def show_loaddata_table(gate_names, as_columns=True):
'''Utility function to create a table that can be read by CP LoadData.
gate_names -- list of gate names to apply
as_columns -- use True to output each gate as a column with 0's and 1's
use False to output only the rows that fall within all gates.
'''
for g in gate_names:
for t in p.gates[g].get_tables():
assert t == p.image_table, 'this function only takes per-image gates'
columns = list(dbconnect.image_key_columns() + dbconnect.well_key_columns()) + p.image_file_cols + p.image_path_cols
if as_columns:
query_columns = columns + ['(%s) AS %s'%(str(p.gates[g]), g) for g in gate_names]
columns += gate_names
data = db.execute('SELECT %s FROM %s'
%(','.join(query_columns), p.image_table))
else:
# display only values within the given gates
where_clause = ' AND '.join([str(p.gates[g]) for g in gate_names])
data = db.execute('SELECT %s FROM %s WHERE %s'
%(','.join(columns), p.image_table, where_clause))
if data == []:
wx.MessageBox('Sorry, no data points fall within the combined selected gates.', 'No data to show')
return None
grid = TableViewer(None, title="Gated Data")
grid.table_from_array(np.array(data, dtype='object'), columns, grouping='image',
key_indices=range(len(dbconnect.image_key_columns())))
grid.Show()
return grid
def get_image_keys_at_row(self, row):
# XXX: needs to be updated to work for per_well data
if self.table_name == p.image_table:
key = self.get_row_key(row)
if key is None:
return None
return [key]
# return [tuple([self.GetValue(row, col) for col in self.key_indices])]
elif self.table_name == p.object_table:
key = self.get_row_key(row)
if key is None:
return None
return [key[:-1]]
else:
# BAD: assumes that columns with the same name as the image key
# columns ARE image key columns (not true if looking at unrelated
# image table)
key = []
for col in dbconnect.image_key_columns():
if col not in self.col_labels:
return None
else:
col_index = self.col_labels.tolist().index(col)
key += [self.GetValue(row, col_index)]
return [tuple(key)]
def get_image_keys_at_row(self, row):
# XXX: needs to be updated to work for per_well data
if self.table_name == p.image_table:
key = self.get_row_key(row)
if key is None:
return None
return [key]
# return [tuple([self.GetValue(row, col) for col in self.key_indices])]
elif self.table_name == p.object_table:
key = self.get_row_key(row)
if key is None:
return None
return [key[:-1]]
else:
# BAD: assumes that columns with the same name as the image key
# columns ARE image key columns (not true if looking at unrelated
# image table)
key = []
for col in dbconnect.image_key_columns():
if col not in self.col_labels:
return None
else:
col_index = self.col_labels.tolist().index(col)
key += [self.GetValue(row, col_index)]
return [tuple(key)]
def do_by_steps(tables, filter_name, area_score=False):
filter_clause = '1 = 1'
join_clause = ''
if filter_name is not None:
filter = p._filters[filter_name]
if isinstance(filter, cpa.sqltools.OldFilter):
join_table = '(%s) as filter' % str(filter)
else:
if p.object_table in tables:
join_table = None
else:
join_table = p.object_table
filter_clause = str(filter)
if join_table:
join_clause = 'JOIN %s USING (%s)' % (join_table, ','.join(image_key_columns()))
wheres = _where_clauses(p, dm, filter_name)
num_clauses = len(wheres)
counts = {}
# iterate over where clauses to go through whole set
for idx, where_clause in enumerate(wheres):
if filter_clause is not None:
where_clause += ' AND ' + filter_clause
if area_score:
data = db.execute('SELECT %s, %s, %s FROM %s '
'%s WHERE %s'
%(UniqueImageClause(p.object_table),
",".join(db.GetColnamesForClassifier()),
_objectify(p, p.area_scoring_column), tables,
join_clause, where_clause),
silent=(idx > 10))
area_score = data[-1] #separate area from data
data = data[:-1]
else:
data = db.execute('SELECT %s, %s FROM %s '
'%s WHERE %s'
%(UniqueImageClause(p.object_table),
",".join(db.GetColnamesForClassifier()), tables,
join_clause, where_clause),
silent=(idx > 10))
cell_data, image_keys = processData(data)
predicted_classes = classifier.Predict(cell_data)
for i in range(0, len(predicted_classes)):
row_cls = tuple(np.append(image_keys[i], predicted_classes[i]))
oneCount = np.array([1])
if area_score:
oneCount = np.append(oneCount, area_score[i])
if row_cls in counts:
counts[row_cls] += oneCount
else:
counts[row_cls] = oneCount
if cb:
cb(min(1, idx/float(num_clauses))) #progress
return counts
def FilterObjectsFromClassN(classNum, classifier, filterKeys, uncertain):
'''
uncertain: allows to search for uncertain (regarding the probs assigned by the classifier) cell images
classNum: 1-based index of the class to retrieve obKeys from
classifier: trained classifier object
filterKeys: (optional) A list of specific imKeys OR obKeys (NOT BOTH)
to classify.
* WARNING: If this list is too long, you may exceed the size limit to
MySQL queries.
* Useful when fetching N objects from a particular class. Use the
DataModel to get batches of random objects, and sift through them
here until N objects of the desired class have been accumulated.
* Also useful for classifying a specific image or group of images.
RETURNS: A list of object keys that fall in the specified class (but not all objects?),
if Properties.area_scoring_column is specified, area sums are also
reported for each class
'''
if filterKeys != [] and filterKeys is not None:
if isinstance(filterKeys, str):
whereclause = filterKeys #+ " AND"
else:
isImKey = len(filterKeys[0]) == len(image_key_columns())
if isImKey:
whereclause = GetWhereClauseForImages(filterKeys) #+ " AND"
else:
whereclause = GetWhereClauseForObjects(filterKeys) #+ " AND"
else:
whereclause = ""
if p.area_scoring_column:
data = db.execute('SELECT %s, %s FROM %s WHERE %s'%(UniqueObjectClause(p.object_table),
",".join(db.GetColnamesForClassifier()),
_objectify(p, p.area_scoring_column), p.object_table, whereclause))
area_score = data[-1] #separate area from data
data = data[:-1]
else:
data = db.execute('SELECT %s, %s FROM %s WHERE %s'%(UniqueObjectClause(p.object_table),
",".join(db.GetColnamesForClassifier()), p.object_table, whereclause))
cell_data, object_keys = processData(data)#, p.check_tables=='yes')
res = [] # list
if uncertain:
# Our requirement: if the two largest scores are smaller than threshold
probabilities = classifier.PredictProba(cell_data) #
threshold = 0.1 # TODO: This threshold should be adjustable
sorted_p = np.sort(probabilities)[:,-2:]# sorted array
diff = sorted_p[:,1] - sorted_p[:,0]
indices = np.where(diff < threshold)[0] # get all indices where this is true
res = [object_keys[i] for i in indices]
else:
predicted_classes = classifier.Predict(cell_data)
res = object_keys[predicted_classes == classNum * np.ones(predicted_classes.shape)].tolist() #convert to list
return map(tuple,res) # ... and then to tuples
def FilterObjectsFromClassN(classNum, classifier, filterKeys, uncertain):
'''
uncertain: allows to search for uncertain (regarding the probs assigned by the classifier) cell images
classNum: 1-based index of the class to retrieve obKeys from
classifier: trained classifier object
filterKeys: (optional) A list of specific imKeys OR obKeys (NOT BOTH)
to classify.
* WARNING: If this list is too long, you may exceed the size limit to
MySQL queries.
* Useful when fetching N objects from a particular class. Use the
DataModel to get batches of random objects, and sift through them
here until N objects of the desired class have been accumulated.
* Also useful for classifying a specific image or group of images.
RETURNS: A list of object keys that fall in the specified class (but not all objects?),
if Properties.area_scoring_column is specified, area sums are also
reported for each class
'''
if filterKeys != [] and filterKeys is not None:
if isinstance(filterKeys, str):
whereclause = filterKeys #+ " AND"
else:
isImKey = len(filterKeys[0]) == len(image_key_columns())
if isImKey:
whereclause = GetWhereClauseForImages(filterKeys) #+ " AND"
else:
whereclause = GetWhereClauseForObjects(filterKeys) #+ " AND"
else:
whereclause = ""
if p.area_scoring_column:
data = db.execute('SELECT %s, %s FROM %s WHERE %s'%(UniqueObjectClause(p.object_table),
",".join(db.GetColnamesForClassifier()),
_objectify(p, p.area_scoring_column), p.object_table, whereclause))
area_score = data[-1] #separate area from data
data = data[:-1]
else:
data = db.execute('SELECT %s, %s FROM %s WHERE %s'%(UniqueObjectClause(p.object_table),
",".join(db.GetColnamesForClassifier()), p.object_table, whereclause))
cell_data, object_keys = processData(data)
res = [] # list
if uncertain:
# Our requirement: if the two largest scores are smaller than threshold
probabilities = classifier.PredictProba(cell_data) #
threshold = 0.1 # TODO: This threshold should be adjustable
sorted_p = np.sort(probabilities)[:,-2:]# sorted array
diff = sorted_p[:,1] - sorted_p[:,0]
indices = np.where(diff < threshold)[0] # get all indices where this is true
res = [object_keys[i] for i in indices]
else:
predicted_classes = classifier.Predict(cell_data)
res = object_keys[predicted_classes == classNum * np.ones(predicted_classes.shape)].tolist() #convert to list
return map(tuple,res) # ... and then to tuples
def show_loaddata_table(gate_names, as_columns=True):
'''Utility function to create a table that can be read by CP LoadData.
gate_names -- list of gate names to apply
as_columns -- use True to output each gate as a column with 0's and 1's
use False to output only the rows that fall within all gates.
