docling 2.46.0__py3-none-any.whl → 2.47.0__py3-none-any.whl

docling/utils/layout_postprocessor.py

@@ -239,15 +239,18 @@ class LayoutPostprocessor:
         final_clusters = self._sort_clusters(
             self.regular_clusters + self.special_clusters, mode="id"
         )
-        for cluster in final_clusters:
-            cluster.cells = self._sort_cells(cluster.cells)
-            # Also sort cells in children if any
-            for child in cluster.children:
-                child.cells = self._sort_cells(child.cells)
 
-        assert self.page.parsed_page is not None
-        self.page.parsed_page.textline_cells = self.cells
-        self.page.parsed_page.has_lines = len(self.cells) > 0
+        # Conditionally process cells if not skipping cell assignment
+        if not self.options.skip_cell_assignment:
+            for cluster in final_clusters:
+                cluster.cells = self._sort_cells(cluster.cells)
+                # Also sort cells in children if any
+                for child in cluster.children:
+                    child.cells = self._sort_cells(child.cells)
+
+            assert self.page.parsed_page is not None
+            self.page.parsed_page.textline_cells = self.cells
+            self.page.parsed_page.has_lines = len(self.cells) > 0
 
         return final_clusters, self.cells
 
@@ -264,36 +267,38 @@ class LayoutPostprocessor:
             if cluster.label in self.LABEL_REMAPPING:
                 cluster.label = self.LABEL_REMAPPING[cluster.label]
 
-        # Initial cell assignment
-        clusters = self._assign_cells_to_clusters(clusters)
-
-        # Remove clusters with no cells (if keep_empty_clusters is False),
-        # but always keep clusters with label DocItemLabel.FORMULA
-        if not self.options.keep_empty_clusters:
-            clusters = [
-                cluster
-                for cluster in clusters
-                if cluster.cells or cluster.label == DocItemLabel.FORMULA
-            ]
-
-        # Handle orphaned cells
-        unassigned = self._find_unassigned_cells(clusters)
-        if unassigned and self.options.create_orphan_clusters:
-            next_id = max((c.id for c in self.all_clusters), default=0) + 1
-            orphan_clusters = []
-            for i, cell in enumerate(unassigned):
-                conf = cell.confidence
-
-                orphan_clusters.append(
-                    Cluster(
-                        id=next_id + i,
-                        label=DocItemLabel.TEXT,
-                        bbox=cell.to_bounding_box(),
-                        confidence=conf,
-                        cells=[cell],
+        # Conditionally assign cells to clusters
+        if not self.options.skip_cell_assignment:
+            # Initial cell assignment
+            clusters = self._assign_cells_to_clusters(clusters)
+
+            # Remove clusters with no cells (if keep_empty_clusters is False),
+            # but always keep clusters with label DocItemLabel.FORMULA
+            if not self.options.keep_empty_clusters:
+                clusters = [
+                    cluster
+                    for cluster in clusters
+                    if cluster.cells or cluster.label == DocItemLabel.FORMULA
+                ]
+
+            # Handle orphaned cells
+            unassigned = self._find_unassigned_cells(clusters)
+            if unassigned and self.options.create_orphan_clusters:
+                next_id = max((c.id for c in self.all_clusters), default=0) + 1
+                orphan_clusters = []
+                for i, cell in enumerate(unassigned):
+                    conf = cell.confidence
+
+                    orphan_clusters.append(
+                        Cluster(
+                            id=next_id + i,
+                            label=DocItemLabel.TEXT,
+                            bbox=cell.to_bounding_box(),
+                            confidence=conf,
+                            cells=[cell],
+                        )
                     )
-                )
-            clusters.extend(orphan_clusters)
+                clusters.extend(orphan_clusters)
 
         # Iterative refinement
         prev_count = len(clusters) + 1
@@ -350,12 +355,15 @@ class LayoutPostprocessor:
                         b=max(c.bbox.b for c in contained),
                     )
 
-                # Collect all cells from children
-                all_cells = []
-                for child in contained:
-                    all_cells.extend(child.cells)
-                special.cells = self._deduplicate_cells(all_cells)
-                special.cells = self._sort_cells(special.cells)
+                # Conditionally collect cells from children
+                if not self.options.skip_cell_assignment:
+                    all_cells = []
+                    for child in contained:
+                        all_cells.extend(child.cells)
+                    special.cells = self._deduplicate_cells(all_cells)
+                    special.cells = self._sort_cells(special.cells)
+                else:
+                    special.cells = []
 
         picture_clusters = [
             c for c in special_clusters if c.label == DocItemLabel.PICTURE

