def test___init__sets_all(self):
PARAM_NUM = 16 # how to get dynamically?
with self.assertRaises(TypeError):
results.Result(*range(PARAM_NUM + 1))
result = results.Result(*range(PARAM_NUM - 1), src={})
self.assertEqual(
PARAM_NUM,
len([x for x in result.__dict__.values() if x is not None]))
def dorequest(self, timeout=TIMEOUT, HttpMethod="POST", parsexsams=True):
"""
Sends the request to the database node and returns a result.Result instance. The
request uses 'POST' requests by default. If the request fails or if stated in the parameter 'HttpMethod',
'GET' requests will be performed.
The returned result will be parsed by default and the model defined in 'specmodel' will be populated by default
(parseexams = True).
"""
self.xml = None
#self.get_xml(self.Source.Requesturl)
url = self.baseurl + self.querypath
urlobj = urlsplit(url)
if urlobj.scheme == 'https':
conn = HTTPSConnection(urlobj.netloc, timeout=timeout)
else:
conn = HTTPConnection(urlobj.netloc, timeout=timeout)
conn.putrequest(HttpMethod, urlobj.path + "?" + urlobj.query)
conn.endheaders()
try:
res = conn.getresponse()
except socket.timeout:
# error handling has to be included
self.status = 408
self.reason = "Socket timeout"
raise TimeOutError
self.status = res.status
self.reason = res.reason
if not parsexsams:
if res.status == 200:
result = r.Result()
result.Content = res.read()
elif res.status == 400 and HttpMethod == 'POST':
# Try to use http-method: GET
result = self.dorequest(HttpMethod='GET',
parsexsams=parsexsams)
else:
result = None
else:
if res.status == 200:
self.xml = res.read()
result = r.Result()
result.Xml = self.xml
result.populate_model()
elif res.status == 400 and HttpMethod == 'POST':
# Try to use http-method: GET
result = self.dorequest(HttpMethod='GET',
parsexsams=parsexsams)
else:
result = None
return result
def get_species(self):
"""
Queries all species from the database-node via TAP-XSAMS request and
Query 'Select Species'. The list of species is saved in object species.
Note: This does not work for all species !
"""
# Some nodes do not understand this query at all
# others do not understand the query SELECT SPECIES
# therefore the following query is a small workaround for some nodes
query=q.Query("SELECT SPECIES WHERE ((InchiKey!='UGFAIRIUMAVXCW'))")
query.set_node(self)
result = r.Result()
result.set_query(query)
result.do_query()
result.populate_model()
try:
self.Molecules = result.data['Molecules']
except:
pass
try:
self.Atoms = result.data['Atoms']
except:
pass
def get_potential_words(self, word, k):
'''Return a list of potential words matching input as a ResultList.
:param word: input to find matches for
:param k: maximum error numbers to search against (recommend 2 max)
'''
result_list = results.ResultList()
startnode = self.rootNode
word = word + self.chr0
for key, next_node in startnode.child_nodes.iteritems():
result = results.Result()
self._fuzzy_search_r(next_node, word, 0, result, result_list, k)
return result_list
def check_for_updates(self, node):
"""
"""
count_updates = 0
counter = 0
#species_list = []
cursor = self.conn.cursor()
cursor.execute("SELECT PF_Name, PF_SpeciesID, PF_VamdcSpeciesID, datetime(PF_Timestamp) FROM Partitionfunctions ")
rows = cursor.fetchall()
num_rows = len(rows)
query = q.Query()
result = results.Result()
for row in rows:
counter += 1
print "%5d/%5d: Check specie %-55s (%-15s): " % (counter, num_rows, row[0], row[1]),
#id = row[1]
vamdcspeciesid = row[2]
# query_string = "SELECT ALL WHERE VAMDCSpeciesID='%s'" % vamdcspeciesid
query_string = "SELECT ALL WHERE SpeciesID=%s" % row[1][6:]
query.set_query(query_string)
query.set_node(node)
result.set_query(query)
try:
changedate = result.getChangeDate()
except:
changedate = None
tstamp = parser.parse(row[3] + " GMT")
if changedate is None:
print " -- UNKNOWN (Could not retrieve information)"
continue
if tstamp < changedate:
print " -- UPDATE AVAILABLE "
count_updates += 1
else:
print " -- up to date"
if count_updates == 0:
print "\r No updates for your entries available"
print "Done"
def update_database(self, add_nodes = None, insert_only = False, update_only = False):
"""
Checks if there are updates available for all entries. Updates will
be retrieved from the resource specified in the database.