'''
for g in gate_names:
for t in p.gates[g].get_tables():
assert t == p.image_table, 'this function only takes per-image gates'
columns = list(
dbconnect.image_key_columns() +
dbconnect.well_key_columns()) + p.image_file_cols + p.image_path_cols
if as_columns:
query_columns = columns + [
'(%s) AS %s' % (str(p.gates[g]), g) for g in gate_names
]
columns += gate_names
data = db.execute('SELECT %s FROM %s' %
(','.join(query_columns), p.image_table))
else:
# display only values within the given gates
where_clause = ' AND '.join([str(p.gates[g]) for g in gate_names])
data = db.execute('SELECT %s FROM %s WHERE %s' %
(','.join(columns), p.image_table, where_clause))
if data == []:
wx.MessageBox(
'Sorry, no data points fall within the combined selected gates.',
'No data to show')
return None
grid = TableViewer(None, title="Gated Data")
grid.table_from_array(np.array(data, dtype='object'),
columns,
grouping='image',
key_indices=range(len(
dbconnect.image_key_columns())))
grid.Show()
return grid
def FilterObjectsFromClassN(classNum, classifier, filterKeys):
'''
classNum: 1-based index of the class to retrieve obKeys from
classifier: trained classifier object
filterKeys: (optional) A list of specific imKeys OR obKeys (NOT BOTH)
to classify.
* WARNING: If this list is too long, you may exceed the size limit to
MySQL queries.
* Useful when fetching N objects from a particular class. Use the
DataModel to get batches of random objects, and sift through them
here until N objects of the desired class have been accumulated.
* Also useful for classifying a specific image or group of images.
RETURNS: A list of object keys that fall in the specified class (but not all objects?),
if Properties.area_scoring_column is specified, area sums are also
reported for each class
'''
if filterKeys != [] and filterKeys is not None:
if isinstance(filterKeys, str):
whereclause = filterKeys #+ " AND"
else:
isImKey = len(filterKeys[0]) == len(image_key_columns())
if isImKey:
whereclause = GetWhereClauseForImages(filterKeys) #+ " AND"
else:
whereclause = GetWhereClauseForObjects(filterKeys) #+ " AND"
else:
whereclause = ""
if p.area_scoring_column:
data = db.execute('SELECT %s, %s FROM %s WHERE %s'%(UniqueObjectClause(p.object_table),
",".join(db.GetColnamesForClassifier()),
_objectify(p, p.area_scoring_column), p.object_table, whereclause))
area_score = data[-1] #separate area from data
data = data[:-1]
else:
data = db.execute('SELECT %s, %s FROM %s WHERE %s'%(UniqueObjectClause(p.object_table),
",".join(db.GetColnamesForClassifier()), p.object_table, whereclause))
number_of_features = len(db.GetColnamesForClassifier())
cell_data = np.array([row[-number_of_features:] for row in data]) #last number_of_features columns in row
object_keys = np.array([row[:-number_of_features] for row in data]) #all elements in row before last (number_of_features) elements
predicted_classes = classifier.Predict(cell_data)
res = object_keys[predicted_classes == classNum * np.ones(predicted_classes.shape)].tolist() #convert to list
return map(tuple,res) # ... and then to tuples
def set_table(self, table_name):
if table_name == p.image_table:
self.grouping = 'Image'
elif table_name == p.object_table:
self.grouping = 'Object'
else:
self.grouping = None
self.table_name = table_name
self.cache = odict()
self.col_labels = np.array(db.GetColumnNames(self.table_name))
self.shown_columns = np.arange(len(self.col_labels))
self.order_by = [self.col_labels[0]]
self.order_direction = 'ASC'
self.key_indices = None
if self.table_name == p.image_table:
self.key_indices = [self.col_labels.tolist().index(v) for v in dbconnect.image_key_columns()]
if self.table_name == p.object_table:
self.key_indices = [self.col_labels.tolist().index(v) for v in dbconnect.object_key_columns()]
def FilterObjectsFromClassN(self, classN = None, keys = None):
'''
Filter the input objects to output the keys of those in classN,
using a defined SVM model classifier.
'''
# Retrieve instance of the database connection
db = dbconnect.DBConnect.getInstance()
object_data = {}
if isinstance(keys, str):
object_data[0] = db.GetCellDataForClassifier(keys)
elif keys != []:
if len(keys) == len(dbconnect.image_key_columns()):
# Retrieve instance of the data model and retrieve objects in the requested image
dm = DataModel.getInstance()
obKeys = dm.GetObjectsFromImage(keys[0])
else:
obKeys = keys
for key in obKeys:
object_data[key] = db.GetCellDataForClassifier(key)
sorted_keys = sorted(object_data.keys())
values_array = np.array([object_data[key] for key in sorted_keys])
scaled_values = self.ScaleData(values_array)
pred_labels = self.model.predict(scaled_values)
# Group the object keys per class
classObjects = {}
for index in range(1, len(self.classBins)+1):
classObjects[float(index)] = []
for index, label in enumerate(pred_labels):
classObjects[np.int(label)+1].append(sorted_keys[index])
# Return either a summary of all classes and their corresponding objects
# or just the objects for a specific class
if classN is None:
return classObjects
else:
return classObjects[classN]
def OnSavePerImageCountsToCSV(self, evt):
defaultFileName = 'Per_Image_Counts.csv'
saveDialog = wx.FileDialog(self, message="Save as:",
defaultDir=os.getcwd(),
defaultFile=defaultFileName,
wildcard='csv|*',
style=(wx.SAVE | wx.FD_OVERWRITE_PROMPT |
wx.FD_CHANGE_DIR))
if saveDialog.ShowModal()==wx.ID_OK:
colHeaders = list(dbconnect.image_key_columns())
pos = len(colHeaders)
if p.plate_id:
colHeaders += [p.plate_id]
if p.well_id:
colHeaders += [p.well_id]
colHeaders += ['total_count']
colHeaders += ['count_'+bin.label for bin in self.GetParent().classBins]
data = list(self.GetParent().keysAndCounts)
for row in data:
if p.table_id:
where = '%s=%s AND %s=%s'%(p.table_id, row[0], p.image_id, row[1])
total = sum(row[2:])
else:
where = '%s=%s'%(p.image_id, row[0])
total = sum(row[1:])
row.insert(pos, total)
# Plate and Well are written separately IF they are found in the props file
# TODO: ANY column could be reported by this mechanism
if p.well_id:
res = db.execute('SELECT %s FROM %s WHERE %s'%(p.well_id, p.image_table, where), silent=True)
well = res[0][0]
row.insert(pos, well)
if p.plate_id:
res = db.execute('SELECT %s FROM %s WHERE %s'%(p.plate_id, p.image_table, where), silent=True)
plate = res[0][0]
row.insert(pos, plate)
self.SaveCSV(saveDialog.GetPath(), data, colHeaders)
saveDialog.Destroy()
def FilterObjectsFromClassN(self, classN=None, keys=None):
'''
Filter the input objects to output the keys of those in classN,
using a defined SVM model classifier.
