[datasets] UnicodeEncode error for € when drawing with default PIL font.

[charlesmindee]

🐛 Bug

UnicodeEncode error for € when drawing with default PIL font.

To Reproduce

Steps to reproduce the behavior:

1. Set font=None in the train loader of references/classification/train_tensorflow.py
2. launch python references/classification/train_tensorflow.py mobilenet_v3_small --show-samples

Traceback (most recent call last):
  File "references/classification/train_tensorflow.py", line 245, in <module>
    main(args)
  File "references/classification/train_tensorflow.py", line 124, in main
    train_set = CharacterGenerator(
  File "/home/laptopmindee/doctr/doctr/datasets/classification/tensorflow.py", line 29, in __init__
    super().__init__(*args, **kwargs)
  File "/home/laptopmindee/doctr/doctr/datasets/classification/base.py", line 58, in __init__
    self._data = [synthesize_char_img(char, font_family=self.font_family) for char in self.vocab]
  File "/home/laptopmindee/doctr/doctr/datasets/classification/base.py", line 58, in <listcomp>
    self._data = [synthesize_char_img(char, font_family=self.font_family) for char in self.vocab]
  File "/home/laptopmindee/doctr/doctr/datasets/classification/base.py", line 36, in synthesize_char_img
    d.text((4, 0), char, font=font, fill=(255, 255, 255))
  File "/home/laptopmindee/venv3.8/lib/python3.8/site-packages/PIL/ImageDraw.py", line 469, in text
    draw_text(ink)
  File "/home/laptopmindee/venv3.8/lib/python3.8/site-packages/PIL/ImageDraw.py", line 429, in draw_text
    mask = font.getmask(
  File "/home/laptopmindee/venv3.8/lib/python3.8/site-packages/PIL/ImageFont.py", line 149, in getmask
    return self.font.getmask(text, mode)
UnicodeEncodeError: 'latin-1' codec can't encode character '\u20ac' in position 0: ordinal not in range(256)

Expected behavior

Display of samples

Environment

DocTR version: 0.3.1a0
TensorFlow version: 2.5.0
PyTorch version: 1.9.0+cu111 (torchvision 0.10.0+cu111)
OpenCV version: 4.5.1
OS: Ubuntu 18.04.5 LTS
Python version: 3.8
Is CUDA available (TensorFlow): Yes
Is CUDA available (PyTorch): Yes
CUDA runtime version: 11.4.100
GPU models and configuration: GPU 0: NVIDIA GeForce RTX 2060
Nvidia driver version: 470.57.02
cuDNN version: Probably one of the following:
/usr/lib/x86_64-linux-gnu/libcudnn.so.7.6.5
/usr/lib/x86_64-linux-gnu/libcudnn.so.8.2.2
/usr/lib/x86_64-linux-gnu/libcudnn_adv_infer.so.8.2.2
/usr/lib/x86_64-linux-gnu/libcudnn_adv_train.so.8.2.2
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_infer.so.8.2.2
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_train.so.8.2.2
/usr/lib/x86_64-linux-gnu/libcudnn_ops_infer.so.8.2.

[fg-mindee]

Correct, I think we should add OS automatic detection and take a basic font that is compatible (Arial for windows & Mac for instance, FreeMono for Linux)

[charlesmindee]

Closed by #418