All resources will be searched for new entries, which will be inserted
if available. Additional resources can be specified via add_nodes.
add_nodes: Single or List of node-instances (nodes.Node)
"""
# counter to identify which entry is currently processed
counter = 0
# counter to count available updates
count_updates = 0
# list of database - nodes which are currently in the local database
dbnodes = []
# create an instance with all available vamdc-nodes
nl = nodes.Nodelist()
# attach additional nodes to the list of dbnodes (for insert)
if not functions.isiterable(add_nodes):
add_nodes = [add_nodes]
for node in add_nodes:
if node is None:
pass
elif not isinstance(node, nodes.Node):
print "Could not attach node. Wrong type, it should be type <nodes.Node>"
else:
dbnodes.append(node)
#--------------------------------------------------------------------
# Check if updates are available for entries
# Get list of species in the database
cursor = self.conn.cursor()
cursor.execute("SELECT PF_Name, PF_SpeciesID, PF_VamdcSpeciesID, datetime(PF_Timestamp), PF_ResourceID FROM Partitionfunctions ")
rows = cursor.fetchall()
num_rows = len(rows)
query = q.Query()
result = results.Result()
if not insert_only:
print("----------------------------------------------------------")
print "Looking for updates"
print("----------------------------------------------------------")
for row in rows:
counter += 1
print "%5d/%5d: Check specie %-55s (%-15s): " % (counter, num_rows, row[0], row[1]),
try:
node = nl.getnode(str(row[4]))
except:
node = None
if node is None:
print " -- RESOURCE NOT AVAILABLE"
continue
else:
if node not in dbnodes:
dbnodes.append(node)
vamdcspeciesid = row[2]
query_string = "SELECT ALL WHERE SpeciesID=%s" % row[1][6:]
query.set_query(query_string)
query.set_node(node)
result.set_query(query)
try:
changedate = result.getChangeDate()
except:
changedate = None
tstamp = parser.parse(row[3] + " GMT")
if changedate is None:
print " -- UNKNOWN (Could not retrieve information)"
continue
if tstamp < changedate:
print " -- UPDATE AVAILABLE "
count_updates += 1
print " -- PERFORM UPDATE -- "
query_string = "SELECT SPECIES WHERE SpeciesID=%s" % row[1][6:]
query.set_query(query_string)
query.set_node(node)
result.set_query(query)
result.do_query()
result.populate_model()
insert_species_data(result.data['Molecules'], update = True)
print " -- UPDATE DONE -- "
else:
print " -- up to date"
if count_updates == 0:
print "\r No updates for your entries available"
print "Done"
else:
cursor.execute("SELECT distinct PF_ResourceID FROM Partitionfunctions ")
rows = cursor.fetchall()
for row in rows:
try:
node = nl.getnode(str(row[0]))
except:
node = None
if node is None:
print " -- RESOURCE NOT AVAILABLE"
continue
else:
if node not in dbnodes:
dbnodes.append(node)
if update_only:
return
# Check if there are new entries available
#---------------------------------------------------------
# Check all dbnodes for new species
counter = 0
insert_molecules_list = []
for node in dbnodes:
print("----------------------------------------------------------")
print "Query '{dbname}' for new species ".format(dbname=node.name)
print("----------------------------------------------------------")
node.get_species()
for id in node.Molecules:
try:
cursor.execute("SELECT PF_Name, PF_SpeciesID, PF_VamdcSpeciesID, PF_Timestamp FROM Partitionfunctions WHERE PF_SpeciesID=?", [(id)])
exist = cursor.fetchone()
if exist is None:
print " %s" % node.Molecules[id]
insert_molecules_list.append(node.Molecules[id])
counter += 1
except Exception, e:
print e
print id
print "There are %d new species available" % counter
print("----------------------------------------------------------")
print "Start insert"
print("----------------------------------------------------------")
self.insert_species_data(insert_molecules_list, node)
print("----------------------------------------------------------")
print "Done"
def insert_species_data(self, species, node, update=False):
"""
Inserts new species into the local database
species: species which will be inserted
node: vamdc-node / type: instance(nodes.node)
update: if True then all entries in the local database with the same
species-id will be deleted before the insert is performed.
