def save_results(conn, files, plate):
# Upload the results as OMERO.table
print("saving results...")
Nuclei = pandas.concat(files, ignore_index=True)
summary = Nuclei.groupby('Image').mean()
# Free memory
del Nuclei
cols = []
for col in summary.columns:
if col == 'Image':
cols.append(ImageColumn(col, '', summary[col]))
elif col == 'Well':
cols.append(WellColumn(col, '', summary[col]))
elif summary[col].dtype == 'int64':
cols.append(LongColumn(col, '', summary[col]))
elif summary[col].dtype == 'float64':
cols.append(DoubleColumn(col, '', summary[col]))
resources = conn.c.sf.sharedResources()
repository_id = resources.repositories().descriptions[0].getId().getValue()
table_name = "idr0002_cellprofiler"
table = resources.newTable(repository_id, table_name)
table.initialize(cols)
table.addData(cols)
# Link the table to the plate
orig_file = table.getOriginalFile()
file_ann = FileAnnotationWrapper(conn)
file_ann.setNs(NSBULKANNOTATIONS)
file_ann._obj.file = OriginalFileI(orig_file.id.val, False)
file_ann.save()
plate.linkAnnotation(file_ann)
table.close()
def get_empty_columns(self, n_columns):
"""
Retrieves a set of empty OmeroTables columns for the analysis results
prefixed by an ImageColumn and RoiColumn to handle these linked
object indexes.
"""
columns = [ImageColumn('Image', '', list()),
RoiColumn('ROI', '', list())]
for i in range(n_columns):
columns.append(DoubleColumn('', '', list()))
return columns
def save_table(conn, images, image_data, script_params, project=None):
"""Summarise ROIs as Table (1 row per Image) linked to Project."""
resources = conn.c.sf.sharedResources()
repository_id = resources.repositories().descriptions[0].getId().getValue()
table_name = "batch_roi_export"
table = resources.newTable(repository_id, table_name)
# Create table
image_ids = [i.id for i in images]
img_column = ImageColumn('Image', '', image_ids)
cols = [DoubleColumn(k, '', image_data[k]) for k in SUMMARY_COL_NAMES]
data = [img_column] + cols
table.initialize(data)
table.addData(data)
table.close()
if project is None:
log("No Project found to link table")
else:
link_table(conn, table, project)
def parse(self):
provider = self.original_file_provider
data = provider.get_original_file_data(self.original_file)
try:
rows = list(csv.reader(data, delimiter=","))
finally:
data.close()
columns = [
ImageColumn("Image", "", list()),
RoiColumn("ROI", "", list()),
StringColumn("Type", "", 12, list()),
]
for row in rows[1:]:
wellnumber = self.well_name_to_number(row[0])
image = self.analysis_ctx.\
image_from_wellnumber(wellnumber)
# TODO: what to do with the field?!
# field = int(row[1])
# image = images[field]
roi = RoiI()
shape = PointI()
shape.cx = rdouble(float(row[2]))
shape.cy = rdouble(float(row[3]))
shape.textValue = rstring(row[4])
roi.addShape(shape)
roi.image = image.proxy()
rid = self.update_service\
.saveAndReturnIds([roi])[0]
columns[0].values.append(image.id.val)
columns[1].values.append(rid)
columns[2].values.append(row[4])
return MeasurementParsingResult([columns])
def save_table(conn, images, export_data, script_params):
"""Summarise ROIs as Table (1 row per Image) linked to Project."""
resources = conn.c.sf.sharedResources()
repository_id = resources.repositories().descriptions[0].getId().getValue()
table_name = "batch_roi_export"
table = resources.newTable(repository_id, table_name)
# Summarise data into columns
image_ids = [i.id for i in images]
col_data = defaultdict(list)
for image in images:
data = get_summary_data_for_image(conn, image, export_data,
script_params)
for key in SUMMARY_COL_NAMES:
col_data[key].append(data[key] if data is not None else 0)
# Create table
img_column = ImageColumn('Image', '', image_ids)
cols = [DoubleColumn(k, '', col_data[k]) for k in SUMMARY_COL_NAMES]
data = [img_column] + cols
table.initialize(data)
table.addData(data)
table.close()
# Link Table to first Project we find
project = None
for image in images:
project = image.getProject()
if project is not None:
break
if project is None:
log("No Project found to link table")
else:
link_table(conn, table, project)
def parse(self):
log.info("Parsing: %s" % self.original_file.name.val)
provider = self.original_file_provider
data = provider.get_original_file_data(self.original_file)
try:
events = ('start', 'end')
well_data = None
n_roi = 0
n_measurements = 0
cells_columns = {'Image': ImageColumn('Image', '', list()),
'Cell': LongColumn('Cell', '', list()),
'ROI': RoiColumn('ROI', '', list())
}
organelles_columns = {'Image': ImageColumn('Image', '', list()),
'Cell': LongColumn('Cell', '', list()),
}
nuclei_columns = {'Image': ImageColumn('Image', '', list()),
'Cell': LongColumn('Cell', '', list()),
'ROI': RoiColumn('ROI', '', list())
}
for event, element in iterparse(data, events=events):
if event == 'start' and element.tag == 'WellData' \
and element.get('cell') != 'Summary':
row = int(element.get('row')) - 1
col = int(element.get('col')) - 1
i = int(element.get('field')) - 1
try:
well, images = self.get_well_images(row, col)
if not images:
continue
image = images[i]
except:
log.exception("ERROR: Failed to get well images")
continue
self.check_sparse_data(cells_columns.values())
self.check_sparse_data(nuclei_columns.values())
self.check_sparse_data(organelles_columns.values())
cell = long(element.get('cell'))
cells_columns['Cell'].values.append(cell)
nuclei_columns['Cell'].values.append(cell)
organelles_columns['Cell'].values.append(cell)
well_data = element
cells_columns['Image'].values.append(image.id.val)
nuclei_columns['Image'].values.append(image.id.val)
organelles_columns['Image'].values.append(image.id.val)
elif well_data is not None and event == 'start' \
and element.tag == 'Measure':
source = element.get('source')
key = element.get('key')
value = float(element.get('value'))
if source == self.CELLS_SOURCE:
columns_list = [cells_columns]
elif source == self.NUCLEI_SOURCE:
columns_list = [nuclei_columns]
elif source == self.ORGANELLES_SOURCE:
columns_list = [organelles_columns]
else:
columns_list = [cells_columns, nuclei_columns,
organelles_columns]
for columns in columns_list:
if key not in columns:
columns[key] = DoubleColumn(key, '', list())
columns[key].values.append(value)
n_measurements += 1
elif event == 'end' and element.tag == 'WellData':
if well_data is not None:
n_roi += 1
well_data.clear()
well_data = None
else:
element.clear()
# Final row sparseness check
self.check_sparse_data(cells_columns.values())
self.check_sparse_data(nuclei_columns.values())
self.check_sparse_data(organelles_columns.values())
log.info("Total ROI: %d" % n_roi)
log.info("Total measurements: %d" % n_measurements)
sets_of_columns = [cells_columns.values(), nuclei_columns.values(),
organelles_columns.values()]
return MeasurementParsingResult(sets_of_columns)
finally:
data.close()