'''
# Retrieve instance of the database connection
db = dbconnect.DBConnect.getInstance()
object_data = {}
if isinstance(keys, str):
object_data[0] = db.GetCellDataForClassifier(keys)
elif keys != []:
if len(keys) == len(dbconnect.image_key_columns()):
# Retrieve instance of the data model and retrieve objects in the requested image
dm = DataModel.getInstance()
obKeys = dm.GetObjectsFromImage(keys[0])
else:
obKeys = keys
for key in obKeys:
object_data[key] = db.GetCellDataForClassifier(key)
sorted_keys = sorted(object_data.keys())
values_array = np.array([object_data[key] for key in sorted_keys])
scaled_values = self.ScaleData(values_array)
pred_labels = self.model.predict(scaled_values)
# Group the object keys per class
classObjects = {}
for index in range(1, len(self.classBins) + 1):
classObjects[float(index)] = []
for index, label in enumerate(pred_labels):
classObjects[np.int(label) + 1].append(sorted_keys[index])
# Return either a summary of all classes and their corresponding objects
# or just the objects for a specific class
if classN is None:
return classObjects
else:
return classObjects[classN]
def set_table(self, table_name):
if table_name == p.image_table:
self.grouping = 'Image'
elif table_name == p.object_table:
self.grouping = 'Object'
else:
self.grouping = None
self.table_name = table_name
self.cache = odict()
self.col_labels = np.array(db.GetColumnNames(self.table_name))
self.shown_columns = np.arange(len(self.col_labels))
self.order_by = [self.col_labels[0]]
self.order_direction = 'ASC'
self.key_indices = None
if self.table_name == p.image_table:
self.key_indices = [
self.col_labels.tolist().index(v)
for v in dbconnect.image_key_columns()
]
if self.table_name == p.object_table:
self.key_indices = [
self.col_labels.tolist().index(v)
for v in dbconnect.object_key_columns()
]
def LoadCSV(self, csvfile, group='Image'):
try:
self.grid.Destroy()
except: pass
try:
# Remove the previous column show/hide menu (should be the third menu)
self.GetMenuBar().Remove(2)
self.colmenu.Destroy()
except: pass
r = csv.reader(open(csvfile))
labels = r.next()
dtable = dbconnect.get_data_table_from_csv_reader(r)
coltypes = db.InferColTypesFromData(dtable, len(labels))
for i in range(len(coltypes)):
if coltypes[i] == 'INT': coltypes[i] = int
elif coltypes[i] == 'FLOAT': coltypes[i] = float
else: coltypes[i] = str
r = csv.reader(open(csvfile))
r.next() # skip col-headers
data = []
for row in r:
data += [[coltypes[i](v) for i,v in enumerate(row)]]
data = np.array(data, dtype=object)
if group == DO_NOT_LINK_TO_IMAGES:
keycols = []
elif group == 'Image':
keycols = range(len(dbconnect.image_key_columns()))
else:
keycols = range(len(dm.GetGroupColumnNames(group)))
self.grid = HugeTableGrid(self, data, labels, key_col_indices=keycols, grouping=group, chMap=p.image_channel_colors)
self.Title = '%s (%s)'%(csvfile, group)
self.file = csvfile
self.CreateColumnMenu()
self.RescaleGrid()
def do_normalization(self):
if not self.validate():
# Should be unreachable
wx.MessageBox('Your normalization settings are invalid. Can\'t perform normalization.')
long_cols = [col for col in self.col_choices.GetCheckedStrings()
if len(col) + 4 > 64]
if long_cols:
dlg = wx.MessageDialog(self, 'The following columns contain more '
'than 64 characters when a normalization suffix (4 '
'characters) is appended. This may cause a problem when '
'writing to the database.\n %s'%('\n'.join(long_cols)),
'Warning', wx.OK|wx.CANCEL|wx.ICON_EXCLAMATION)
if dlg.ShowModal() == wx.ID_CANCEL:
return
dlg.Destroy()
imkey_cols = dbconnect.image_key_columns()
obkey_cols = dbconnect.object_key_columns()
wellkey_cols = dbconnect.well_key_columns()
im_clause = dbconnect.UniqueImageClause
well_clause = dbconnect.UniqueWellClause
input_table = self.table_choice.GetStringSelection()
meas_cols = self.col_choices.GetCheckedStrings()
wants_norm_meas = self.norm_meas_checkbox.IsChecked()
wants_norm_factor = self.norm_factor_checkbox.IsChecked()
output_table = self.output_table.Value
FIRST_MEAS_INDEX = len(imkey_cols + (wellkey_cols or tuple()))
if p.db_type == 'mysql':
BATCH_SIZE = 100
else:
BATCH_SIZE = 1
if input_table == p.object_table:
FIRST_MEAS_INDEX += 1 # Original
if wellkey_cols:
if input_table == p.image_table:
WELL_KEY_INDEX = len(imkey_cols)
else:
WELL_KEY_INDEX = len(imkey_cols) + 1
if db.table_exists(output_table):
dlg = wx.MessageDialog(self, 'Are you sure you want to overwrite the table "%s"?'%(output_table),
"Overwrite table?", wx.YES_NO|wx.NO_DEFAULT|wx.ICON_EXCLAMATION)
if dlg.ShowModal() == wx.ID_NO:
dlg.Destroy()
return
dlg.Destroy()
#
# First Get the data from the db.
#
if input_table == p.image_table:
if wellkey_cols:
# If there are well columns, fetch them.
query = "SELECT %s, %s, %s FROM %s"%(
im_clause(), well_clause(), ', '.join(meas_cols),
input_table)
else:
query = "SELECT %s, %s FROM %s"%(
im_clause(), ', '.join(meas_cols),
input_table)
elif input_table == p.object_table:
if p.image_table and wellkey_cols:
# If we have x and y from cells, we can use that for classifier
if p.cell_x_loc and p.cell_y_loc:
FIRST_MEAS_INDEX += 2 # Cell X and Y Location are fixed to for classifier
# If there are well columns, fetch them from the per-image table.
query = "SELECT %s, %s, %s, %s, %s FROM %s, %s WHERE %s"%(
dbconnect.UniqueObjectClause(p.object_table),
well_clause(p.image_table),
p.cell_x_loc,
p.cell_y_loc,
', '.join(['%s.%s'%(p.object_table, col) for col in meas_cols]),
p.image_table, p.object_table,
' AND '.join(['%s.%s=%s.%s'%(p.image_table, c, p.object_table, c)
for c in imkey_cols]) )
else:
# If there are well columns, fetch them from the per-image table.
query = "SELECT %s, %s, %s FROM %s, %s WHERE %s"%(
dbconnect.UniqueObjectClause(p.object_table),
well_clause(p.image_table),
', '.join(['%s.%s'%(p.object_table, col) for col in meas_cols]),
p.image_table, p.object_table,
' AND '.join(['%s.%s=%s.%s'%(p.image_table, c, p.object_table, c)
for c in imkey_cols]) )
else:
if p.cell_x_loc and p.cell_y_loc:
FIRST_MEAS_INDEX += 2 # Cell X and Y Location are fixed to for classifier
query = "SELECT %s, %s, %s, %s FROM %s"%(
im_clause(), p.cell_x_loc, p.cell_y_loc, ', '.join(meas_cols),
input_table)
else:
query = "SELECT %s, %s FROM %s"%(
im_clause(), ', '.join(meas_cols),
input_table)
if p.negative_control: # if the user defined negative control, we can use that to fetch the wellkeys
neg_query = query + ' AND ' + p.negative_control # fetch all the negative control elements
if wellkey_cols:
query += " ORDER BY %s"%(well_clause(p.image_table))
dlg = wx.ProgressDialog('Computing normalized values',
'Querying database for raw data.',
parent=self,
style = wx.PD_CAN_ABORT|wx.PD_APP_MODAL)
dlg.Pulse()
#
# MAKE THE QUERY
#
input_data = np.array(db.execute(query), dtype=object)
if p.negative_control:
import pandas as pd
negative_control = pd.DataFrame(db.execute(neg_query), dtype=float)
logging.info("# of objects in negative control: " + str(negative_control.shape[0]))
logging.info("# of objects queried: " + str(input_data.shape[0]))
neg_mean_plate = negative_control.groupby([WELL_KEY_INDEX]).mean()
neg_std_plate = negative_control.groupby([WELL_KEY_INDEX]).std()
output_columns = np.ones(input_data[:,FIRST_MEAS_INDEX:].shape) * np.nan
output_factors = np.ones(input_data[:,FIRST_MEAS_INDEX:].shape) * np.nan
for colnum, col in enumerate(input_data[:,FIRST_MEAS_INDEX:].T):
keep_going, skip = dlg.Pulse("Normalizing column %d of %d"%(colnum+1, len(meas_cols)))
if not keep_going:
dlg.Destroy()
return
norm_data = col.copy()
for step_num, step_panel in enumerate(self.norm_steps):
d = step_panel.get_configuration_dict()
if d[norm.P_GROUPING] in (norm.G_QUADRANT, norm.G_WELL_NEIGHBORS):
# Reshape data if normalization step is plate sensitive.
assert p.plate_id and p.well_id
well_keys = input_data[:, range(WELL_KEY_INDEX, FIRST_MEAS_INDEX - 2) ]
wellkeys_and_vals = np.hstack((well_keys, np.array([norm_data]).T))
new_norm_data = []
for plate, plate_grp in groupby(wellkeys_and_vals, lambda(row): row[0]):
keys_and_vals = np.array(list(plate_grp))
plate_data, wks, ind = FormatPlateMapData(keys_and_vals)
pnorm_data = norm.do_normalization_step(plate_data, **d)
new_norm_data += pnorm_data.flatten()[ind.flatten().tolist()].tolist()
norm_data = new_norm_data
elif d[norm.P_GROUPING] == norm.G_PLATE:
assert p.plate_id and p.well_id
if d[norm.P_AGG_TYPE] == norm.M_NEGCTRL:
mean_plate_col = neg_mean_plate[colnum + FIRST_MEAS_INDEX]
std_plate_col = neg_std_plate[colnum + FIRST_MEAS_INDEX]
print mean_plate_col
print std_plate_col
well_keys = input_data[:, range(WELL_KEY_INDEX, FIRST_MEAS_INDEX - 2)]
wellkeys_and_vals = np.hstack((well_keys, np.array([norm_data]).T))
new_norm_data = []
# print wellkeys_and_vals
for plate, plate_grp in groupby(wellkeys_and_vals, lambda(row): row[0]):
plate_data = np.array(list(plate_grp))[:,-1].flatten()
pnorm_data = norm.do_normalization_step(plate_data, **d)
if d[norm.P_AGG_TYPE] == norm.M_NEGCTRL:
try:
plate_mean = mean_plate_col[plate]
plate_std = std_plate_col[plate]
except:
plate_mean = mean_plate_col[int(plate)]
plate_std = std_plate_col[int(plate)]
try:
pnorm_data = (pnorm_data - plate_mean) / plate_std
print pnorm_data
except:
logging.error("Plate std is zero, division by zero!")
