doctr-synth-generator 0.2.0__tar.gz → 0.2.1__tar.gz

{doctr_synth_generator-0.2.0 → doctr_synth_generator-0.2.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: doctr-synth-generator
-Version: 0.2.0
+Version: 0.2.1
 Summary: A synthetic data generator for training OCR models
 Author-email: Felix Dittrich <felixdittrich92@gmail.com>
 Maintainer: Felix Dittrich
@@ -250,7 +250,7 @@ Dynamic: license-file
 ![Build Status](https://github.com/felixdittrich92/docTR-Synth-Generator/workflows/builds/badge.svg)
 [![codecov](https://codecov.io/gh/felixdittrich92/docTR-Synth-Generator/graph/badge.svg?token=31MDR20JGI)](https://codecov.io/gh/felixdittrich92/docTR-Synth-Generator)
 [![CodeFactor](https://www.codefactor.io/repository/github/felixdittrich92/doctr-synth-generator/badge)](https://www.codefactor.io/repository/github/felixdittrich92/doctr-synth-generator)
-[![Pypi](https://img.shields.io/badge/pypi-v0.1.0-blue.svg)](https://pypi.org/project/docTR-Synth-Generator/)
+[![Pypi](https://img.shields.io/badge/pypi-v0.2.1-blue.svg)](https://pypi.org/project/docTR-Synth-Generator/)
 
 # docTR-Synth-Generator
 A tool to generate synthetic OCR datasets - made for docTR

{doctr_synth_generator-0.2.0 → doctr_synth_generator-0.2.1}/README.md

@@ -2,7 +2,7 @@
 ![Build Status](https://github.com/felixdittrich92/docTR-Synth-Generator/workflows/builds/badge.svg)
 [![codecov](https://codecov.io/gh/felixdittrich92/docTR-Synth-Generator/graph/badge.svg?token=31MDR20JGI)](https://codecov.io/gh/felixdittrich92/docTR-Synth-Generator)
 [![CodeFactor](https://www.codefactor.io/repository/github/felixdittrich92/doctr-synth-generator/badge)](https://www.codefactor.io/repository/github/felixdittrich92/doctr-synth-generator)
-[![Pypi](https://img.shields.io/badge/pypi-v0.1.0-blue.svg)](https://pypi.org/project/docTR-Synth-Generator/)
+[![Pypi](https://img.shields.io/badge/pypi-v0.2.1-blue.svg)](https://pypi.org/project/docTR-Synth-Generator/)
 
 # docTR-Synth-Generator
 A tool to generate synthetic OCR datasets - made for docTR

{doctr_synth_generator-0.2.0 → doctr_synth_generator-0.2.1}/doctr_synth_generator.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: doctr-synth-generator
-Version: 0.2.0
+Version: 0.2.1
 Summary: A synthetic data generator for training OCR models
 Author-email: Felix Dittrich <felixdittrich92@gmail.com>
 Maintainer: Felix Dittrich
@@ -250,7 +250,7 @@ Dynamic: license-file
 ![Build Status](https://github.com/felixdittrich92/docTR-Synth-Generator/workflows/builds/badge.svg)
 [![codecov](https://codecov.io/gh/felixdittrich92/docTR-Synth-Generator/graph/badge.svg?token=31MDR20JGI)](https://codecov.io/gh/felixdittrich92/docTR-Synth-Generator)
 [![CodeFactor](https://www.codefactor.io/repository/github/felixdittrich92/doctr-synth-generator/badge)](https://www.codefactor.io/repository/github/felixdittrich92/doctr-synth-generator)
-[![Pypi](https://img.shields.io/badge/pypi-v0.1.0-blue.svg)](https://pypi.org/project/docTR-Synth-Generator/)
+[![Pypi](https://img.shields.io/badge/pypi-v0.2.1-blue.svg)](https://pypi.org/project/docTR-Synth-Generator/)
 
