docling-ibm-models 1.1.1__py3-none-any.whl → 1.1.3__py3-none-any.whl

docling_ibm_models/tableformer/data_management/matching_post_processor.py

@@ -383,6 +383,71 @@ class MatchingPostProcessor:
         clean_matches = json.loads(clean_matches_string)
         return clean_matches
 
+    def _find_overlapping(self, table_cells):
+
+        def correct_overlap(box1, box2):
+            # Extract coordinates from the bounding boxes
+            x1_min, y1_min, x1_max, y1_max = box1["bbox"]
+            x2_min, y2_min, x2_max, y2_max = box2["bbox"]
+
+            # Calculate the overlap in both x and y directions
+            overlap_x = min(x1_max, x2_max) - max(x1_min, x2_min)
+            overlap_y = min(y1_max, y2_max) - max(y1_min, y2_min)
+
+            # If there is no overlap, return the original boxes
+            if overlap_x <= 0 or overlap_y <= 0:
+                return box1, box2
+
+            # Decide how to push the boxes apart
+            if overlap_x < overlap_y:
+                # Push horizontally
+                if x1_min < x2_min:
+                    # Move box1 to the left and box2 to the right
+                    box1["bbox"][2] -= overlap_x
+                    box2["bbox"][0] += overlap_x
+                else:
+                    # Move box2 to the left and box1 to the right
+                    box2["bbox"][2] -= overlap_x
+                    box1["bbox"][0] += overlap_x
+            else:
+                # Push vertically
+                if y1_min < y2_min:
+                    # Move box1 up and box2 down
+                    box1["bbox"][3] -= overlap_y
+                    box2["bbox"][1] += overlap_y
+                else:
+                    # Move box2 up and box1 down
+                    box2["bbox"][3] -= overlap_y
+                    box1["bbox"][1] += overlap_y
+
+            return box1, box2
+
+        def do_boxes_overlap(box1, box2):
+            # print("{} - {}".format(box1["bbox"], box2["bbox"]))
+            # Extract coordinates from the bounding boxes
+            x1_min, y1_min, x1_max, y1_max = box1["bbox"]
+            x2_min, y2_min, x2_max, y2_max = box2["bbox"]
+            # Check if one box is to the left of the other
+            if x1_max < x2_min or x2_max < x1_min:
+                return False
+            # Check if one box is above the other
+            if y1_max < y2_min or y2_max < y1_min:
+                return False
+            return True
+
+        def find_overlapping_pairs_indexes(bboxes):
+            overlapping_indexes = []
+            # Compare each box with every other box (combinations)
+            for i in range(len(bboxes)):
+                for j in range(i + 1, len(bboxes)):
+                    if do_boxes_overlap(bboxes[i], bboxes[j]):
+                        bboxes[i], bboxes[j] = correct_overlap(bboxes[i], bboxes[j])
+
+            return overlapping_indexes, bboxes
+
+        overlapping_indexes, table_cells = find_overlapping_pairs_indexes(table_cells)
+        return table_cells
+
     def _align_table_cells_to_pdf(self, table_cells, pdf_cells, matches):
         r"""
         USED in 8.a step
@@ -1261,7 +1326,9 @@ class MatchingPostProcessor:
             dedupl_table_cells, key=lambda k: k["cell_id"]
         )
 
-        if len(pdf_cells) > 300:
+        if (
+            len(pdf_cells) > 300
+        ):  # For performance, skip this step if there are too many pdf_cells
             aligned_table_cells2 = dedupl_table_cells_sorted
         else:
             aligned_table_cells2 = self._align_table_cells_to_pdf(
@@ -1281,6 +1348,10 @@ class MatchingPostProcessor:
         table_cells_wo = po2
         max_cell_id = po3
 
+        # As the last step - correct cell bboxes in a way that they don't overlap:
+        if len(table_cells_wo) <= 300:  # For performance reasons
+            table_cells_wo = self._find_overlapping(table_cells_wo)
+
         self._log().debug("*** final_matches_wo")
         self._log().debug(final_matches_wo)
         self._log().debug("*** table_cells_wo")

docling_ibm_models/tableformer/data_management/tf_predictor.py

@@ -99,8 +99,13 @@ class TFPredictor:
     Table predictions for the in-memory Docling API
     """
 
