for v in vars]
print >> sys.stderr, "INNER QUERY: ", "%s %% %s" % (query,vars)
#for v in vars:
# print >> sys.stderr, "%s : %s" % (type(v),repr(v))
results = db.dbh.execute(query,vars)
return results
def query_lowlevel(db, param_dict):
vars = []
clauses = []
# if column = _philo_id I can do a special query here
for column, values in param_dict.items():
norm_path = db.locals[
"db_path"] + "/frequencies/normalized_" + column + "_frequencies"
for v in values:
parsed = parse_query(v)
print >> sys.stderr, "METADATA_TOKENS:", parsed
grouped = group_terms(parsed)
print >> sys.stderr, "METADATA_SYNTAX:", grouped
expanded = expand_grouped_query(grouped, norm_path)
print >> sys.stderr, "METADATA_SYNTAX:", expanded
sql_clause = make_grouped_sql_clause(expanded, column)
print >> sys.stderr, "SQL_SYNTAX:", sql_clause
clauses.append(sql_clause)
# clause,some_vars = make_clause(column,parsed,norm_path)
# print >> sys.stderr, "METADATA_QUERY:",clause,some_vars
# clauses.append(clause)
# vars += some_vars
if clauses:
query = "SELECT philo_id FROM toms WHERE " + " AND ".join(
"(%s)" % c for c in clauses) + " order by rowid;"
else:
query = "SELECT philo_id FROM toms order by rowid;"
# vars = [v.decode("utf-8") for v in vars]
print >> sys.stderr, "INNER QUERY: ", "%s %% %s" % (query, vars)
#for v in vars:
# print >> sys.stderr, "%s : %s" % (type(v),repr(v))
results = db.dbh.execute(query, vars)
return results
def query(db,terms,corpus_file=None,corpus_size=0,method=None,method_arg=None,limit=3000,filename="", query_debug=False):
sys.stdout.flush()
tstart = datetime.now()
parsed = parse_query(terms)
grouped = group_terms(parsed)
split = split_terms(grouped)
# print >> sys.stderr, "QUERY FORMATTED at ", datetime.now() - tstart
words_per_hit = len(split)
# print >> sys.stderr, "QUERY SPLIT at ", datetime.now() - tstart, repr(split)
origpid = os.getpid()
if not filename:
hfile = str(origpid) + ".hitlist"
dir = db.path + "/hitlists/"
filename = filename or (dir + hfile)
hl = open(filename, "w")
err = open("/dev/null", "w")
freq_file = db.path+"/frequencies/normalized_word_frequencies"
if (query_debug):
print >> sys.stderr, "FORKING"
pid = os.fork()
if pid == 0:
os.umask(0)
os.chdir(dir)
os.setsid()
pid = os.fork()
if pid > 0:
os._exit(0)
else:
#now we're detached from the parent, and can do our work.
if query_debug:
print >> sys.stderr, "WORKER DETACHED at ", datetime.now() - tstart
# args = ["search4", db.path,"--limit",str(limit)]
args = ["corpus_search"]
if corpus_file:
args.extend(("-c", corpus_file));
# if corpus_file and corpus_size:
# args.extend(("--corpusfile", corpus_file , "--corpussize" , str(corpus_size)))
if method and method_arg:
args.extend(("-m",method,"-a",str(method_arg)))
args.extend(("-o","binary",db.path,));
worker = subprocess.Popen(args,stdin=subprocess.PIPE,stdout=hl,stderr=err)
# if query_debug:
# print >> sys.stderr, "WORKER STARTED:"," ".join(args);
query_log_fh = filename + ".terms"
print >> sys.stderr, "LOGGING TERMS to " + filename + ".terms"
logger = subprocess.Popen(["tee",query_log_fh],stdin=subprocess.PIPE,stdout = worker.stdin)
# print >> sys.stderr, "EXPANDING"
expand_query_not(split,freq_file,logger.stdin)
logger.stdin.close()
# if query_debug == True:
# print >> sys.stderr, "DEBUGGING"
# query_log_fh = filename + ".terms"
# print >> sys.stderr, "LOGGING to " + filename + ".terms"
# logger = subprocess.Popen(["tee",query_log_fh],stdin=subprocess.PIPE,stdout = worker.stdin)
# print >> sys.stderr, "EXPANDING"
# expand_query_not(split,freq_file,logger.stdin)
# logger.stdin.close()
# else:
# expand_query_not(split,freq_file,worker.stdin)
worker.stdin.close()
returncode = worker.wait()
if returncode == -11:
print >> sys.stderr, "SEGFAULT"
seg_flag = open(filename + ".error","w")
seg_flag.close()
#do something to mark query as finished
flag = open(filename + ".done","w")
flag.write(" ".join(args) + "\n")
flag.close()
# print >> sys.stderr, "SUBPROC DONE at ", datetime.now() - tstart
os._exit(0)
else:
hl.close()
return HitList.HitList(filename,words_per_hit,db)
return grep_proc
def invert_grep_exact(token, in_fh, dest_fh):
#don't strip accent or case, exact match only.