 # docTR-Synth-Generator
 A tool to generate synthetic OCR datasets - made for docTR

{doctr_synth_generator-0.2.0 → doctr_synth_generator-0.2.1}/generator/components/corpus_downloader.py

@@ -263,5 +263,5 @@ def generate_numeric_tokens(n: int, seed: int | None = None) -> list[str]:
         elif kind == "percent":
             tokens.append(f"{rng.randint(0, 100)}%")
         else:  # phone
-            tokens.append(f"+{rng.randint(1, 99)} {rng.randint(100, 999)} {rng.randint(100000, 9999999)}")
+            tokens.append(f"+{rng.randint(1, 99)}{rng.randint(100, 999)}{rng.randint(100000, 9999999)}")
     return tokens

{doctr_synth_generator-0.2.0 → doctr_synth_generator-0.2.1}/generator/doctr_dataset.py

@@ -20,11 +20,31 @@ from .dataset_generator import SyntheticDatasetGenerator
 
 try:  # docTR's single-class name ("words"); fall back to the same literal if absent.
     from doctr.file_utils import CLASS_NAME  # type: ignore[import-not-found]
+    from doctr.utils import Sample  # type: ignore[import-not-found]
 except Exception:  # pragma: no cover - docTR not installed
+    from dataclasses import dataclass
+
     CLASS_NAME = "words"
 
+    @dataclass
+    class Sample:  # type: ignore[no-redef]
+        """Canonical data container for all transforms."""
+
+        image: Any
+        mask: Any | None = None
+        target: np.ndarray | dict[str, np.ndarray] | None = None
+
+        def replace(self, **kwargs) -> "Sample":
+            return Sample(
+                image=kwargs.get("image", self.image),
+                mask=kwargs.get("mask", self.mask),
+                target=kwargs.get("target", self.target),
+            )
+
+
 __all__ = [
     "CLASS_NAME",
+    "Sample",
     "polygons_to_target",
     "render_recognition_sample",
     "render_detection_sample",
@@ -118,7 +138,7 @@ def _torch():
 def _pil_to_tensor(img: Image.Image):
     """PIL RGB -> ``CxHxW`` float32 tensor in ``[0, 1]`` (matches docTR's reader)."""
     torch = _torch()
-    arr = np.asarray(img.convert("RGB"), dtype=np.uint8)
+    arr = np.asarray(img.convert("RGB"), dtype=np.uint8, copy=True)
     return torch.from_numpy(arr).permute(2, 0, 1).contiguous().float().div_(255.0)
 
 
@@ -175,11 +195,13 @@ class _BaseSynthDataset:
         self._seed_sample(index)
         img, target = self._render(index)
         tensor = _pil_to_tensor(img)
+        sample = Sample(image=tensor, target=target)
         if self.img_transforms is not None:
-            tensor = self.img_transforms(tensor)
+            sample = self.img_transforms(sample)
         if self.sample_transforms is not None:
-            tensor, target = self.sample_transforms(tensor, target)
-        return tensor, target
+            sample = self.sample_transforms(sample)
+        # Keep compatibility with the existing collate_fn
+        return sample.image, sample.target
 
     @staticmethod
     def collate_fn(samples):

doctr_synth_generator-0.2.1/generator/version.py

@@ -0,0 +1 @@
+__version__ = 'v0.2.1'

{doctr_synth_generator-0.2.0 → doctr_synth_generator-0.2.1}/setup.py

@@ -9,7 +9,7 @@ from pathlib import Path
 from setuptools import setup
 
 PKG_NAME = "doctr-synth-generator"
-VERSION = os.getenv("BUILD_VERSION", "0.2.0a0")
+VERSION = os.getenv("BUILD_VERSION", "0.2.1a0")
 
 
 if __name__ == "__main__":

doctr_synth_generator-0.2.0/generator/version.py

@@ -1 +0,0 @@
-__version__ = 'v0.2.0'