-    def __init__(self, config):
+    def __init__(self, config, num_threads: int = None):
         r"""
+        The number of threads is decided, in the following order, by:
+        1. The init method parameter `num_threads`, if it is set.
+        2. The envvar "OMP_NUM_THREADS", if it is set.
+        3. The default value 4.
+
         Parameters
         ----------
         config : dict
@@ -123,6 +128,13 @@ class TFPredictor:
         self._post_processor = MatchingPostProcessor(config)
 
         self._init_word_map()
+
+        # Set the number of torch threads
+        if num_threads is None:
+            num_threads = int(os.environ.get("OMP_NUM_THREADS", 4))
+        self._num_threads = num_threads
+        torch.set_num_threads(num_threads)
+
         # Load the model
         self._model = self._load_model()
         self._model.eval()

docling_ibm_models/tableformer/models/common/base_model.py

@@ -257,7 +257,9 @@ class BaseModel(ABC):
                 self._log().info(
                     "Loading model checkpoint file: {}".format(checkpoint_file)
                 )
-                saved_model = torch.load(checkpoint_file, map_location=self._device)
+                saved_model = torch.load(
+                    checkpoint_file, map_location=self._device, weights_only=False
+                )
                 return saved_model, checkpoint_file
             except RuntimeError:
                 self._log().error("Cannot load file: {}".format(checkpoint_file))

docling_ibm_models/tableformer/models/table04_rs/encoder04_rs.py

@@ -30,7 +30,7 @@ class Encoder04(nn.Module):
         self.enc_image_size = enc_image_size
         self._encoder_dim = enc_dim
 
-        resnet = torchvision.models.resnet18(pretrained=False)
+        resnet = torchvision.models.resnet18()
         modules = list(resnet.children())[:-3]
 
         self._resnet = nn.Sequential(*modules)

{docling_ibm_models-1.1.1.dist-info → docling_ibm_models-1.1.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: docling-ibm-models
-Version: 1.1.1
+Version: 1.1.3
 Summary: This package contains the AI models used by the Docling PDF conversion package
 License: MIT
 Keywords: docling,convert,document,pdf,layout model,segmentation,table structure,table former
@@ -26,8 +26,7 @@ Requires-Dist: lxml (>=4.9.1,<5.0.0)
 Requires-Dist: mean_average_precision (>=2021.4.26.0,<2022.0.0.0)
 Requires-Dist: numpy (>=1.24.4,<2.0.0)
 Requires-Dist: onnxruntime (>=1.16.2,<2.0.0)
-Requires-Dist: opencv-python (>=4.9.0.80,<5.0.0.0) ; sys_platform != "linux"
-Requires-Dist: opencv-python-headless (>=4.9.0.80,<5.0.0.0) ; sys_platform == "linux"
+Requires-Dist: opencv-python-headless (>=4.9.0.80,<5.0.0.0)
 Requires-Dist: torch (>=2.2.2,<3.0.0)
 Requires-Dist: torchvision (>=0.17.2)
 Requires-Dist: tqdm (>=4.64.0,<5.0.0)
@@ -44,7 +43,7 @@ Description-Content-Type: text/markdown
 
 # Docling IBM models
 
-AI modules to support the Dockling PDF document conversion project.
+AI modules to support the Docling PDF document conversion project.
 
 - TableFormer is an AI module that recognizes the structure of a table and the bounding boxes of the table content.
 - Layout model is an AI model that provides among other things ability to detect tables on the page. This package contains inference code for Layout model.