{docling-2.46.0.dist-info → docling-2.47.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docling
-Version: 2.46.0
+Version: 2.47.0
 Summary: SDK and CLI for parsing PDF, DOCX, HTML, and more, to a unified document representation for powering downstream workflows such as gen AI applications.
 Author-email: Christoph Auer <cau@zurich.ibm.com>, Michele Dolfi <dol@zurich.ibm.com>, Maxim Lysak <mly@zurich.ibm.com>, Nikos Livathinos <nli@zurich.ibm.com>, Ahmed Nassar <ahn@zurich.ibm.com>, Panos Vagenas <pva@zurich.ibm.com>, Peter Staar <taa@zurich.ibm.com>
 License-Expression: MIT
@@ -59,6 +59,7 @@ Provides-Extra: vlm
 Requires-Dist: transformers<5.0.0,>=4.46.0; extra == "vlm"
 Requires-Dist: accelerate<2.0.0,>=1.2.1; extra == "vlm"
 Requires-Dist: mlx-vlm<1.0.0,>=0.3.0; (python_version >= "3.10" and sys_platform == "darwin" and platform_machine == "arm64") and extra == "vlm"
+Requires-Dist: vllm<1.0.0,>=0.10.0; (python_version >= "3.10" and sys_platform == "linux") and extra == "vlm"
 Provides-Extra: rapidocr
 Requires-Dist: rapidocr-onnxruntime<2.0.0,>=1.4.0; python_version < "3.13" and extra == "rapidocr"
 Requires-Dist: onnxruntime<2.0.0,>=1.7.0; extra == "rapidocr"

