style: Added isort ordering & dedicated CI check (mindee#577)

* chore: Added isort config

* ci: Added isort CI job

* docs: Updated CONTRIBUTING

* style: Reordered imports

* refactor: Moved doctr/models/data_utils to doctr/utils/data

* chore: Fixed circular import

.github/workflows/style.yml

@@ -26,6 +26,24 @@ jobs:
           flake8 --version
           flake8 ./
 
+  isort-py3:
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix:
+        os: [ubuntu-latest]
+        python: [3.7]
+    steps:
+      - uses: actions/checkout@v2
+      - name: Set up Python
+        uses: actions/setup-python@v1
+        with:
+          python-version: ${{ matrix.python }}
+          architecture: x64
+      - name: Run isort
+        run: |
+          pip install isort
+          isort **/*.py -c -v
+
   mypy-py3:
     runs-on: ${{ matrix.os }}
     strategy:

CONTRIBUTING.md

@@ -74,6 +74,15 @@ flake8 ./
 ```
 This will read the `.flake8` setting file and let you know whether your commits need some adjustments.
 
+#### Import order
+
+In order to ensure there is a common import order convention, run [isort](https://github.com/PyCQA/isort) as follows:
+
+```shell
+isort **/*.py
+```
+This will reorder the imports of your local files.
+
 #### Annotation typing
 
 Additionally, to catch type-related issues and have a cleaner codebase, annotation typing are expected. After installing [mypy](https://github.com/python/mypy), you can run the verifications as follows:

api/app/config.py

@@ -4,8 +4,8 @@
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
 import os
-import doctr
 
+import doctr
 
 PROJECT_NAME: str = 'docTR API template'
 PROJECT_DESCRIPTION: str = 'Template API for Optical Character Recognition'

api/app/main.py

@@ -4,12 +4,11 @@
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
 import time
-from fastapi import FastAPI, Request
-from fastapi.openapi.utils import get_openapi
 
 from app import config as cfg
-from app.routes import recognition, detection, ocr
-
+from app.routes import detection, ocr, recognition
+from fastapi import FastAPI, Request
+from fastapi.openapi.utils import get_openapi
 
 app = FastAPI(title=cfg.PROJECT_NAME, description=cfg.PROJECT_DESCRIPTION, debug=cfg.DEBUG, version=cfg.VERSION)
 

api/app/routes/detection.py

@@ -3,13 +3,13 @@
 # This program is licensed under the Apache License version 2.
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
-from fastapi import APIRouter, UploadFile, File
 from typing import List
 
-from doctr.io import decode_img_as_tensor
-from app.vision import det_predictor
 from app.schemas import DetectionOut
+from app.vision import det_predictor
+from fastapi import APIRouter, File, UploadFile
 
+from doctr.io import decode_img_as_tensor
 
 router = APIRouter()
 

api/app/routes/ocr.py

@@ -3,13 +3,13 @@
 # This program is licensed under the Apache License version 2.
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
-from fastapi import APIRouter, UploadFile, File
 from typing import List
 
-from doctr.io import decode_img_as_tensor
-from app.vision import predictor
 from app.schemas import OCROut
+from app.vision import predictor
+from fastapi import APIRouter, File, UploadFile
 
+from doctr.io import decode_img_as_tensor
 
 router = APIRouter()
 

api/app/routes/recognition.py

@@ -3,12 +3,11 @@
 # This program is licensed under the Apache License version 2.
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
-from fastapi import APIRouter, UploadFile, File
-
-from doctr.io import decode_img_as_tensor
-from app.vision import reco_predictor
 from app.schemas import RecognitionOut
+from app.vision import reco_predictor
+from fastapi import APIRouter, File, UploadFile
 
+from doctr.io import decode_img_as_tensor
 
 router = APIRouter()
 

api/app/schemas.py

@@ -3,9 +3,10 @@
 # This program is licensed under the Apache License version 2.
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
-from pydantic import BaseModel, Field
 from typing import Tuple
 
+from pydantic import BaseModel, Field
+
 
 # Recognition output
 class RecognitionOut(BaseModel):

api/app/vision.py

@@ -11,7 +11,6 @@
 
 from doctr.models import ocr_predictor
 
-
 predictor = ocr_predictor(pretrained=True)
 det_predictor = predictor.det_predictor
 reco_predictor = predictor.reco_predictor

api/tests/conftest.py

@@ -5,9 +5,8 @@
 
 import pytest
 import requests
-from httpx import AsyncClient
-
 from app.main import app
+from httpx import AsyncClient
 
 
 @pytest.fixture(scope="session")

api/tests/routes/test_detection.py

@@ -3,9 +3,10 @@
 # This program is licensed under the Apache License version 2.
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
-import pytest
 import numpy as np
+import pytest
 from scipy.optimize import linear_sum_assignment
+
 from doctr.utils.metrics import box_iou
 
 

api/tests/routes/test_ocr.py

@@ -3,9 +3,10 @@
 # This program is licensed under the Apache License version 2.
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
-import pytest
 import numpy as np
+import pytest
 from scipy.optimize import linear_sum_assignment
+
 from doctr.utils.metrics import box_iou
 
 

demo/app.py

@@ -4,13 +4,14 @@
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
 import os
-import streamlit as st
+
 import matplotlib.pyplot as plt
+import streamlit as st
 
 os.environ["TF_CPP_MIN_LOG_LEVEL"] = "2"
 
-import tensorflow as tf
 import cv2
+import tensorflow as tf
 
 gpu_devices = tf.config.experimental.list_physical_devices('GPU')
 if any(gpu_devices):