new_norm_data += pnorm_data.tolist()
norm_data = new_norm_data
else:
norm_data = norm.do_normalization_step(norm_data, **d)
output_columns[:, colnum] = np.array(norm_data)
output_factors[:,
colnum] = col.astype(float) / np.array(norm_data,
dtype=float)
dlg.Destroy()
return # Abort here for coding
norm_table_cols = []
# Write new table
db.execute('DROP TABLE IF EXISTS %s' % (output_table))
if input_table == p.image_table:
norm_table_cols += dbconnect.image_key_columns()
col_defs = ', '.join([
'%s %s' % (col, db.GetColumnTypeString(p.image_table, col))
for col in dbconnect.image_key_columns()
])
elif input_table == p.object_table:
norm_table_cols += obkey_cols
col_defs = ', '.join([
'%s %s' % (col, db.GetColumnTypeString(p.object_table, col))
for col in obkey_cols
])
if wellkey_cols:
norm_table_cols += wellkey_cols
col_defs += ', ' + ', '.join([
'%s %s' % (col, db.GetColumnTypeString(p.image_table, col))
for col in wellkey_cols
def prompt_user_to_link_table(parent, table):
'''Prompts the user for information about the given table so it may be
linked into the tables that CPA already accesses.
returns the given table name or None if the user cancels
'''
dlg = wx.SingleChoiceDialog(parent, 'What kind of data is in this table (%s)?'%(table),
'Select table type', ['per-well', 'per-image', 'per-object', 'other'],
wx.CHOICEDLG_STYLE)
show_table_button = wx.Button(dlg, -1, 'Show table')
dlg.Sizer.Children[2].GetSizer().Insert(0, show_table_button, 0, wx.ALL, 10)
dlg.Sizer.Children[2].GetSizer().InsertStretchSpacer(1, 1)
def on_show_table(evt):
from tableviewer import TableViewer
tableview = TableViewer(get_main_frame_or_none())
tableview.Show()
tableview.load_db_table(table)
show_table_button.Bind(wx.EVT_BUTTON, on_show_table)
if dlg.ShowModal() != wx.ID_OK:
dlg.Destroy()
return None
new_table_type = dlg.GetStringSelection()
if new_table_type == 'per-well':
link_table_to_try = p.image_table
link_cols_to_try = dbconnect.well_key_columns()
elif new_table_type == 'per-image':
dlg = wx.MessageDialog(parent, 'Does this per-image table represent a '
'new set of images in your experiment?',
'New per-image table', wx.YES_NO)
if dlg.ShowModal() == wx.ID_YES:
wx.MessageDialog('Sorry, CPA does not currently support multiple\n'
'per-image tables unless they are referring to the\n'
'same images.\n\n'
'Please see the manual for more information',
'Multiple per-image tables not supported')
dlg.Destroy()
return None
link_table_to_try = p.image_table
link_cols_to_try = dbconnect.image_key_columns()
elif new_table_type == 'per-object':
dlg = wx.MessageDialog(parent, 'Does this per-object table represent a '
'new set of objects in your experiment?',
'New per-object table', wx.YES_NO)
if dlg.ShowModal() == wx.ID_YES:
wx.MessageDialog('Sorry, CPA does not currently support multiple\n'
'per-object tables unless they are referring to the\n'
'same objects.\n\n'
'Please see the manual for more information',
'Multiple per-object tables not supported')
if p.object_table:
if table == p.object_table:
raise
link_table_to_try = p.object_table
link_cols_to_try = dbconnect.object_key_columns()
else:
# There should never be an object table without another object
# table existing first. Connecting this table to the image_table is
# asking for trouble.
return None
else:
dlg = wx.SingleChoiceDialog(parent, 'Which of your tables is "%s" linked '
'to?'%(table), 'Select linking table',
db.get_linkable_tables(), wx.CHOICEDLG_STYLE)
if dlg.ShowModal() != wx.ID_OK:
dlg.Destroy()
return None
link_table_to_try = dlg.GetStringSelection()
link_cols_to_try = []
dlg = LinkTablesDialog(parent, table, link_table_to_try,
link_cols_to_try, link_cols_to_try)
if dlg.ShowModal() != wx.ID_OK:
dlg.Destroy()
return None
col_pairs = dlg.get_column_pairs()
src_cols = [col_pair[0][1] for col_pair in col_pairs]
dest_cols = [col_pair[1][1] for col_pair in col_pairs]
db.do_link_tables(table, link_table_to_try, src_cols, dest_cols)
# return the newly linked table
return table
coltypes = db.InferColTypesFromData(dtable, len(labels))
for i in range(len(coltypes)):
if coltypes[i] == 'INT': coltypes[i] = int
elif coltypes[i] == 'FLOAT': coltypes[i] = float
else: coltypes[i] = str
r = csv.reader(open(csvfile))
r.next() # skip col-headers
data = []
for row in r:
data += [[coltypes[i](v) for i,v in enumerate(row)]]
data = np.array(data, dtype=object)
group = 'Image'
if len(sys.argv)==4:
group = sys.argv[3]
if group == 'Image':
keycols = range(len(dbconnect.image_key_columns()))
else:
keycols = range(len(dm.GetGroupColumnNames(group)))
grid = DataGrid(data, labels, grouping=group,
key_col_indices=keycols,
chMap=p.image_channel_colors,
title=csvfile, autosave=False)
grid.Show()
app.MainLoop()
def do_normalization(self):
if not self.validate():
# Should be unreachable
wx.MessageBox(
'Your normalization settings are invalid. Can\'t perform normalization.'
)
long_cols = [
col for col in self.col_choices.GetCheckedStrings()
if len(col) + 4 > 64
]
if long_cols:
dlg = wx.MessageDialog(
self, 'The following columns contain more '
'than 64 characters when a normalization suffix (4 '
'characters) is appended. This may cause a problem when '
'writing to the database.\n %s' % ('\n'.join(long_cols)),
'Warning', wx.OK | wx.CANCEL | wx.ICON_EXCLAMATION)
if dlg.ShowModal() == wx.ID_CANCEL:
return
dlg.Destroy()
imkey_cols = dbconnect.image_key_columns()
obkey_cols = dbconnect.object_key_columns()
wellkey_cols = dbconnect.well_key_columns()
im_clause = dbconnect.UniqueImageClause
well_clause = dbconnect.UniqueWellClause
input_table = self.table_choice.GetStringSelection()
meas_cols = self.col_choices.GetCheckedStrings()
wants_norm_meas = self.norm_meas_checkbox.IsChecked()
wants_norm_factor = self.norm_factor_checkbox.IsChecked()
output_table = self.output_table.Value
FIRST_MEAS_INDEX = len(imkey_cols + (wellkey_cols or tuple()))
if p.db_type == 'mysql':
BATCH_SIZE = 100
else:
BATCH_SIZE = 1
if input_table == p.object_table:
FIRST_MEAS_INDEX += 1 # Original
if wellkey_cols:
if input_table == p.image_table:
WELL_KEY_INDEX = len(imkey_cols)
else:
WELL_KEY_INDEX = len(imkey_cols) + 1
if db.table_exists(output_table):
dlg = wx.MessageDialog(
self, 'Are you sure you want to overwrite the table "%s"?' %
(output_table), "Overwrite table?",
wx.YES_NO | wx.NO_DEFAULT | wx.ICON_EXCLAMATION)
if dlg.ShowModal() == wx.ID_NO:
dlg.Destroy()
return
dlg.Destroy()
