compositional-explanations 0.0.3__tar.gz → 0.0.3.2__tar.gz

{compositional_explanations-0.0.3/src/compositional_explanations.egg-info → compositional_explanations-0.0.3.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: compositional_explanations
-Version: 0.0.3
+Version: 0.0.3.2
 Summary: Package to compute Compositional Explanations both in their optimal and beam format.
 Author-email: Biagio La Rosa <bilarosa@ucsc.edu>
 License: MIT

{compositional_explanations-0.0.3 → compositional_explanations-0.0.3.2}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "compositional_explanations"
-version = "0.0.3"
+version = "0.0.3.2"
 authors = [
   { name="Biagio La Rosa", email="bilarosa@ucsc.edu" },
 ]

{compositional_explanations-0.0.3 → compositional_explanations-0.0.3.2}/src/compositional_explanations/metrics.py

@@ -7,7 +7,7 @@ import functools
 
 import torch
 
-from utils import mask_utils
+from package_utils import mask_utils
 
 
 @functools.lru_cache(maxsize=10)
@@ -43,17 +43,19 @@ def iou(vector1, vector2):
     return score
 
 
+
 def compute_iou_from_masks(*, formula, masks,  bitmaps):
-    """Compute the intersection over union between two vectors.
+    """Compute the IoU between a formula mask and a bitmap.
     Args:
-        vector1 (torch.Tensor): A tensor of shape (N, H, W) where N is the
-            number of sample.
-        vector2 (torch.Tensor): A tensor of shape (N, H, W) where N is the
-            number of sample.
+        formula (str): A string representing the formula.
+        masks (dict): A dictionary mapping concepts to their corresponding binary masks.
+        bitmaps (torch.Tensor): A tensor of shape (N, H, W) where N is the number of samples.
     Returns:
-        iou (float): The intersection over union between the two vectors.
+        iou (float): The intersection over union between the formula mask and the bitmap.
     """
     formula_mask = mask_utils.get_formula_mask(formula, masks)
+    if formula_mask.device != bitmaps.device:
+        formula_mask = formula_mask.to(bitmaps.device)
     intersection = torch.count_nonzero(formula_mask & bitmaps).item()
     v1_size = compute_hits(formula_mask).item()
     v2_size = compute_hits(bitmaps).item()

{compositional_explanations-0.0.3 → compositional_explanations-0.0.3.2}/src/compositional_explanations/optimal.py

@@ -8,8 +8,8 @@ import torch
 
 from compositional_explanations import formula as F
 from compositional_explanations import path_heuristic
-from utils import optimal_utils, mask_utils
-from utils.constants import INDEX_OR, INDEX_AND, INDEX_NOT
+from package_utils import optimal_utils, mask_utils, general_utils
+from package_utils.constants import INDEX_OR, INDEX_AND, INDEX_NOT
 
 
 def update_frontier_by_ancestors(
@@ -924,7 +924,7 @@ def compute_optimal_explanations(*, masks, bitmaps, disjoint_info, length, devic
     bitmaps = bitmaps.to(device)
     
     masks_info = mask_utils.get_dataset_info(masks, cache_dir)
-    max_size_mask = torch.numel(masks[0])
+    max_size_mask = torch.numel(general_utils.parse_mask_by_type(masks[0]))
 
     # Get quantities
     neuron_quantities, concept_quantities = optimal_utils.get_optimal_heuristic_info(

{compositional_explanations-0.0.3 → compositional_explanations-0.0.3.2}/src/compositional_explanations/optimal_sample_heuristic.py

@@ -3,9 +3,9 @@
 
 import numpy as np
 
-from utils import optimal_utils
+from package_utils import optimal_utils
 from compositional_explanations import formula as F
-from utils.constants import (
+from package_utils.constants import (
     TOP_INDEX_SAMPLE,
     BOTTOM_INDEX_SAMPLE,
     BOTTOM_INDEX_SUM,

{compositional_explanations-0.0.3 → compositional_explanations-0.0.3.2}/src/compositional_explanations/optimal_sum_heuristic.py

@@ -2,9 +2,9 @@
 
 import numpy as np
 
-from utils import optimal_utils
+from package_utils import optimal_utils
 from compositional_explanations import formula as F
-from utils.constants import TOP_INDEX_SUM, BOTTOM_INDEX_SUM
+from package_utils.constants import TOP_INDEX_SUM, BOTTOM_INDEX_SUM
 
 
 def estimate_disjoint_label_info(label, *, left_quantities, right_quantities):

{compositional_explanations-0.0.3 → compositional_explanations-0.0.3.2/src/compositional_explanations.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: compositional_explanations
-Version: 0.0.3
+Version: 0.0.3.2
 Summary: Package to compute Compositional Explanations both in their optimal and beam format.
 Author-email: Biagio La Rosa <bilarosa@ucsc.edu>
 License: MIT

{compositional_explanations-0.0.3 → compositional_explanations-0.0.3.2}/src/compositional_explanations.egg-info/SOURCES.txt

@@ -13,7 +13,7 @@ src/compositional_explanations.egg-info/SOURCES.txt
 src/compositional_explanations.egg-info/dependency_links.txt
 src/compositional_explanations.egg-info/requires.txt
 src/compositional_explanations.egg-info/top_level.txt
-src/utils/constants.py
-src/utils/general_utils.py
-src/utils/mask_utils.py
-src/utils/optimal_utils.py
+src/package_utils/constants.py
+src/package_utils/general_utils.py
+src/package_utils/mask_utils.py
+src/package_utils/optimal_utils.py

{compositional_explanations-0.0.3 → compositional_explanations-0.0.3.2}/src/compositional_explanations.egg-info/top_level.txt

@@ -1,2 +1,2 @@
 compositional_explanations
-utils
+package_utils

{compositional_explanations-0.0.3/src/utils → compositional_explanations-0.0.3.2/src/package_utils}/mask_utils.py

@@ -4,7 +4,7 @@ import pickle
 
 import torch
 
-from utils import general_utils
+from package_utils import general_utils
 from compositional_explanations import formula as F
 
 
@@ -24,6 +24,7 @@ def get_dataset_info(masks, info_directory):
             print("Loading dataset info from cache...")
             common_elements, unique_elements, uncoverable_elements = pickle.load(file)
     else:
+        print("Computing dataset info...")
         # Compute unique elements
         sum_elements = torch.zeros_like(
             general_utils.parse_mask_by_type(masks[1]), dtype=torch.int32

{compositional_explanations-0.0.3/src/utils → compositional_explanations-0.0.3.2/src/package_utils}/optimal_utils.py

@@ -7,8 +7,8 @@ import torch
 import numpy as np
 
 from compositional_explanations import formula as F
-from utils import general_utils
-from utils.constants import (
+from package_utils import general_utils
+from package_utils.constants import (
     QUANTITIES,
     INDEX_TUPLE_MAX,
     INDEX_TUPLE_MIN,