docs/source/conf.py

@@ -6,15 +6,16 @@
 
 # -- Path setup --------------------------------------------------------------
 
-import sphinx_rtd_theme
-
 # If extensions (or modules to document with autodoc) are in another directory,
 # add these directories to sys.path here. If the directory is relative to the
 # documentation root, use os.path.abspath to make it absolute, like shown here.
 #
 import os
 import sys
 from datetime import datetime
+
+import sphinx_rtd_theme
+
 sys.path.insert(0, os.path.abspath('../..'))
 import doctr
 

doctr/__init__.py

@@ -1,7 +1,3 @@
+from . import datasets, io, models, transforms, utils
 from .file_utils import is_tf_available, is_torch_available
 from .version import __version__  # noqa: F401
-from . import datasets
-from . import io
-from . import transforms
-from . import utils
-from . import models

doctr/datasets/__init__.py

@@ -1,14 +1,14 @@
-from .utils import *
-from .vocabs import *
-from .funsd import *
+from doctr.file_utils import is_tf_available
+
+from .classification import *
 from .cord import *
 from .detection import *
-from .recognition import *
+from .funsd import *
 from .ocr import *
+from .recognition import *
 from .sroie import *
-from .classification import *
-
-from doctr.file_utils import is_tf_available
+from .utils import *
+from .vocabs import *
 
 if is_tf_available():
     from .loader import *

doctr/datasets/classification/base.py

@@ -3,13 +3,15 @@
 # This program is licensed under the Apache License version 2.
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
-from PIL import ImageDraw, Image, ImageFont
-from typing import Any, Tuple, Optional, Callable, List
 import logging
 import platform
+from typing import Any, Callable, List, Optional, Tuple
+
+from PIL import Image, ImageDraw, ImageFont
 
 from doctr.io.image import tensor_from_pil
 from doctr.utils.fonts import get_font
+
 from ..datasets import AbstractDataset
 
 

doctr/datasets/classification/pytorch.py

@@ -4,8 +4,8 @@
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
 from torch.utils.data._utils.collate import default_collate
-from .base import _CharacterGenerator
 
+from .base import _CharacterGenerator
 
 __all__ = ['CharacterGenerator']
 

doctr/datasets/classification/tensorflow.py

@@ -4,8 +4,8 @@
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
 import tensorflow as tf
-from .base import _CharacterGenerator
 
+from .base import _CharacterGenerator
 
 __all__ = ['CharacterGenerator']
 

doctr/datasets/cord.py

@@ -3,15 +3,17 @@
 # This program is licensed under the Apache License version 2.
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
-import os
 import json
-import numpy as np
+import os
 from pathlib import Path
-from typing import List, Dict, Any, Tuple, Optional, Callable
+from typing import Any, Callable, Dict, List, Optional, Tuple
+
+import numpy as np
 
-from .datasets import VisionDataset
 from doctr.utils.geometry import fit_rbbox
 
+from .datasets import VisionDataset
+
 __all__ = ['CORD']
 
 

doctr/datasets/datasets/base.py

@@ -5,11 +5,10 @@
 
 import os
 from pathlib import Path
+from typing import Any, Callable, List, Optional, Tuple, Union
 from zipfile import ZipFile
-from typing import List, Any, Optional, Tuple, Callable, Union
-
-from doctr.models.data_utils import download_from_url
 
+from doctr.utils.data import download_from_url
 
 __all__ = ['_AbstractDataset', '_VisionDataset']
 

doctr/datasets/datasets/pytorch.py

@@ -4,12 +4,13 @@
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
 import os
-from typing import List, Any, Tuple
+from typing import Any, List, Tuple
+
 import torch
 
 from doctr.io import read_img_as_tensor
-from .base import _AbstractDataset, _VisionDataset
 
+from .base import _AbstractDataset, _VisionDataset
 
 __all__ = ['AbstractDataset', 'VisionDataset']
 

doctr/datasets/datasets/tensorflow.py

@@ -4,12 +4,13 @@
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
 import os
-from typing import List, Any, Tuple
+from typing import Any, List, Tuple
+
 import tensorflow as tf
 
 from doctr.io import read_img_as_tensor
-from .base import _AbstractDataset, _VisionDataset
 
+from .base import _AbstractDataset, _VisionDataset
 
 __all__ = ['AbstractDataset', 'VisionDataset']
 

doctr/datasets/detection.py

@@ -3,14 +3,16 @@
 # This program is licensed under the Apache License version 2.
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
-import os
 import json
+import os
+from typing import Any, Callable, Dict, List, Optional, Tuple
+
 import numpy as np
-from typing import List, Tuple, Dict, Any, Optional, Callable
 
-from .datasets import AbstractDataset
 from doctr.utils.geometry import fit_rbbox
 
+from .datasets import AbstractDataset
+
 __all__ = ["DetectionDataset"]
 
 

doctr/datasets/funsd.py

@@ -3,11 +3,12 @@
 # This program is licensed under the Apache License version 2.
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
-import os
 import json
-import numpy as np
+import os
 from pathlib import Path
-from typing import List, Dict, Any, Tuple, Optional, Callable
+from typing import Any, Callable, Dict, List, Optional, Tuple
+
+import numpy as np
 
 from .datasets import VisionDataset
 

doctr/datasets/loader.py

@@ -4,9 +4,10 @@
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
 import math
-import tensorflow as tf
+from typing import Callable, Optional
+
 import numpy as np
-from typing import Optional, Callable
+import tensorflow as tf
 
 from doctr.utils.multithreading import multithread_exec