"""
# create a list of names. New names have not to be in that list
names_black_list = []
cursor = self.conn.cursor()
cursor.execute("SELECT PF_Name FROM Partitionfunctions")
rows = cursor.fetchall()
for row in rows:
names_black_list.append(row[0])
#----------------------------------------------------------
# Create a list of species for which transitions will be
# retrieved and inserted in the database.
# Species have to be in the Partitionfunctions - table
if not functions.isiterable(species):
species = [species]
#--------------------------------------------------------------
for specie in species:
num_transitions = {}
# will contain a list of names which belong to one specie
species_names = {}
# list will contain species whose insert-failed
species_with_error = []
# check if specie is of type Molecule
if isinstance(specie, specmodel.Molecule):
speciesid = specie.SpeciesID
vamdcspeciesid = specie.VAMDCSpeciesID
formula = specie.OrdinaryStructuralFormula
else:
try:
if isinstance(specie, str) and len(specie) == 27:
vamdcspeciesid = specie
speciesid = None
except:
print "Specie is not of wrong type"
print "Type Molecule or string (Inchikey) is allowed"
continue
if speciesid:
print "Processing: {speciesid}".format(speciesid = speciesid)
else:
print "Processing: {vamdcspeciesid}".format(vamdcspeciesid = vamdcspeciesid)
try:
# Create query string
query_string = "SELECT ALL WHERE VAMDCSpeciesID='%s'" % vamdcspeciesid
query = q.Query()
result = results.Result()
# Get data from the database
query.set_query(query_string)
query.set_node(node)
result.set_query(query)
result.do_query()
result.populate_model()
except:
print " -- Error: Could not fetch and process data"
continue
#---------------------------------------
cursor = self.conn.cursor()
cursor.execute('BEGIN TRANSACTION')
#------------------------------------------------------------------------------------------------------
# if update is allowed then all entries in the database for the given species-id will be
# deleted, and thus replaced by the new data
if update:
cursor.execute("SELECT PF_Name FROM Partitionfunctions WHERE PF_SpeciesID = ?", (speciesid, ))
rows = cursor.fetchall()
for row in rows:
names_black_list.remove(row[0])
cursor.execute("DELETE FROM Transitions WHERE T_Name = ?", (row[0], ))
cursor.execute("DELETE FROM Partitionfunctions WHERE PF_Name = ?", (row[0], ))
#------------------------------------------------------------------------------------------------------
#------------------------------------------------------------------------------------------------------
# Insert all transitions
num_transitions_found = len(result.data['RadiativeTransitions'])
counter_transitions = 0
for trans in result.data['RadiativeTransitions']:
counter_transitions+=1
print "\r insert transition %d of %d" % (counter_transitions, num_transitions_found),
# data might contain transitions for other species (if query is based on ichikey/vamdcspeciesid).
# Insert transitions only if they belong to the correct specie
if result.data['RadiativeTransitions'][trans].SpeciesID == speciesid or speciesid is None:
id = str(result.data['RadiativeTransitions'][trans].SpeciesID)
# if an error has occured already then there will be no further insert
if id in species_with_error:
continue
formula = str(result.data['Molecules'][id].OrdinaryStructuralFormula)
# Get upper and lower state from the states table
try:
upper_state = result.data['States']["%s" % result.data['RadiativeTransitions'][trans].UpperStateRef]
lower_state = result.data['States']["%s" % result.data['RadiativeTransitions'][trans].LowerStateRef]
except (KeyError, AttributeError):
print " -- Error: State is missing"
species_with_error.append(id)
continue
# Get string which identifies the vibrational states involved in the transition
t_state = self.getvibstatelabel(upper_state, lower_state)
# Get hyperfinestructure info if hfsInfo is None
# only then the hfsInfo has not been inserted in the species name
# (there can be multiple values in the complete dataset
t_hfs = ''
try:
for pc in result.data['RadiativeTransitions'][trans].ProcessClass:
if str(pc)[:3] == 'hyp':
t_hfs = str(pc)
except Exception, e:
print "Error: %s", e
t_name = "%s; %s; %s" % (formula, t_state, t_hfs)
t_name = t_name.strip()
# check if name is in the list of forbidden names and add counter if so
i = 1
while t_name in names_black_list:
t_name = "%s#%d" % (t_name.split('#')[0], i)
