nodes/set-soft-ComfyUI-ImageMisc
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[Added][ConsolidateMetrics] Resume consolidation

Browse Source
This commit is contained in:
Salvador E. Tropea
2025-11-24 19:26:20 -03:00
parent a33906ede8
commit 70e753333f
2 changed files with 104 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files
example_workflows/Compare_Masks_Dataset7.json
+1
View File
File diff suppressed because one or more lines are too long
src/nodes/nodes_img.py
+103 -17
View File
@@ -102,6 +102,8 @@ SOD_NAMES = {'mae': "MAE", 'max_f_mes': "Max F-measure", 'adp_f_mes': "Adp F-mea
REVERSE_SOD_NAMES = {v: k for k, v in SOD_NAMES.items()}
IS_VECT_NAME = re.compile(r"(F|FP|FR|E)\((\d+)\)")
VECT_NAMES = {'e', 'f', 'fp', 'fr'}
DATASET_NAMES = ['Fmax dataset', 'Fmean dataset', 'Emax dataset', 'Emean dataset']
# A dictionary to cache loaded fonts
font_cache = {}
@@ -160,7 +162,7 @@ def expand_header_keys(header):
key = key.upper()
h.extend([f"{key}({c})" for c in range(F_POINTS)])
else:
h.append(SOD_NAMES[key])
h.append(SOD_NAMES.get(key, key))
return h
@@ -990,6 +992,10 @@ class ConsolidateMetrics(ComfyNodeABC):
"If this is a directory the file\nwill be named `consolidated.csv` inside it"
}),
},
"optional": {
"metric_ids": (IO.STRING, ),
"res_destination": (IO.STRING, ),
},
}
INPUT_IS_LIST = True
@@ -1028,7 +1034,7 @@ class ConsolidateMetrics(ComfyNodeABC):
# Read the single value from the current position.
scalar_str_value = row_data[data_col_idx]
# Convert to float and store it in the dictionary.
metric_values[key] = float(scalar_str_value)
metric_values[key] = float(scalar_str_value) if scalar_str_value else 0
# Advance the column index by one.
data_col_idx += 1
return metric_values
@@ -1037,6 +1043,7 @@ class ConsolidateMetrics(ComfyNodeABC):
existing_data = {}
header = []
metric_keys_ordered = []
totals = {}
if dest_path.is_file():
try:
@@ -1057,7 +1064,7 @@ class ConsolidateMetrics(ComfyNodeABC):
if index == "0":
metric_keys_ordered.append(vect_name.lower())
else:
metric_keys_ordered.append(REVERSE_SOD_NAMES.get(h))
metric_keys_ordered.append(REVERSE_SOD_NAMES.get(h, h))
# Load existing rows, stopping at any blank line (which precedes totals).
for row in reader:
@@ -1071,16 +1078,29 @@ class ConsolidateMetrics(ComfyNodeABC):
metric_values = self.compact_row(row[1:], metric_keys_ordered)
existing_data[filename] = metric_values
totals_row = next(reader, [''])
if totals_row[0] == 'Total':
totals = self.compact_row(totals_row[1:], metric_keys_ordered)
for row in reader:
if not row:
continue
if len(row) >= 2:
totals[row[0]] = row[1]
except (IOError, StopIteration, IndexError, ValueError) as e:
logger.warning(f"Could not properly read existing file at {dest_path}. It will be overwritten. Error: {e}")
raise
existing_data = {} # Reset on read error
return existing_data, header, metric_keys_ordered
return existing_data, header, metric_keys_ordered, totals
def execute(self, metrics, img_name, destination):
def execute(self, metrics, img_name, destination, metric_ids=None, res_destination=None):
n_metrics = len(metrics)
n_img_name = len(img_name)
n_destination = len(destination)
n_res_destination = 0 if res_destination is None else len(res_destination)
n_metric_ids = 0 if metric_ids is None else len(metric_ids)
logger.debug(f"metrics: {n_metrics}")
logger.debug(f"img_name: {n_img_name} {img_name}")
@@ -1093,13 +1113,24 @@ class ConsolidateMetrics(ComfyNodeABC):
raise ValueError(f"Got {n_metrics} metrics and {n_img_name} file names. They must match.")
if n_metrics % n_destination != 0:
raise ValueError(f"Got {n_metrics} metrics and {n_destination} destinations. They must be multiples.")
if n_res_destination > 1:
raise ValueError(f"Only one resume destination currently supported, got {n_res_destination}")
if n_res_destination and n_metric_ids != n_metrics:
raise ValueError(f"Got {n_metrics} metrics and {n_metric_ids} metric IDs. They must match.")