{docling-2.46.0.dist-info → docling-2.47.0.dist-info}/RECORD

@@ -9,12 +9,12 @@ docling/backend/csv_backend.py,sha256=2g9famYG2W-ID9jEdZPxc6O8QGv1vWQfjN8pL-QMBE
 docling/backend/docling_parse_backend.py,sha256=9rUo1vPxX6QLzGqF-2B2iEYglZg6YQ3Uea00XrLluTg,7918
 docling/backend/docling_parse_v2_backend.py,sha256=3ckTfke8IICjaImlIzc3TRhG7KDuxDDba0AuCEcjA-M,9500
 docling/backend/docling_parse_v4_backend.py,sha256=MbCMxNGmoW4iuev9tX1Vt4jtIeak2kC9Uac3xQSRxeo,7509
-docling/backend/html_backend.py,sha256=zJH4wkcyftvoA-ixC4MH-xjwl-TGTN9BvZT7Hhla2mc,34701
+docling/backend/html_backend.py,sha256=qPXmMiKxskSDJJK5e0a46xhkSiATgjyi02eMrY_ahR8,38323
 docling/backend/md_backend.py,sha256=qCI7SD9hnWWGrkG_drpzQv2Z7DVBG4Tsq3hhTsYV790,22562
 docling/backend/mets_gbs_backend.py,sha256=EA8sY6tbmGiysKGYPPZiNlK-i7Adn8bLTo-7Ym15hTU,12774
 docling/backend/msexcel_backend.py,sha256=cq8MQ2RSh6pqCiVrldjOerSww7dOPTWmCQoCBI57i6w,18579
 docling/backend/mspowerpoint_backend.py,sha256=wJgB2JStEPfD7MPpWQlpPN7bffPxaHFUnKD4wj8SLxU,15114
-docling/backend/msword_backend.py,sha256=DxMgPcq-Ao1vq7X2v8qqWeMs9MryPw_Jw3YRAAdXBtM,44904
+docling/backend/msword_backend.py,sha256=fKeAMGGR5ABimedo_ofCQAybzdqmqWA3A3mpLl7X6qY,49129
 docling/backend/noop_backend.py,sha256=EOPbD86FzZPX-K_DpNrJh0_lC0bZz--4DpG-OagDNGY,1688
 docling/backend/pdf_backend.py,sha256=Wcd1NSrAMjXK8VicTki5p-j-JLofklt07eF0kIG17_0,3361
 docling/backend/pypdfium2_backend.py,sha256=8dVniLHgiTdJuDbYr66kPp6Ccv5ZDlqDMEbA2xIfS7U,13370
@@ -29,23 +29,23 @@ docling/backend/xml/jats_backend.py,sha256=LPj33EFdi2MRCakkLWrRLlUAc-B-949f8zp5g
 docling/backend/xml/uspto_backend.py,sha256=nyAMr5ht7dclxkVDwsKNeiOhLQrUtRLS8JdscB2AVJg,70924
 docling/chunking/__init__.py,sha256=h83TDs0AuOV6oEPLAPrn9dpGKiU-2Vg6IRNo4cv6GDA,346
 docling/cli/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-docling/cli/main.py,sha256=-W_vdKvSm5gZUZyvRpFH0YMI_1iJrP5sJOZ5_1bLorw,30359
-docling/cli/models.py,sha256=9yLGp6QRJGpR86U3SjmWAXDt3MvBaJLLY4xDVdsu3O8,4160
+docling/cli/main.py,sha256=yRgDoc60wm5py2ztq6-q5BPnxfLJZ1EA8Bce-EcpLPs,30952
+docling/cli/models.py,sha256=5C3CZz3HZXoCrBl92Is62KMCtUqsZK-oygj1hqzJ8vo,6008
 docling/cli/tools.py,sha256=QhtRxQG0TVrfsMqdv5i7J0_qQy1ZZyWYnHPwJl7b5oY,322
 docling/datamodel/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 docling/datamodel/accelerator_options.py,sha256=wv6dOFTVAwr9onkE-0pfUqX_fDb6gX53iPPE6o8nKjI,2511
 docling/datamodel/asr_model_specs.py,sha256=Wg7z3zm_wXIWu122iPVy0RMECsA_JCFHrlFF-xxHoVQ,2187
-docling/datamodel/base_models.py,sha256=Ifd8PPHs4sW7ScwSqpa-y3rwgPbde_iw13Y2NUCPfU8,11944
+docling/datamodel/base_models.py,sha256=OI2-tBjH3PZMF_Zyyc4eezJ4gFXIBiKT4BYKYy6n81E,11924
 docling/datamodel/document.py,sha256=zsxFYXvo6GtwGNogSDoBB1TFvkm7IOrP_VnqXNqBhJs,17329
 docling/datamodel/layout_model_specs.py,sha256=GSkJ-Z_0PVgwWGi7C7TsxbzRjlrWS9ZrHJjHumv-Z5U,2339
-docling/datamodel/pipeline_options.py,sha256=vOLpuVF-d4nmr-L16EmmhGFn25SDsgExCfX5kPiyISg,10470
+docling/datamodel/pipeline_options.py,sha256=x0RlEdTiEU9gH27YDRov1ZVMpTlx4BnqEoEtmOHd08k,10584
 docling/datamodel/pipeline_options_asr_model.py,sha256=7X068xl-qpbyPxC7-TwX7Q6tLyZXGT5h1osZ_xLNLM0,1454
-docling/datamodel/pipeline_options_vlm_model.py,sha256=eH-Cj_8aic9FdX4xGlBcf5_R9e152JAL2LhtY8d0rhw,2498