#
# First Get the data from the db.
#
if input_table == p.image_table:
if wellkey_cols:
# If there are well columns, fetch them.
query = "SELECT %s, %s, %s FROM %s" % (im_clause(
), well_clause(), ', '.join(meas_cols), input_table)
else:
query = "SELECT %s, %s FROM %s" % (
im_clause(), ', '.join(meas_cols), input_table)
elif input_table == p.object_table:
if p.image_table and wellkey_cols:
# If we have x and y from cells, we can use that for classifier
if p.cell_x_loc and p.cell_y_loc:
FIRST_MEAS_INDEX += 2 # Cell X and Y Location are fixed to for classifier
# If there are well columns, fetch them from the per-image table.
query = "SELECT %s, %s, %s, %s, %s FROM %s, %s WHERE %s" % (
dbconnect.UniqueObjectClause(
p.object_table), well_clause(p.image_table),
p.cell_x_loc, p.cell_y_loc, ', '.join([
'%s.%s' % (p.object_table, col)
for col in meas_cols
]), p.image_table, p.object_table, ' AND '.join([
'%s.%s=%s.%s' %
(p.image_table, c, p.object_table, c)
for c in imkey_cols
]))
else:
# If there are well columns, fetch them from the per-image table.
query = "SELECT %s, %s, %s FROM %s, %s WHERE %s" % (
dbconnect.UniqueObjectClause(p.object_table),
well_clause(p.image_table), ', '.join([
'%s.%s' % (p.object_table, col)
for col in meas_cols
]), p.image_table, p.object_table, ' AND '.join([
'%s.%s=%s.%s' %
(p.image_table, c, p.object_table, c)
for c in imkey_cols
]))
else:
if p.cell_x_loc and p.cell_y_loc:
FIRST_MEAS_INDEX += 2 # Cell X and Y Location are fixed to for classifier
query = "SELECT %s, %s, %s, %s FROM %s" % (
im_clause(), p.cell_x_loc, p.cell_y_loc,
', '.join(meas_cols), input_table)
else:
query = "SELECT %s, %s FROM %s" % (
im_clause(), ', '.join(meas_cols), input_table)
if p.negative_control: # if the user defined negative control, we can use that to fetch the wellkeys
neg_query = query + ' AND ' + p.negative_control # fetch all the negative control elements
if wellkey_cols:
query += " ORDER BY %s" % (well_clause(p.image_table))
dlg = wx.ProgressDialog('Computing normalized values',
'Querying database for raw data.',
parent=self,
style=wx.PD_CAN_ABORT | wx.PD_APP_MODAL)
dlg.Pulse()
#
# MAKE THE QUERY
#
input_data = np.array(db.execute(query), dtype=object)
if p.negative_control:
import pandas as pd
negative_control = pd.DataFrame(db.execute(neg_query), dtype=float)
logging.info("# of objects in negative control: " +
str(negative_control.shape[0]))
logging.info("# of objects queried: " + str(input_data.shape[0]))
neg_mean_plate = negative_control.groupby([WELL_KEY_INDEX]).mean()
neg_std_plate = negative_control.groupby([WELL_KEY_INDEX]).std()
output_columns = np.ones(input_data[:,
FIRST_MEAS_INDEX:].shape) * np.nan
output_factors = np.ones(input_data[:,
FIRST_MEAS_INDEX:].shape) * np.nan
for colnum, col in enumerate(input_data[:, FIRST_MEAS_INDEX:].T):
keep_going, skip = dlg.Pulse("Normalizing column %d of %d" %
(colnum + 1, len(meas_cols)))
if not keep_going:
dlg.Destroy()
return
norm_data = col.copy()
for step_num, step_panel in enumerate(self.norm_steps):
d = step_panel.get_configuration_dict()
if d[norm.P_GROUPING] in (norm.G_QUADRANT,
norm.G_WELL_NEIGHBORS):
# Reshape data if normalization step is plate sensitive.
assert p.plate_id and p.well_id
well_keys = input_data[:,
range(WELL_KEY_INDEX,
FIRST_MEAS_INDEX - 2)]
wellkeys_and_vals = np.hstack(
(well_keys, np.array([norm_data]).T))
new_norm_data = []
for plate, plate_grp in groupby(wellkeys_and_vals, lambda
(row): row[0]):
keys_and_vals = np.array(list(plate_grp))
plate_data, wks, ind = FormatPlateMapData(
keys_and_vals)
pnorm_data = norm.do_normalization_step(
plate_data, **d)
new_norm_data += pnorm_data.flatten()[
ind.flatten().tolist()].tolist()
norm_data = new_norm_data
elif d[norm.P_GROUPING] == norm.G_PLATE:
assert p.plate_id and p.well_id
if d[norm.P_AGG_TYPE] == norm.M_NEGCTRL:
mean_plate_col = neg_mean_plate[colnum +
FIRST_MEAS_INDEX]
std_plate_col = neg_std_plate[colnum +
FIRST_MEAS_INDEX]
print mean_plate_col
print std_plate_col
well_keys = input_data[:,
range(WELL_KEY_INDEX,
FIRST_MEAS_INDEX - 2)]
wellkeys_and_vals = np.hstack(
(well_keys, np.array([norm_data]).T))
new_norm_data = []
# print wellkeys_and_vals
for plate, plate_grp in groupby(wellkeys_and_vals, lambda
(row): row[0]):
plate_data = np.array(list(plate_grp))[:, -1].flatten()
pnorm_data = norm.do_normalization_step(
plate_data, **d)
if d[norm.P_AGG_TYPE] == norm.M_NEGCTRL:
try:
plate_mean = mean_plate_col[plate]
plate_std = std_plate_col[plate]
except:
plate_mean = mean_plate_col[int(plate)]
plate_std = std_plate_col[int(plate)]
try:
pnorm_data = (pnorm_data -
plate_mean) / plate_std
print pnorm_data
except:
logging.error(
"Plate std is zero, division by zero!")
new_norm_data += pnorm_data.tolist()
norm_data = new_norm_data
def do_by_steps(tables, filter_name, area_score=False):
filter_clause = '1 = 1'
join_clause = ''
if filter_name is not None:
filter = p._filters[filter_name]
if isinstance(filter, cpa.sqltools.OldFilter):
join_table = '(%s) as filter' % str(filter)
else:
if p.object_table in tables:
join_table = None
else:
join_table = p.object_table
filter_clause = str(filter)
if join_table:
join_clause = 'JOIN %s USING (%s)' % (join_table, ','.join(
image_key_columns()))
wheres = _where_clauses(p, dm, filter_name)
num_clauses = len(wheres)
counts = {}
# iterate over where clauses to go through whole set
for idx, where_clause in enumerate(wheres):
if filter_clause is not None:
where_clause += ' AND ' + filter_clause
if area_score:
data = db.execute('SELECT %s, %s, %s FROM %s '
'%s WHERE %s' %
(UniqueImageClause(p.object_table), ",".join(
db.GetColnamesForClassifier()),
_objectify(p, p.area_scoring_column),
tables, join_clause, where_clause),
silent=(idx > 10))
area_score = data[-1] #separate area from data
data = data[:-1]
else:
data = db.execute('SELECT %s, %s FROM %s '
'%s WHERE %s' %
(UniqueObjectClause(p.object_table),
",".join(db.GetColnamesForClassifier()),
tables, join_clause, where_clause),
silent=(idx > 10))
cell_data, image_keys = processData(data)
for i in range(cell_data.shape[0]):
for j in range(cell_data.shape[1]):
try:
float(cell_data[i, j])
except:
print(i, j, cell_data[i, j], type(cell_data[i, j]))
predicted_classes = classifier.Predict(cell_data)
for i in range(0, len(predicted_classes)):
row_cls = tuple(
np.append(image_keys[i][0], predicted_classes[i]))
oneCount = np.array([1])
if area_score:
oneCount = np.append(oneCount, area_score[i])
if row_cls in counts:
counts[row_cls] += oneCount
else:
counts[row_cls] = oneCount
if cb:
cb(min(1, idx / float(num_clauses))) #progress
return counts
def prompt_user_to_link_table(parent, table):
'''Prompts the user for information about the given table so it may be
linked into the tables that CPA already accesses.
returns the given table name or None if the user cancels
'''
dlg = wx.SingleChoiceDialog(parent, 'What kind of data is in this table (%s)?'%(table),
'Select table type', ['per-well', 'per-image', 'per-object', 'other'],
wx.CHOICEDLG_STYLE)
show_table_button = wx.Button(dlg, -1, 'Show table')
dlg.Sizer.Children[2].GetSizer().Insert(0, show_table_button, 0, wx.ALL, 10)
dlg.Sizer.Children[2].GetSizer().InsertStretchSpacer(1, 1)
def on_show_table(evt):
from tableviewer import TableViewer
tableview = TableViewer(get_main_frame_or_none())
tableview.Show()
tableview.load_db_table(table)
show_table_button.Bind(wx.EVT_BUTTON, on_show_table)
if dlg.ShowModal() != wx.ID_OK:
dlg.Destroy()
return None
new_table_type = dlg.GetStringSelection()
if new_table_type == 'per-well':
link_table_to_try = p.image_table
link_cols_to_try = dbconnect.well_key_columns()
elif new_table_type == 'per-image':
dlg = wx.MessageDialog(parent, 'Does this per-image table represent a '
'new set of images in your experiment?',
'New per-image table', wx.YES_NO)
if dlg.ShowModal() == wx.ID_YES:
wx.MessageDialog(parent,'Sorry, CPA does not currently support multiple\n'
'per-image tables unless they are referring to the\n'
'same images.\n\n'
'Please see the manual for more information',
'Multiple per-image tables not supported')
dlg.Destroy()
return None
link_table_to_try = p.image_table
link_cols_to_try = dbconnect.image_key_columns()
elif new_table_type == 'per-object':
dlg = wx.MessageDialog(parent, 'Does this per-object table represent a '
'new set of objects in your experiment?',
'New per-object table', wx.YES_NO)
if dlg.ShowModal() == wx.ID_YES:
wx.MessageDialog(parent,'Sorry, CPA does not currently support multiple\n'
'per-object tables unless they are referring to the\n'
'same objects.\n\n'
'Please see the manual for more information',
'Multiple per-object tables not supported')
if p.object_table:
if table == p.object_table:
raise
link_table_to_try = p.object_table
link_cols_to_try = dbconnect.object_key_columns()
else:
# There should never be an object table without another object
# table existing first. Connecting this table to the image_table is
# asking for trouble.
return None
else:
dlg = wx.SingleChoiceDialog(parent, 'Which of your tables is "%s" linked '
'to?'%(table), 'Select linking table',
db.get_linkable_tables(), wx.CHOICEDLG_STYLE)
if dlg.ShowModal() != wx.ID_OK:
dlg.Destroy()
return None
link_table_to_try = dlg.GetStringSelection()
link_cols_to_try = []
dlg = LinkTablesDialog(parent, table, link_table_to_try,
link_cols_to_try, link_cols_to_try)
if dlg.ShowModal() != wx.ID_OK:
dlg.Destroy()
return None
col_pairs = dlg.get_column_pairs()
src_cols = [col_pair[0][1] for col_pair in col_pairs]
dest_cols = [col_pair[1][1] for col_pair in col_pairs]
db.do_link_tables(table, link_table_to_try, src_cols, dest_cols)
# return the newly linked table
return table
def do_normalization(self):
if not self.validate():
# Should be unreachable
wx.MessageBox('Your normalization settings are invalid. Can\'t perform normalization.')