slice_size = n_metrics // n_destination
res = []
totals = {}
for c, d in enumerate(destination):
start = c * slice_size
end = start + slice_size
res.append(self.consolidate(metrics[start:end], img_name[start:end], d))
data, total = self.consolidate(metrics[start:end], img_name[start:end], d)
res.append(data)
if n_res_destination:
totals[metric_ids[c]] = total
if totals:
self.update_resume(res_destination[0], totals)
return (res, )
@@ -1108,7 +1139,7 @@ class ConsolidateMetrics(ComfyNodeABC):
if destination is None:
# We don't even know where to consolidate data
return [{}]
return [{}], {}
# Resolve the final destination path for the CSV file.
dest_path = Path(get_output_directory(), destination)
@@ -1118,14 +1149,14 @@ class ConsolidateMetrics(ComfyNodeABC):
if metrics[0] is None or img_name[0] is None:
# This is normal when all images are processed and we aren't blocking
# In this case return what we already have on disk
existing_data, header, metric_keys_ordered = self.load_current_data(dest_path)
return [v for v in existing_data.values()]
existing_data, header, metric_keys_ordered, totals = self.load_current_data(dest_path)
return [v for v in existing_data.values()], totals
# Ensure the parent directory exists.
dest_path.parent.mkdir(exist_ok=True)
# --- 2. Load Existing Data from CSV (if it exists) ---
existing_data, header, metric_keys_ordered = self.load_current_data(dest_path)
existing_data, header, metric_keys_ordered, _ = self.load_current_data(dest_path)
# --- 3. Consolidate New Metrics ---
@@ -1136,7 +1167,7 @@ class ConsolidateMetrics(ComfyNodeABC):
if not existing_data:
logger.warning("[Warning] No metrics to consolidate. Aborting file write.")
return [{}]
return [{}], {}
# --- 4. Prepare for Writing (Sort and Define Header if New) ---
@@ -1191,25 +1222,80 @@ class ConsolidateMetrics(ComfyNodeABC):
Fmax = averages.get('f').max() if 'f' in averages else 0.0
Emax = averages.get('e').max() if 'e' in averages else 0.0
if Fmax or Fmax:
if Fmax or Emax:
writer.writerow([])
if Fmax:
# This is the maximum for the average F-measure
# Is more representative for the dataset than the average of the maximums of each image
writer.writerow(['Fmax dataset', f"{Fmax:.4f}"])
key = 'Fmax dataset'
val = f"{Fmax:.4f}"
writer.writerow([key, val])
averages[key] = val
Ftot = averages.get('f').sum()
writer.writerow(['Fmean dataset', f"{Ftot/F_POINTS:.4f}"])
key = 'Fmean dataset'
val = f"{Ftot/F_POINTS:.4f}"
writer.writerow([key, val])
averages[key] = val
# Do we have E(th)?
if Emax:
writer.writerow(['Emax dataset', f"{Emax:.4f}"])
key = 'Emax dataset'
val = f"{Emax:.4f}"
writer.writerow([key, val])
averages[key] = val
Etot = averages.get('e').sum()
writer.writerow(['Emean dataset', f"{Etot/F_POINTS:.4f}"])
key = 'Emean dataset'
val = f"{Etot/F_POINTS:.4f}"
writer.writerow([key, val])
averages[key] = val
logger.info(f"Metrics consolidated and saved to {dest_path}")
return [v for v in existing_data.values()]
return [v for v in existing_data.values()], averages
def update_resume(self, destination, totals):
# Resolve the final destination path for the CSV file.
dest_path = Path(get_output_directory(), destination)
if dest_path.is_dir():
dest_path = dest_path / 'consolidated.csv'
# Ensure the parent directory exists.
dest_path.parent.mkdir(exist_ok=True)
# Load Existing Data from CSV (if it exists)
existing_data, header, metric_keys_ordered, _ = self.load_current_data(dest_path)
# Consolidate New Metrics
# Add or update the new metrics into our dictionary of existing data.
existing_data.update(totals)
if not existing_data:
return
# Prepare for Writing (Sort and Define Header if New)
# If the file was new, define the header and key order now.
if not header:
metric_keys_ordered = list(SOD_NAMES.keys()) + DATASET_NAMES + list(VECT_NAMES)
header = ["ID"] + expand_header_keys(metric_keys_ordered)
# Sort the consolidated data alphabetically by ID.
sorted_ids = sorted(existing_data.keys())
# Write Consolidated File
with open(dest_path, 'w', newline='') as f:
writer = csv.writer(f)
# Write the header.
writer.writerow(expand_header(header))
# Write the sorted data rows.
for id in sorted_ids:
metric_dict = existing_data[id]
row_data = [id] + expand_data_row(metric_dict, metric_keys_ordered)
writer.writerow(row_data)
logger.info(f"Resumed metrics consolidated and saved to {dest_path}")
return
# Most code from Gemini 3 Pro