+docling/datamodel/pipeline_options_vlm_model.py,sha256=AcqqThSW74hwQ6x7pazzm57LnJiUqB7gQi5wFayGlbk,2628
 docling/datamodel/settings.py,sha256=c0MTw6pO5be_BKxHKYl4SaBJAw_qL-aapxp-g5HHj1A,2084
-docling/datamodel/vlm_model_specs.py,sha256=--jZexGeu-s_lWp7y_WwWEf6CD1J4XqADrS1-OY_pWM,4737
+docling/datamodel/vlm_model_specs.py,sha256=dFObfYlPyN7AbTCudsubsWvWTTx4F4Xz9GEJPkEV2_M,8175
 docling/models/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 docling/models/api_vlm_model.py,sha256=-zisU32pgDRbychyG6-neB0qweNbPaYnLXwiGT7SEdI,2859
-docling/models/base_model.py,sha256=NNjIapqCruAEAWR-CCdsNgXc2QkwiPYAcaQ_ZYe1W28,2978
+docling/models/base_model.py,sha256=tXFM7zJwF6Kn2EhtaB4QmgK4O2ruv1C7SjdBgM5QKak,6225
 docling/models/base_ocr_model.py,sha256=kT8TylASOpPlY60rIG6VL6_eLVsfg5KvEVnZHzDWtR0,8193
 docling/models/code_formula_model.py,sha256=XRugm4EwifLRc-TrAk-glKlktJP-nAPneKh2EOovkJU,11308
 docling/models/document_picture_classifier.py,sha256=9JvoWeH5uQBC7levjM8zptk7UT-b8EQnD-2EnxTjTT4,6202
@@ -53,10 +53,10 @@ docling/models/easyocr_model.py,sha256=ECPBd-48cCw5s935NsPJO_C_1QuK_yAUGloMM77Wq
 docling/models/layout_model.py,sha256=Nfbo6keMB4vVjGoZdFMqD9CmZcWh-0bE3LkRjJTDJQ0,9146
 docling/models/ocr_mac_model.py,sha256=y-1DSFDbACHpEwNTfQwzN9ab8r5j5rBFNPtQ48BzsrA,5396
 docling/models/page_assemble_model.py,sha256=TvN1naez7dUodLxpUUBzpuMCpqZBTf6YSpewxgjzmrg,6323
-docling/models/page_preprocessing_model.py,sha256=x8MI4mvjizqEqAb5511dtrNRCJSb-lSmwHw0tmHPFiI,5103
+docling/models/page_preprocessing_model.py,sha256=rHNX1uP1ScTjVUlsxZ0eamK2uNUqI94WpnyrP10Pj6k,5277
 docling/models/picture_description_api_model.py,sha256=o3EkV5aHW_6WzE_fdj_VRnNCrS_btclO_ZCLAUqrfl0,2377
 docling/models/picture_description_base_model.py,sha256=kLthLhdlgwhootQ4_xhhcAk6A-vso5-qcsFJ3TcYfO0,2991
-docling/models/picture_description_vlm_model.py,sha256=yfyAFOy8RjxQJrafPMSAMrrpaYu3anahjRX6tCnVcs0,4028
+docling/models/picture_description_vlm_model.py,sha256=5BJvaF3PHuL9lCVYqPv9krh3h_7YwNSdKYw1EVEj13k,4156
 docling/models/rapid_ocr_model.py,sha256=AMdc66s_iWO4p6nQ0LNjQMUYVxrDSxMyLNPpjPYt6N8,5916
 docling/models/readingorder_model.py,sha256=bZoXHaSwUsa8niSmJrbCuy784ixCeBXT-RQBUfgHJ4A,14925
 docling/models/table_structure_model.py,sha256=RFXo73f2q4XuKyaSqbxpznh7JVtlLcT0FsOWl9oZbSg,12518
@@ -70,22 +70,23 @@ docling/models/plugins/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3
 docling/models/plugins/defaults.py,sha256=OAHWW2tCcUXSyDMFxV_lXVRjSBJ1n6z-Eb3R8cDucU4,886
 docling/models/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 docling/models/utils/hf_model_download.py,sha256=scBEfsM4yl7xPzqe7UtPvDh9RfQZQnuOhqQKilYBHls,984
-docling/models/vlm_models_inline/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-docling/models/vlm_models_inline/hf_transformers_model.py,sha256=Rwdr7neDpn5ehtrp6n7G21fcPBK2m9Har_6BFNdyw-Q,8359
-docling/models/vlm_models_inline/mlx_model.py,sha256=YYYmopsITlX17JVS5KhLlb1IQSEVoSECNx_fXLHNpAc,5880
+docling/models/vlm_models_inline/__init__.py,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
+docling/models/vlm_models_inline/hf_transformers_model.py,sha256=G0RpKwdzm5NiqIBHG5nWLwBsrDfDebzErzRkyXppZPw,12134