i += 1
# update list of distinct species names.
if id in species_names:
if not t_name in species_names[id]:
species_names[id].append(t_name)
num_transitions[t_name] = 0
else:
species_names[id] = [t_name]
num_transitions[t_name] = 0
frequency = float(result.data['RadiativeTransitions'][trans].FrequencyValue)
try:
uncertainty = "%lf" % float(result.data['RadiativeTransitions'][trans].FrequencyAccuracy)
except TypeError:
print " -- Error uncertainty not available"
species_with_error.append(id)
continue
# Get statistical weight if present
try:
weight = int(upper_state.TotalStatisticalWeight)
except:
print " -- Error statistical weight not available"
species_with_error.append(id)
continue
# Get nuclear spin isomer (ortho/para) if present
#print "%s; %s" % (result.data['RadiativeTransitions'][trans].Id, upper_state.Id)
try:
nsiName = upper_state.NuclearSpinIsomerName
except AttributeError:
nsiName = None
# Insert transition into database
try:
cursor.execute("""INSERT INTO Transitions (
T_Name,
T_Frequency,
T_EinsteinA,
T_Uncertainty,
T_EnergyLower,
T_UpperStateDegeneracy,
T_HFS,
T_UpperStateQuantumNumbers,
T_LowerStateQuantumNumbers) VALUES
(?, ?,?,?,?, ?,?, ?,?)""",
(t_name,
"%lf" % frequency,
"%g" % float(result.data['RadiativeTransitions'][trans].TransitionProbabilityA),
uncertainty, "%lf" % float(lower_state.StateEnergyValue),
weight,
#upper_state.QuantumNumbers.case,
t_hfs,
str(upper_state.QuantumNumbers.qn_string),
str(lower_state.QuantumNumbers.qn_string),
))
num_transitions[t_name] += 1
except Exception, e:
print "Transition has not been inserted:\n Error: %s" % e
def insert_radiativetransitions(self, species, node):
"""
"""
# will contain a list of names which belong to one specie
species_names = {}
#----------------------------------------------------------
# Create a list of species for which transitions will be
# retrieved and inserted in the database.
# Species have to be in the Partitionfunctions - table
if not functions.isiterable(species):
species = [species]
species_list = []
cursor = self.conn.cursor()
for specie in species:
cursor.execute("SELECT PF_Name, PF_SpeciesID, PF_VamdcSpeciesID, PF_HFS FROM Partitionfunctions WHERE PF_SpeciesID=? or PF_VamdcSpeciesID=?",
(specie, specie))
rows = cursor.fetchall()
for row in rows:
species_list.append([row[0], row[1], row[2], row[3]])
#--------------------------------------------------------------
for specie in species_list:
num_transitions = {}
#------------------------------------
# Retrieve data from the database
id = specie[1]
vamdcspeciesid = specie[2]
hfs = specie[3]
name = specie[0]
# name should be formated like 'formula; state-info; hfs-info'
name_array = name.split(';')
formula = name_array[0].strip()
try:
stateInfo = name_array[1].strip()
except:
stateInfo = ''
# get hfs-flag from the name.
try:
hfsInfo = name_array[2].strip()
except:
hfsInfo = ''
# Create query string
query_string = "SELECT ALL WHERE VAMDCSpeciesID='%s'" % vamdcspeciesid
if hfs is not None and hfs.strip() != '':
query_string += " and RadTransCode='%s'" % hfs
query = q.Query()
result = results.Result()
# Get data from the database
query.set_query(query_string)
query.set_node(node)
result.set_query(query)
result.do_query()
result.populate_model()
#---------------------------------------
cursor = self.conn.cursor()
cursor.execute('BEGIN TRANSACTION')
cursor.execute("DELETE FROM Transitions WHERE T_Name = ?", (name,))
for trans in result.data['RadiativeTransitions']:
# data might contain transitions for other species (if query is based on ichikey/vamdcspeciesid).
# Insert transitions only if they belong to the correct specie
if result.data['RadiativeTransitions'][trans].SpeciesID == id:
# Get upper and lower state from the states table
upper_state = result.data['States']["%s" % result.data['RadiativeTransitions'][trans].UpperStateRef]
lower_state = result.data['States']["%s" % result.data['RadiativeTransitions'][trans].LowerStateRef]
# Get string which identifies the vibrational states involved in the transition
try:
if upper_state.QuantumNumbers.vibstate == lower_state.QuantumNumbers.vibstate:
t_state = str(upper_state.QuantumNumbers.vibstate).strip()
else:
#vup = upper_state.QuantumNumbers.vibstate.split(",")
#vlow = lower_state.QuantumNumbers.vibstate.split(",")
v_dict = {}
for label in list(set(upper_state.QuantumNumbers.qn_dict.keys() + lower_state.QuantumNumbers.qn_dict.keys())):
if isVibrationalStateLabel(label):
try:
value_up = upper_state.QuantumNumbers.qn_dict[label]
except:
value_up = 0
try:
value_low = lower_state.QuantumNumbers.qn_dict[label]
except:
value_low = 0
v_dict[label] = [value_up, value_low]
v_string = ''
valup_string = ''
vallow_string = ''
for v in v_dict:
v_string += "%s," % v
valup_string += "%s," % v_dict[v][0]
vallow_string += "%s," % v_dict[v][1]
if len(v_dict) > 1:
t_state = "(%s)=(%s)-(%s)" % (v_string[:-1], valup_string[:-1], vallow_string[:-1])
else:
t_state = "%s=%s-%s" % (v_string[:-1], valup_string[:-1], vallow_string[:-1])
#t_state = '(%s)-(%s)' % (upper_state.QuantumNumbers.vibstate,lower_state.QuantumNumbers.vibstate)
except:
t_state = ''
# go to the next transition if state does not match
if t_state != stateInfo and stateInfo is not None and stateInfo != '':
continue
# Get hyperfinestructure info if hfsInfo is None
# only then the hfsInfo has not been inserted in the species name
# (there can be multiple values in the complete dataset
if hfsInfo == '':
t_hfs = ''
try:
for pc in result.data['RadiativeTransitions'][trans].ProcessClass:
if str(pc)[:3] == 'hyp':
t_hfs = str(pc)
except Exception, e:
print "Error: %s", e
else:
t_hfs = hfsInfo
# if hfs is not None and empty then only Transitions without hfs-flag
# should be processed
if hfs is not None and hfs != t_hfs:
continue
t_name = "%s; %s; %s" % (formula, t_state, t_hfs)
t_name = t_name.strip()
# update list of distinct species names.
if id in species_names:
if not t_name in species_names[id]:
species_names[id].append(t_name)
num_transitions[t_name] = 0
else:
species_names[id] = [t_name]
num_transitions[t_name] = 0
frequency = float(result.data['RadiativeTransitions'][trans].FrequencyValue)
uncertainty = "%lf" % float(result.data['RadiativeTransitions'][trans].FrequencyAccuracy)
# Get statistical weight if present
if upper_state.TotalStatisticalWeight:
weight = int(upper_state.TotalStatisticalWeight)
else:
weight = None
# Get nuclear spin isomer (ortho/para) if present
try:
nsiName = upper_state.NuclearSpinIsomerName
except AttributeError:
nsiName = None
# Insert transition into database
try:
cursor.execute("""INSERT INTO Transitions (
T_Name,
T_Frequency,
T_EinsteinA,
T_Uncertainty,
T_EnergyLower,
T_UpperStateDegeneracy,
T_HFS,
T_UpperStateQuantumNumbers,
T_LowerStateQuantumNumbers) VALUES
(?, ?,?,?,?, ?,?, ?,?)""",
(t_name,
"%lf" % frequency,
"%g" % float(result.data['RadiativeTransitions'][trans].TransitionProbabilityA),
uncertainty, "%lf" % float(lower_state.StateEnergyValue),
weight,
#upper_state.QuantumNumbers.case,
t_hfs,
str(upper_state.QuantumNumbers.qn_string),
str(lower_state.QuantumNumbers.qn_string),
))
num_transitions[t_name] += 1
except Exception, e:
print "Transition has not been inserted:\n Error: %s" % e
def _fuzzy_search_r(self, node, word, pos, result, result_list, k, alrdy_trans=False):
'''recursive function to do search work