long_cols = [col for col in self.col_choices.GetCheckedStrings()
if len(col) + 4 > 64]
if long_cols:
dlg = wx.MessageDialog(self, 'The following columns contain more '
'than 64 characters when a normalization suffix (4 '
'characters) is appended. This may cause a problem when '
'writing to the database.\n %s'%('\n'.join(long_cols)),
'Warning', wx.OK|wx.CANCEL|wx.ICON_EXCLAMATION)
if dlg.ShowModal() == wx.ID_CANCEL:
return
dlg.Destroy()
imkey_cols = dbconnect.image_key_columns()
obkey_cols = dbconnect.object_key_columns()
wellkey_cols = dbconnect.well_key_columns()
im_clause = dbconnect.UniqueImageClause
well_clause = dbconnect.UniqueWellClause
input_table = self.table_choice.GetStringSelection()
meas_cols = self.col_choices.GetCheckedStrings()
wants_norm_meas = self.norm_meas_checkbox.IsChecked()
wants_norm_factor = self.norm_factor_checkbox.IsChecked()
output_table = self.output_table.Value
FIRST_MEAS_INDEX = len(imkey_cols + (wellkey_cols or tuple()))
if p.db_type == 'mysql':
BATCH_SIZE = 100
else:
BATCH_SIZE = 1
if input_table == p.object_table:
FIRST_MEAS_INDEX += 1 # Original
if wellkey_cols:
if input_table == p.image_table:
WELL_KEY_INDEX = len(imkey_cols)
else:
WELL_KEY_INDEX = len(imkey_cols) + 1
if db.table_exists(output_table):
dlg = wx.MessageDialog(self, 'Are you sure you want to overwrite the table "%s"?'%(output_table),
"Overwrite table?", wx.YES_NO|wx.NO_DEFAULT|wx.ICON_EXCLAMATION)
if dlg.ShowModal() == wx.ID_NO:
dlg.Destroy()
return
dlg.Destroy()
#
# First Get the data from the db.
#
if input_table == p.image_table:
if wellkey_cols:
# If there are well columns, fetch them.
query = "SELECT %s, %s, %s FROM %s"%(
im_clause(), well_clause(), ', '.join(meas_cols),
input_table)
else:
query = "SELECT %s, %s FROM %s"%(
im_clause(), ', '.join(meas_cols),
input_table)
elif input_table == p.object_table:
if p.image_table and wellkey_cols:
# If we have x and y from cells, we can use that for classifier
if p.cell_x_loc and p.cell_y_loc:
FIRST_MEAS_INDEX += 2 # Cell X and Y Location are fixed to for classifier
# If there are well columns, fetch them from the per-image table.
query = "SELECT %s, %s, %s, %s, %s FROM %s, %s WHERE %s"%(
dbconnect.UniqueObjectClause(p.object_table),
well_clause(p.image_table),
p.cell_x_loc,
p.cell_y_loc,
', '.join(['%s.%s'%(p.object_table, col) for col in meas_cols]),
p.image_table, p.object_table,
' AND '.join(['%s.%s=%s.%s'%(p.image_table, c, p.object_table, c)
for c in imkey_cols]) )
else:
# If there are well columns, fetch them from the per-image table.
query = "SELECT %s, %s, %s FROM %s, %s WHERE %s"%(
dbconnect.UniqueObjectClause(p.object_table),
well_clause(p.image_table),
', '.join(['%s.%s'%(p.object_table, col) for col in meas_cols]),
p.image_table, p.object_table,
' AND '.join(['%s.%s=%s.%s'%(p.image_table, c, p.object_table, c)
for c in imkey_cols]) )
else:
if p.cell_x_loc and p.cell_y_loc:
FIRST_MEAS_INDEX += 2 # Cell X and Y Location are fixed to for classifier
query = "SELECT %s, %s, %s, %s FROM %s"%(
im_clause(), p.cell_x_loc, p.cell_y_loc, ', '.join(meas_cols),
input_table)
else:
query = "SELECT %s, %s FROM %s"%(
im_clause(), ', '.join(meas_cols),
input_table)
if p.negative_control: # if the user defined negative control, we can use that to fetch the wellkeys
neg_query = query + ' AND ' + p.negative_control # fetch all the negative control elements
if wellkey_cols:
query += " ORDER BY %s"%(well_clause(p.image_table))
dlg = wx.ProgressDialog('Computing normalized values',
'Querying database for raw data.',
parent=self,
style = wx.PD_CAN_ABORT|wx.PD_APP_MODAL)
dlg.Pulse()
#
# MAKE THE QUERY
#
input_data = np.array(db.execute(query), dtype=object)
if p.negative_control:
import pandas as pd
negative_control = pd.DataFrame(db.execute(neg_query), dtype=float)
logging.info("# of objects in negative control: " + str(negative_control.shape[0]))
logging.info("# of objects queried: " + str(input_data.shape[0]))
neg_mean_plate = negative_control.groupby([WELL_KEY_INDEX]).mean()
neg_std_plate = negative_control.groupby([WELL_KEY_INDEX]).std()
output_columns = np.ones(input_data[:,FIRST_MEAS_INDEX:].shape) * np.nan
output_factors = np.ones(input_data[:,FIRST_MEAS_INDEX:].shape) * np.nan
for colnum, col in enumerate(input_data[:,FIRST_MEAS_INDEX:].T):
keep_going, skip = dlg.Pulse("Normalizing column %d of %d"%(colnum+1, len(meas_cols)))
if not keep_going:
dlg.Destroy()
return
norm_data = col.copy()
for step_num, step_panel in enumerate(self.norm_steps):
d = step_panel.get_configuration_dict()
if d[norm.P_GROUPING] in (norm.G_QUADRANT, norm.G_WELL_NEIGHBORS):
# Reshape data if normalization step is plate sensitive.
assert p.plate_id and p.well_id
well_keys = input_data[:, range(WELL_KEY_INDEX, FIRST_MEAS_INDEX - 2) ]
wellkeys_and_vals = np.hstack((well_keys, np.array([norm_data]).T))
new_norm_data = []
for plate, plate_grp in groupby(wellkeys_and_vals, lambda row: row[0]):
keys_and_vals = np.array(list(plate_grp))
plate_data, wks, ind = FormatPlateMapData(keys_and_vals)
pnorm_data = norm.do_normalization_step(plate_data, **d)
new_norm_data += pnorm_data.flatten()[ind.flatten().tolist()].tolist()
norm_data = new_norm_data
elif d[norm.P_GROUPING] == norm.G_PLATE:
assert p.plate_id and p.well_id
if d[norm.P_AGG_TYPE] == norm.M_NEGCTRL:
mean_plate_col = neg_mean_plate[colnum + FIRST_MEAS_INDEX]
std_plate_col = neg_std_plate[colnum + FIRST_MEAS_INDEX]
print(mean_plate_col)
print(std_plate_col)
well_keys = input_data[:, range(WELL_KEY_INDEX, FIRST_MEAS_INDEX - 2)]
wellkeys_and_vals = np.hstack((well_keys, np.array([norm_data]).T))
new_norm_data = []
# print wellkeys_and_vals
for plate, plate_grp in groupby(wellkeys_and_vals, lambda row: row[0]):
plate_data = np.array(list(plate_grp))[:,-1].flatten()
pnorm_data = norm.do_normalization_step(plate_data, **d)
if d[norm.P_AGG_TYPE] == norm.M_NEGCTRL:
try:
plate_mean = mean_plate_col[plate]
plate_std = std_plate_col[plate]
except:
plate_mean = mean_plate_col[int(plate)]
plate_std = std_plate_col[int(plate)]
try:
pnorm_data = (pnorm_data - plate_mean) / plate_std
print(pnorm_data)
except:
logging.error("Plate std is zero, division by zero!")