grep_command = ["egrep", "-v", "[[:blank:]]%s$" % token[1:-1]]
print >> sys.stderr, grep_command
grep_proc = subprocess.Popen(grep_command,stdin=in_fh,stdout=dest_fh)
#can't wait because input isn't ready yet.
return grep_proc
if __name__ == "__main__":
path = sys.argv[1]
terms = sys.argv[2:]
parsed = parse_query(" ".join(terms))
print >> sys.stderr, "PARSED:", parsed
grouped = group_terms(parsed)
print >> sys.stderr, "GROUPED:", grouped
split = split_terms(grouped)
print >> sys.stderr, "parsed %d terms:" % len(split), split
class Fake_DB: pass
fake_db = Fake_DB()
fake_db.locals = {"db_path":path + "/data/"}
fake_db.path = path + "/data/"
fake_db.encoding = "utf-8"
freq_file = path + "/data/frequencies/normalized_word_frequencies"
# freq_file = "/Library/WebServer/Documents/philologic/plain_text_test/data/frequencies/normalized_word_frequencies"
expand_query_not(split,freq_file,sys.stdout)
hits = query(fake_db," ".join(terms),query_debug = True)
def query(db,
terms,
corpus_file=None,
corpus_size=0,
method=None,
method_arg=None,
limit=3000,
filename="",
query_debug=False):
sys.stdout.flush()
tstart = datetime.now()
parsed = parse_query(terms)
grouped = group_terms(parsed)
split = split_terms(grouped)
# print >> sys.stderr, "QUERY FORMATTED at ", datetime.now() - tstart
words_per_hit = len(split)
# print >> sys.stderr, "QUERY SPLIT at ", datetime.now() - tstart, repr(split)
origpid = os.getpid()
if not filename:
hfile = str(origpid) + ".hitlist"
dir = db.locals["db_path"] + "/hitlists/"
filename = filename or (dir + hfile)
hl = open(filename, "w")
err = open("/dev/null", "w")
freq_file = db.locals[
"db_path"] + "/frequencies/normalized_word_frequencies"
pid = os.fork()
if pid == 0:
os.umask(0)
os.chdir(dir)
os.setsid()
pid = os.fork()
if pid > 0:
os._exit(0)
else:
#now we're detached from the parent, and can do our work.
print >> sys.stderr, "WORKER DETACHED at ", datetime.now() - tstart
args = ["search4", db.path, "--limit", str(limit)]
if corpus_file and corpus_size:
args.extend(("--corpusfile", corpus_file, "--corpussize",
str(corpus_size)))
if method and method_arg:
args.extend((method, str(method_arg)))
worker = subprocess.Popen(args,
stdin=subprocess.PIPE,
stdout=hl,
stderr=err)
print >> sys.stderr, "WORKER STARTED"
if query_debug == True:
print >> sys.stderr, "DEBUGGING"
query_log_fh = filename + ".terms"
print >> sys.stderr, "LOGGING to " + filename + ".terms"
logger = subprocess.Popen(["tee", query_log_fh],
stdin=subprocess.PIPE,
stdout=worker.stdin)
print >> sys.stderr, "EXPANDING"
expand_query(split, freq_file, logger.stdin)
logger.stdin.close()
else:
expand_query(split, freq_file, worker.stdin)
worker.stdin.close()
returncode = worker.wait()
if returncode == -11:
print >> sys.stderr, "SEGFAULT"
seg_flag = open(filename + ".error", "w")
seg_flag.close()
#do something to mark query as finished
flag = open(filename + ".done", "w")
flag.write(" ".join(args) + "\n")
flag.close()
# print >> sys.stderr, "SUBPROC DONE at ", datetime.now() - tstart
os._exit(0)
else:
hl.close()
return HitList.HitList(filename, words_per_hit, db)