feat: Support for Python 3.14 (#2530)

* fix dependencies for py314

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* add metadata and CI tests

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* add back gliner

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* update error message about python 3.14 availability

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* skip tests which cannot run on py 3.14

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* fix lint

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* remove vllm from py 3.14 deps

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* safe import for vllm

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* update lock

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* remove torch.compile()

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* update checkbox results after docling-core changes

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* cannot run mlx example in CI

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* add test for rapidocr backends and skip onnxruntime on py3.14

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* fix other occurances of torch.compile()

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* allow torch.compile for Python <3.14. proper support will be introduced with new torch releases

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

---------

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

tests/data/groundtruth/docling_v2/right_to_left_03.md

@@ -16,9 +16,9 @@
 
 استاندارد اجباری است؟
 
-بلی
+- [ ] بلی
 
-خير
+- [x] خير
 
 مرجع صادرکننده استاندارد
 
@@ -26,7 +26,7 @@
 
 آيا توليدکننده محصول، استاندارد مذکور را اخذ نموده است؟
 
-بلی        خير
+- [x] بلی        خير
 
 ## -3 پذيرش در بورس
 

tests/test_asr_pipeline.py

@@ -1,3 +1,4 @@
+import sys
 from pathlib import Path
 from unittest.mock import Mock, patch
 
@@ -10,6 +11,11 @@ from docling.datamodel.pipeline_options import AsrPipelineOptions
 from docling.document_converter import AudioFormatOption, DocumentConverter
 from docling.pipeline.asr_pipeline import AsrPipeline
 
+pytestmark = pytest.mark.skipif(
+    sys.version_info >= (3, 14),
+    reason="Python 3.14 is not yet supported by whisper dependencies.",
+)
+
 
 @pytest.fixture
 def test_audio_path():

tests/test_e2e_ocr_conversion.py

@@ -70,13 +70,19 @@ def test_e2e_conversions():
         (EasyOcrOptions(force_full_page_ocr=True), False),
     ]
 
-    # rapidocr is only available for Python >=3.6,<3.13
-    if sys.version_info < (3, 13):
-        engines.append((RapidOcrOptions(), False))
-        engines.append((RapidOcrOptions(force_full_page_ocr=True), False))
+    for rapidocr_backend in ["onnxruntime", "torch"]:
+        if sys.version_info >= (3, 14) and rapidocr_backend == "onnxruntime":
+            # skip onnxruntime backend on Python 3.14
+            continue
+
+        engines.append((RapidOcrOptions(backend=rapidocr_backend), False))
+        engines.append(
+            (RapidOcrOptions(backend=rapidocr_backend, force_full_page_ocr=True), False)
+        )
         engines.append(
             (
                 RapidOcrOptions(
+                    backend=rapidocr_backend,
                     force_full_page_ocr=True,
                     rec_font_path="test",
                     rapidocr_params={"Rec.font_path": None},  # overwrites rec_font_path