+docling/models/vlm_models_inline/mlx_model.py,sha256=VP05v97mqzmaG4o9bOpJcxIlEqvNzAapJ15Zz3E3ACI,10169
+docling/models/vlm_models_inline/vllm_model.py,sha256=_EnK1nfpAPJky7aRlyp8SUIghiZOQO8AkDN_hHqXLZg,8615
 docling/pipeline/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 docling/pipeline/asr_pipeline.py,sha256=tQkhu9fXdkSuYIL22xzV2YRUlQh-9qktHBbs2qeXhJI,9070
 docling/pipeline/base_pipeline.py,sha256=VYVYndifTPSD2GWHKjfi4Y76M5qgt1DiygO-jowKsqM,9919
 docling/pipeline/simple_pipeline.py,sha256=TXZOwR7hZRji462ZTIpte0VJjzbxvNVE8dbLFANDhSU,2253
 docling/pipeline/standard_pdf_pipeline.py,sha256=yFishq4Cu01BiBGHk3Irr7ogcTQKeSC0QZImQVAhIaY,12740
-docling/pipeline/threaded_standard_pdf_pipeline.py,sha256=miPIyprtzPFYG94n6PmUgK4Nh7rqACYEGkWrlTbrZAc,26133
-docling/pipeline/vlm_pipeline.py,sha256=0lj8tbXNpYF8OLBoLqP2BZfFpTHi40RoHVfvO_Nah4Q,15349
+docling/pipeline/threaded_standard_pdf_pipeline.py,sha256=NgdZxpfpElnvCgGlrQ8kSvq44LNzJcc6wOqD-AMrKZ0,26132
+docling/pipeline/vlm_pipeline.py,sha256=PlllB2ZbhuUQilJ5_W8RhuBUbLtXdSjnrYnTVzHF0Vc,15955
 docling/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 docling/utils/accelerator_utils.py,sha256=DSajLxVx1JEVT0zt5de26llciLNlVfIDfSa2zYCFJzQ,2909
 docling/utils/api_image_request.py,sha256=_CgdzmPqdsyXmyYUFGLZcXcoH586qC6A1p5vsNbj1Q0,1416
 docling/utils/export.py,sha256=VwVUnYDk3mhGmISDbVm306fwpGNnoojouStBD4UajXI,4673
 docling/utils/glm_utils.py,sha256=TKOWQqWAHsX_w4fvoAA7_2xCi_urhnp1DsmjY8_sk5w,12274
-docling/utils/layout_postprocessor.py,sha256=LFLbBE-o3kWu79d8ZcyHlZPIqzQfCabZCIPTJ51lZsY,24657
+docling/utils/layout_postprocessor.py,sha256=sE9UR3Nv4iOk26uoIsN3bFioE7ScfAjj0orDBDneLXg,25166
 docling/utils/locks.py,sha256=RzqQtD5UispgV71pGN_nU6GYfeN11BN0Sh_Dq9ycqGo,52
 docling/utils/model_downloader.py,sha256=3vijCsAIVwWqehGBDRxRq7mJ3yRb9-zBsG00iqjqegU,4076
 docling/utils/ocr_utils.py,sha256=nmresYyfin0raanpQc_GGeU3WoLsfExf6SEXNIQ7Djg,2325
@@ -93,9 +94,9 @@ docling/utils/orientation.py,sha256=jTyLxyT31FlOodZoBMlADHNQK2lAWKYVs5z7pXd_6Cg,
 docling/utils/profiling.py,sha256=YaMGoB9MMZpagF9mb5ndoHj8Lpb9aIdb7El-Pl7IcFs,1753
 docling/utils/utils.py,sha256=kJtIYuzXeOyJHYlxmLAo7dGM5rEsDa1i84qEsUj1nio,1908
 docling/utils/visualization.py,sha256=tY2ylE2aiQKkmzlSLnFW-HTfFyqUUMguW18ldd1PLfo,2868
-docling-2.46.0.dist-info/licenses/LICENSE,sha256=mBb7ErEcM8VS9OhiGHnQ2kk75HwPhr54W1Oiz3965MY,1088
-docling-2.46.0.dist-info/METADATA,sha256=fm7KVaUwGryyuRk7R_AkNSHo1BogY8-ra9gpCWXbnCA,10459
-docling-2.46.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-docling-2.46.0.dist-info/entry_points.txt,sha256=hzVlbeE0aMSTQ9S0-NTYN0Hmgsn6qL_EA2qX4UbkAuY,149
-docling-2.46.0.dist-info/top_level.txt,sha256=vkIywP-USjFyYo1AIRQbWQQaL3xB5jf8vkCYdTIfNic,8
-docling-2.46.0.dist-info/RECORD,,
+docling-2.47.0.dist-info/licenses/LICENSE,sha256=mBb7ErEcM8VS9OhiGHnQ2kk75HwPhr54W1Oiz3965MY,1088
+docling-2.47.0.dist-info/METADATA,sha256=4cBB6CG8LFZzsDFRenI4f09ypOTCswZk2mVwczxEcVs,10569
+docling-2.47.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+docling-2.47.0.dist-info/entry_points.txt,sha256=hzVlbeE0aMSTQ9S0-NTYN0Hmgsn6qL_EA2qX4UbkAuY,149
+docling-2.47.0.dist-info/top_level.txt,sha256=vkIywP-USjFyYo1AIRQbWQQaL3xB5jf8vkCYdTIfNic,8
+docling-2.47.0.dist-info/RECORD,,