new_norm_data += pnorm_data.tolist()
norm_data = new_norm_data
else:
norm_data = norm.do_normalization_step(norm_data, **d)
output_columns[:,colnum] = np.array(norm_data)
output_factors[:,colnum] = col.astype(float) / np.array(norm_data,dtype=float)
dlg.Destroy()
return # Abort here for coding
norm_table_cols = []
# Write new table
db.execute('DROP TABLE IF EXISTS %s'%(output_table))
if input_table == p.image_table:
norm_table_cols += dbconnect.image_key_columns()
col_defs = ', '.join(['%s %s'%(col, db.GetColumnTypeString(p.image_table, col))
for col in dbconnect.image_key_columns()])
elif input_table == p.object_table:
norm_table_cols += obkey_cols
col_defs = ', '.join(['%s %s'%(col, db.GetColumnTypeString(p.object_table, col))
for col in obkey_cols])
if wellkey_cols:
norm_table_cols += wellkey_cols
col_defs += ', '+ ', '.join(['%s %s'%(col, db.GetColumnTypeString(p.image_table, col))
for col in wellkey_cols])
if input_table == p.object_table:
if p.cell_x_loc and p.cell_y_loc:
norm_table_cols += [p.cell_x_loc, p.cell_y_loc]
col_defs += ', %s %s'%(p.cell_x_loc, db.GetColumnTypeString(p.object_table, p.cell_x_loc)) + ', ' + '%s %s'%(p.cell_y_loc, db.GetColumnTypeString(p.object_table, p.cell_y_loc))
if wants_norm_meas:
col_defs += ', '+ ', '.join(['%s_NmM %s'%(col, db.GetColumnTypeString(input_table, col))
for col in meas_cols])
if wants_norm_factor:
col_defs += ', '+ ', '.join(['%s_NmF %s'%(col, db.GetColumnTypeString(input_table, col))
for col in meas_cols])
for col in meas_cols:
if wants_norm_meas:
norm_table_cols += ['%s_NmM'%(col)]
if wants_norm_factor:
norm_table_cols += ['%s_NmF'%(col)]
db.execute('CREATE TABLE %s (%s)'%(output_table, col_defs))
dlg = wx.ProgressDialog('Writing to "%s"'%(output_table),
"Writing normalized values to database",
maximum = output_columns.shape[0],
parent=self,
style = wx.PD_CAN_ABORT|wx.PD_APP_MODAL|wx.PD_ELAPSED_TIME|wx.PD_ESTIMATED_TIME|wx.PD_REMAINING_TIME)
cmd = 'INSERT INTO %s VALUES '%(output_table)
cmdi = cmd
for i, (val, factor) in enumerate(zip(output_columns, output_factors)):
cmdi += '(' + ','.join(['"%s"']*len(norm_table_cols)) + ')'
if wants_norm_meas and wants_norm_factor:
cmdi = cmdi%tuple(list(input_data[i, :FIRST_MEAS_INDEX]) +
['NULL' if (np.isnan(x) or np.isinf(x)) else x for x in val] +
['NULL' if (np.isnan(x) or np.isinf(x)) else x for x in factor])
elif wants_norm_meas:
cmdi = cmdi%tuple(list(input_data[i, :FIRST_MEAS_INDEX]) +
['NULL' if (np.isnan(x) or np.isinf(x)) else x for x in val])
elif wants_norm_factor:
cmdi = cmdi%tuple(list(input_data[i, :FIRST_MEAS_INDEX]) +
['NULL' if (np.isnan(x) or np.isinf(x)) else x for x in factor])
if (i+1) % BATCH_SIZE == 0 or i==len(output_columns)-1:
db.execute(str(cmdi))
cmdi = cmd
# update status dialog
(keep_going, skip) = dlg.Update(i)
if not keep_going:
break
else:
cmdi += ',\n'
dlg.Destroy()
db.Commit()
#
# Update table linkage
#
if db.get_linking_tables(input_table, output_table) is not None:
db.do_unlink_table(output_table)
if input_table == p.image_table:
db.do_link_tables(output_table, input_table, imkey_cols, imkey_cols)
elif input_table == p.object_table:
db.do_link_tables(output_table, input_table, obkey_cols, obkey_cols)
#
# Show the resultant table
#
import tableviewer
tv = tableviewer.TableViewer(ui.get_main_frame_or_none())
tv.Show()
tv.load_db_table(output_table)
def do_normalization(self):
if not self.validate():
# Should be unreachable
wx.MessageBox(
'Your normalization settings are invalid. Can\'t perform normalization.'
)
long_cols = [
col for col in self.col_choices.GetCheckedStrings()
if len(col) + 4 > 64
]
if long_cols:
dlg = wx.MessageDialog(
self, 'The following columns contain more '
'than 64 characters when a normalization suffix (4 '
'characters) is appended. This may cause a problem when '
'writing to the database.\n %s' % ('\n'.join(long_cols)),
'Warning', wx.OK | wx.CANCEL | wx.ICON_EXCLAMATION)
if dlg.ShowModal() == wx.ID_CANCEL:
return
dlg.Destroy()
imkey_cols = dbconnect.image_key_columns()
obkey_cols = dbconnect.object_key_columns()
wellkey_cols = dbconnect.well_key_columns()
im_clause = dbconnect.UniqueImageClause
well_clause = dbconnect.UniqueWellClause
input_table = self.table_choice.GetStringSelection()
meas_cols = self.col_choices.GetCheckedStrings()
wants_norm_meas = self.norm_meas_checkbox.IsChecked()
wants_norm_factor = self.norm_factor_checkbox.IsChecked()
output_table = self.output_table.Value
FIRST_MEAS_INDEX = len(imkey_cols + (wellkey_cols or tuple()))
if p.db_type == 'mysql':
BATCH_SIZE = 100
else:
BATCH_SIZE = 1
if input_table == p.object_table:
FIRST_MEAS_INDEX += 1 # Original
if wellkey_cols:
if input_table == p.image_table:
WELL_KEY_INDEX = len(imkey_cols)
else:
WELL_KEY_INDEX = len(imkey_cols) + 1
if db.table_exists(output_table):
dlg = wx.MessageDialog(
self, 'Are you sure you want to overwrite the table "%s"?' %
(output_table), "Overwrite table?",
wx.YES_NO | wx.NO_DEFAULT | wx.ICON_EXCLAMATION)
if dlg.ShowModal() == wx.ID_NO:
dlg.Destroy()
return
dlg.Destroy()
#
# First Get the data from the db.
#
if input_table == p.image_table:
if wellkey_cols:
# If there are well columns, fetch them.
query = "SELECT %s, %s, %s FROM %s" % (im_clause(
), well_clause(), ', '.join(meas_cols), input_table)
else:
query = "SELECT %s, %s FROM %s" % (
im_clause(), ', '.join(meas_cols), input_table)
elif input_table == p.object_table:
if p.image_table and wellkey_cols:
# If we have x and y from cells, we can use that for classifier
if p.cell_x_loc and p.cell_y_loc:
FIRST_MEAS_INDEX += 2 # Cell X and Y Location are fixed to for classifier
# If there are well columns, fetch them from the per-image table.
query = "SELECT %s, %s, %s, %s, %s FROM %s, %s WHERE %s" % (
dbconnect.UniqueObjectClause(
p.object_table), well_clause(p.image_table),
p.cell_x_loc, p.cell_y_loc, ', '.join([
'%s.%s' % (p.object_table, col)
for col in meas_cols
]), p.image_table, p.object_table, ' AND '.join([
'%s.%s=%s.%s' %
(p.image_table, c, p.object_table, c)
for c in imkey_cols
]))
else:
# If there are well columns, fetch them from the per-image table.
query = "SELECT %s, %s, %s FROM %s, %s WHERE %s" % (
dbconnect.UniqueObjectClause(p.object_table),
well_clause(p.image_table), ', '.join([
'%s.%s' % (p.object_table, col)
for col in meas_cols
]), p.image_table, p.object_table, ' AND '.join([
'%s.%s=%s.%s' %
(p.image_table, c, p.object_table, c)
for c in imkey_cols
]))
else:
if p.cell_x_loc and p.cell_y_loc:
FIRST_MEAS_INDEX += 2 # Cell X and Y Location are fixed to for classifier
query = "SELECT %s, %s, %s, %s FROM %s" % (
im_clause(), p.cell_x_loc, p.cell_y_loc,
', '.join(meas_cols), input_table)
else:
query = "SELECT %s, %s FROM %s" % (
im_clause(), ', '.join(meas_cols), input_table)
if p.negative_control: # if the user defined negative control, we can use that to fetch the wellkeys
neg_query = query + ' AND ' + p.negative_control # fetch all the negative control elements
if wellkey_cols:
query += " ORDER BY %s" % (well_clause(p.image_table))
dlg = wx.ProgressDialog('Computing normalized values',
'Querying database for raw data.',
parent=self,
style=wx.PD_CAN_ABORT | wx.PD_APP_MODAL)
dlg.Pulse()
#
# MAKE THE QUERY
#
input_data = np.array(db.execute(query), dtype=object)
if p.negative_control:
import pandas as pd
negative_control = pd.DataFrame(db.execute(neg_query), dtype=float)
logging.info("# of objects in negative control: " +
str(negative_control.shape[0]))
logging.info("# of objects queried: " + str(input_data.shape[0]))
neg_mean_plate = negative_control.groupby([WELL_KEY_INDEX]).mean()
neg_std_plate = negative_control.groupby([WELL_KEY_INDEX]).std()
output_columns = np.ones(input_data[:,
FIRST_MEAS_INDEX:].shape) * np.nan
output_factors = np.ones(input_data[:,
FIRST_MEAS_INDEX:].shape) * np.nan
for colnum, col in enumerate(input_data[:, FIRST_MEAS_INDEX:].T):
keep_going, skip = dlg.Pulse("Normalizing column %d of %d" %
(colnum + 1, len(meas_cols)))
if not keep_going:
dlg.Destroy()
return
norm_data = col.copy()
for step_num, step_panel in enumerate(self.norm_steps):
d = step_panel.get_configuration_dict()
if d[norm.P_GROUPING] in (norm.G_QUADRANT,
norm.G_WELL_NEIGHBORS):