{docling_ibm_models-1.1.1.dist-info → docling_ibm_models-1.1.3.dist-info}/RECORD

@@ -4,17 +4,17 @@ docling_ibm_models/tableformer/common.py,sha256=RV2ptqgkfz1OIoN-WqiSeln0pkZ_7zTO
 docling_ibm_models/tableformer/data_management/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 docling_ibm_models/tableformer/data_management/data_transformer.py,sha256=lNKkAk0VALbixapCuDDSIQKtA0QPCGQF8AGO3D64new,18263
 docling_ibm_models/tableformer/data_management/functional.py,sha256=UrXsEm4DSc1QXdUPb0tZ7nvbg7mGVjpQhX3pGL6C5bA,20633
-docling_ibm_models/tableformer/data_management/matching_post_processor.py,sha256=HYG-wx5PQC38hTQfyXr3zUnZ6--aSs55x8g9PdsROGU,54207
+docling_ibm_models/tableformer/data_management/matching_post_processor.py,sha256=-82B4xUJ9uxMDcsX2DJINTy3J0OB9rKXzHKtf-J3GHI,57205
 docling_ibm_models/tableformer/data_management/tf_cell_matcher.py,sha256=DFu428Cr84maT9WehdoZkpkJKeahwe5JlclvTC6fuVY,20870
 docling_ibm_models/tableformer/data_management/tf_dataset.py,sha256=6_qSsYt6qoE2JBzUNrJfCDX3Kgg7tyrv3kimGLdEQ5o,49890
-docling_ibm_models/tableformer/data_management/tf_predictor.py,sha256=3e7ICgs3r9YclOLWuoOvqSveQHcERvZYmOBMtmslJqU,38502
+docling_ibm_models/tableformer/data_management/tf_predictor.py,sha256=LxRme9AWLZhQw7xP5Tpxwa5XFTY66m5IFS5v9VC30GA,38978
 docling_ibm_models/tableformer/data_management/transforms.py,sha256=_i1HXkX8LAuHbeGRrg8kF9yFNJRQZOKmWzxKt559ABQ,13268
 docling_ibm_models/tableformer/models/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 docling_ibm_models/tableformer/models/common/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-docling_ibm_models/tableformer/models/common/base_model.py,sha256=_Pn6hjIx49DVTU-po6qsR788RhD7Q4FhVyBqaGl0tMw,9972
+docling_ibm_models/tableformer/models/common/base_model.py,sha256=SbCjeEvDmGnyoKYhB5pYeg2LFVQdArglfrhqkuW1nUw,10030
 docling_ibm_models/tableformer/models/table04_rs/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 docling_ibm_models/tableformer/models/table04_rs/bbox_decoder_rs.py,sha256=mMZSGk0PjQ4_fhuP44_WJVzfxyLky6S3zlVxCrNKRgc,5991
-docling_ibm_models/tableformer/models/table04_rs/encoder04_rs.py,sha256=vlDW890mCIzHdgaGXFZ4avlnCmuUCPHUaJ30P9N6eWo,1991
+docling_ibm_models/tableformer/models/table04_rs/encoder04_rs.py,sha256=iExmqJ0Pn0lJU3nWb_x8abTn42GctMqE55_YA2ppgvc,1975
 docling_ibm_models/tableformer/models/table04_rs/tablemodel04_rs.py,sha256=7iGkrTNLzjC1yn1zuA3N6DvBvbrcO_BR5tmHG3RKmXs,12159
 docling_ibm_models/tableformer/models/table04_rs/transformer_rs.py,sha256=4106qxxH0w92CVOFzFuCb87tRMvqAUP3X3F1WT5Z47A,6371
 docling_ibm_models/tableformer/otsl.py,sha256=k8l1hYWvcCkcnWbLxuBUYEcigYBFTRqiM2GBAHcUDok,21024
@@ -26,7 +26,7 @@ docling_ibm_models/tableformer/utils/app_profiler.py,sha256=13dvwo5byzfP2ejqGBFw
 docling_ibm_models/tableformer/utils/torch_utils.py,sha256=uN0rK9mSXy1ewBnBnILrWebJhhVU4N-XJZBqNiLJwlQ,8893
 docling_ibm_models/tableformer/utils/utils.py,sha256=8Bxf1rEn977lFbY9NX0r5xh9PvxIRipQZX_EZW92XfA,10980
 docling_ibm_models/tableformer/utils/variance.py,sha256=USjRwaMsCmzvc6PeWskaAJnUjbliRVd_MqNKLjMDQw8,4675
-docling_ibm_models-1.1.1.dist-info/LICENSE,sha256=ACwmltkrXIz5VsEQcrqljq-fat6ZXAMepjXGoe40KtE,1069
-docling_ibm_models-1.1.1.dist-info/METADATA,sha256=u1QQMa4RVoxKF7rW30Aa7WeY-JIll0fWxQtLUZNJbCU,7035
-docling_ibm_models-1.1.1.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
-docling_ibm_models-1.1.1.dist-info/RECORD,,
+docling_ibm_models-1.1.3.dist-info/LICENSE,sha256=ACwmltkrXIz5VsEQcrqljq-fat6ZXAMepjXGoe40KtE,1069
+docling_ibm_models-1.1.3.dist-info/METADATA,sha256=omUhpVOQHmnNTPqIyHfFWDlzelyrLvDumi6bCu_tCNA,6931
+docling_ibm_models-1.1.3.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
+docling_ibm_models-1.1.3.dist-info/RECORD,,