explanation:
this performs a fuzzy search for correct words that match the given word input
:param node: current trienode
:param word: remainder of current word. ends in chr0
:param pos: position in trienode data
:param result: current built-up result (i.e. correct word)
:param result_list: pointer to list of results
:param k: allowed edits (total remaining)
:param alrdy_trans: only allow one transposition
'''
#pdb.set_trace()
if node.data[pos:] == word:
# we found a match! Also we must be on a leaf, so nowhere else to go
result.data += node.data[pos:-1]
result.metadata = node.metadata
result_list.add(result)
return
# need to add current char to result
result.data += node.data[pos]
# for edits that advance along the trie, two cases depending on
# if at end of current node or not:
if pos >= (len(node.data)-1): # if at end of node
for next_char, next_node in node.child_nodes.iteritems():
if k > 0:
# DELETION:
new_result = results.Result(result, 'deletion')
self._fuzzy_search_r(next_node, word, 0, new_result, result_list, k-1)
# SUBSTITUTION:
if len(word) > 0 and node.data[pos] != word[0]:
new_result = results.Result(result, 'substitution')
self._fuzzy_search_r(next_node, word[1:], 0, new_result, result_list, k-1)
# CORRECT:
if len(word) > 0 and node.data[pos] == word[0]:
new_result = results.Result(result, 'continue')
self._fuzzy_search_r(next_node, word[1:], 0, new_result, result_list, k)
else: # if not at end of node
if k > 0:
# DELETION:
new_result = results.Result(result, 'deletion')
self._fuzzy_search_r(node, word, pos+1, new_result, result_list, k-1)
# SUBSTITUTION:
if len(word) > 0 and node.data[pos] != word[0]:
new_result = results.Result(result, 'substitution')
self._fuzzy_search_r(node, word[1:], pos+1, new_result, result_list, k-1)
# CORRECT:
if len(word) > 0 and node.data[pos] == word[0]:
new_result = results.Result(result, 'continue')
self._fuzzy_search_r(node, word[1:], pos+1, new_result, result_list, k)
if k > 0:
# INSERTION - doesn't step forward so doesn't matter which if
# we are at the end of a node or not
new_result = results.Result(result, 'insertion')
new_result.data = new_result.data[:-1]
self._fuzzy_search_r(node, word[1:], pos, new_result, result_list, k-1)
# TRANSPOSITION - same:
if not alrdy_trans and len(word) >= 2:
#TODO: why only allow one?
new_word = word[1] + word[0] + word[2:]
new_result = results.Result(result, 'transposition')
new_result.data = new_result.data[:-1]
self._fuzzy_search_r(node, new_word, pos, new_result, result_list, k-1, alrdy_trans=True)
def test_single_result():
r = results.Result()
r.first_name = "Abc"
assert list(r.keys()) == ["first_name"]
assert r.first_name == "Abc"
assert hasattr(r, "first_name")
#Save result as a .json file
dataset = os.path.basename(
os.path.dirname(
os.path.dirname(
os.path.dirname(
os.path.dirname(os.path.dirname(args.tfrecords_dir))))))
round = 1
if args.round is not None:
round = args.round
if cfg.component_N == False:
n_traces = 1
else:
n_traces = 3
result = results.Result(cfg.window_size, n_traces, cfg.n_clusters,
cfg.model, truePositives, falsePositives,
trueNegatives, falseNegatives, accuracy, precision,
recall, f1, locationAccuracy, round, dataset)
now = datetime.datetime.now()
now_str = str(now.year) + str(now.month).zfill(2) + str(
now.day).zfill(2) + str(now.hour).zfill(2) + str(
now.minute).zfill(2) + str(now.second).zfill(2)
filename = "output/eval_" + now_str + "_" + str(
round) + "_" + dataset + "_" + str(cfg.window_size) + "_" + str(
cfg.n_clusters) + "_" + str(
cfg.n_traces) + "_" + cfg.model + ".json"
result.export_json(filename)
#if args.redirect_stdout_stderr:
# stdout_stderr_file.close()
def test_get_confusion_matrix(self):
cm = results.Result(*range(6), ytrue=[0, 1], ypred=[0, 0],
src={}).get_confusion_matrix()
self.assertTrue(np.all(cm == np.array([[1, 0], [1, 0]])))