# Reshape data if normalization step is plate sensitive.
assert p.plate_id and p.well_id
well_keys = input_data[:,
range(WELL_KEY_INDEX,
FIRST_MEAS_INDEX - 2)]
wellkeys_and_vals = np.hstack(
(well_keys, np.array([norm_data]).T))
new_norm_data = []
for plate, plate_grp in groupby(wellkeys_and_vals,
lambda row: row[0]):
keys_and_vals = np.array(list(plate_grp))
plate_data, wks, ind = FormatPlateMapData(
keys_and_vals)
pnorm_data = norm.do_normalization_step(
plate_data, **d)
new_norm_data += pnorm_data.flatten()[
ind.flatten().tolist()].tolist()
norm_data = new_norm_data
elif d[norm.P_GROUPING] == norm.G_PLATE:
assert p.plate_id and p.well_id
if d[norm.P_AGG_TYPE] == norm.M_NEGCTRL:
mean_plate_col = neg_mean_plate[colnum +
FIRST_MEAS_INDEX]
std_plate_col = neg_std_plate[colnum +
FIRST_MEAS_INDEX]
print(mean_plate_col)
print(std_plate_col)
well_keys = input_data[:,
range(WELL_KEY_INDEX,
FIRST_MEAS_INDEX - 2)]
wellkeys_and_vals = np.hstack(
(well_keys, np.array([norm_data]).T))
new_norm_data = []
# print wellkeys_and_vals
for plate, plate_grp in groupby(wellkeys_and_vals,
lambda row: row[0]):
plate_data = np.array(list(plate_grp))[:, -1].flatten()
pnorm_data = norm.do_normalization_step(
plate_data, **d)
if d[norm.P_AGG_TYPE] == norm.M_NEGCTRL:
try:
plate_mean = mean_plate_col[plate]
plate_std = std_plate_col[plate]
except:
plate_mean = mean_plate_col[int(plate)]
plate_std = std_plate_col[int(plate)]
try:
pnorm_data = (pnorm_data -
plate_mean) / plate_std
print(pnorm_data)
except:
logging.error(
"Plate std is zero, division by zero!")
new_norm_data += pnorm_data.tolist()
norm_data = new_norm_data
else:
norm_data = norm.do_normalization_step(norm_data, **d)
output_columns[:, colnum] = np.array(norm_data)
output_factors[:,
colnum] = col.astype(float) / np.array(norm_data,
dtype=float)
dlg.Destroy()
return # Abort here for coding
norm_table_cols = []
# Write new table
db.execute('DROP TABLE IF EXISTS %s' % (output_table))
if input_table == p.image_table:
norm_table_cols += dbconnect.image_key_columns()
col_defs = ', '.join([
'%s %s' % (col, db.GetColumnTypeString(p.image_table, col))
for col in dbconnect.image_key_columns()
])
elif input_table == p.object_table:
norm_table_cols += obkey_cols
col_defs = ', '.join([
'%s %s' % (col, db.GetColumnTypeString(p.object_table, col))
for col in obkey_cols
])
if wellkey_cols:
norm_table_cols += wellkey_cols
col_defs += ', ' + ', '.join([
'%s %s' % (col, db.GetColumnTypeString(p.image_table, col))
for col in wellkey_cols
])
if input_table == p.object_table:
if p.cell_x_loc and p.cell_y_loc:
norm_table_cols += [p.cell_x_loc, p.cell_y_loc]
col_defs += ', %s %s' % (
p.cell_x_loc,
db.GetColumnTypeString(p.object_table, p.cell_x_loc)
) + ', ' + '%s %s' % (p.cell_y_loc,
db.GetColumnTypeString(
p.object_table, p.cell_y_loc))
if wants_norm_meas:
col_defs += ', ' + ', '.join([
'%s_NmM %s' % (col, db.GetColumnTypeString(input_table, col))
for col in meas_cols
])
if wants_norm_factor:
col_defs += ', ' + ', '.join([
'%s_NmF %s' % (col, db.GetColumnTypeString(input_table, col))
for col in meas_cols
])
for col in meas_cols:
if wants_norm_meas:
norm_table_cols += ['%s_NmM' % (col)]
if wants_norm_factor:
norm_table_cols += ['%s_NmF' % (col)]
db.execute('CREATE TABLE %s (%s)' % (output_table, col_defs))
dlg = wx.ProgressDialog('Writing to "%s"' % (output_table),
"Writing normalized values to database",
maximum=output_columns.shape[0],
parent=self,
style=wx.PD_CAN_ABORT | wx.PD_APP_MODAL
| wx.PD_ELAPSED_TIME | wx.PD_ESTIMATED_TIME
| wx.PD_REMAINING_TIME)
cmd = 'INSERT INTO %s VALUES ' % (output_table)
cmdi = cmd
for i, (val, factor) in enumerate(zip(output_columns, output_factors)):
cmdi += '(' + ','.join(['"%s"'] * len(norm_table_cols)) + ')'
if wants_norm_meas and wants_norm_factor:
cmdi = cmdi % tuple(
list(input_data[i, :FIRST_MEAS_INDEX]) + [
'NULL' if (np.isnan(x) or np.isinf(x)) else x
for x in val
] + [
'NULL' if (np.isnan(x) or np.isinf(x)) else x
for x in factor
])
elif wants_norm_meas:
cmdi = cmdi % tuple(
list(input_data[i, :FIRST_MEAS_INDEX]) + [
'NULL' if (np.isnan(x) or np.isinf(x)) else x
for x in val
])
elif wants_norm_factor:
cmdi = cmdi % tuple(
list(input_data[i, :FIRST_MEAS_INDEX]) + [
'NULL' if (np.isnan(x) or np.isinf(x)) else x
for x in factor
])
if (i + 1) % BATCH_SIZE == 0 or i == len(output_columns) - 1:
db.execute(str(cmdi))
cmdi = cmd
# update status dialog
(keep_going, skip) = dlg.Update(i)
if not keep_going:
break
else:
cmdi += ',\n'
dlg.Destroy()
db.Commit()
#
# Update table linkage
#
if db.get_linking_tables(input_table, output_table) is not None:
db.do_unlink_table(output_table)
if input_table == p.image_table:
db.do_link_tables(output_table, input_table, imkey_cols,
imkey_cols)
elif input_table == p.object_table:
db.do_link_tables(output_table, input_table, obkey_cols,
obkey_cols)
#
# Show the resultant table
#
import tableviewer
tv = tableviewer.TableViewer(ui.get_main_frame_or_none())
tv.Show()
tv.load_db_table(output_table)
new_norm_data += pnorm_data.tolist()
norm_data = new_norm_data
else:
norm_data = norm.do_normalization_step(norm_data, **d)
output_columns[:,colnum] = np.array(norm_data)
output_factors[:,colnum] = col.astype(float) / np.array(norm_data,dtype=float)
dlg.Destroy()
return # Abort here for coding
norm_table_cols = []
# Write new table
db.execute('DROP TABLE IF EXISTS %s'%(output_table))
if input_table == p.image_table:
norm_table_cols += dbconnect.image_key_columns()
col_defs = ', '.join(['%s %s'%(col, db.GetColumnTypeString(p.image_table, col))
for col in dbconnect.image_key_columns()])
elif input_table == p.object_table:
norm_table_cols += obkey_cols
col_defs = ', '.join(['%s %s'%(col, db.GetColumnTypeString(p.object_table, col))
for col in obkey_cols])
if wellkey_cols:
norm_table_cols += wellkey_cols
col_defs += ', '+ ', '.join(['%s %s'%(col, db.GetColumnTypeString(p.image_table, col))
for col in wellkey_cols])
if input_table == p.object_table:
if p.cell_x_loc and p.cell_y_loc:
norm_table_cols += [p.cell_x_loc, p.cell_y_loc]
col_defs += ', %s %s'%(p.cell_x_loc, db.GetColumnTypeString(p.object_table, p.cell_x_loc)) + ', ' + '%s %s'%(p.cell_y_loc, db.GetColumnTypeString(p.object_table, p.cell_y_loc))