ruff lint

Author: jonchang

OCR/ocr/api.py

@@ -1,4 +1,5 @@
 """Module for FastAPI interface functions."""
+
 import base64
 
 import uvicorn

OCR/ocr/services/alignment/backends/four_point_transform.py

@@ -16,6 +16,7 @@ class FourPointTransform:
     Attributes:
         image (np.ndarray): The input image as a NumPy array.
     """
+
     def __init__(self, image: Path | np.ndarray):
         """Initializes the FourPointTransform object with an image.
 

OCR/ocr/services/alignment/backends/image_homography.py

@@ -1,4 +1,5 @@
 """Aligns two images using image homography algorithms."""
+
 from pathlib import Path
 
 import numpy as np
@@ -17,6 +18,7 @@ class ImageHomography:
         match_ratio (float): The ratio used for Lowe's ratio test to filter good matches.
         _sift (cv.SIFT): The SIFT detector used to find keypoints and descriptors.
     """
+
     def __init__(self, template: Path | np.ndarray, match_ratio=0.3):
         """Initializes the ImageHomography object with a template image.
 

OCR/ocr/services/alignment/backends/random_perspective_transform.py

@@ -9,7 +9,7 @@
 
 class RandomPerspectiveTransform:
     """Class to generates a random perspective transform based on a template image.
-    
+
     This class allows you to apply random distortions to an image by computing
     a perspective transformation matrix and warping the image accordingly.
 
@@ -85,7 +85,7 @@ def transform(self, transformer: np.ndarray) -> np.ndarray:
     def random_transform(self, distortion_scale: float) -> np.ndarray:
         """Warp the template image with specified distortion_scale.
 
-        This method internally calls `make_transform` to generate the transformation matrix 
+        This method internally calls `make_transform` to generate the transformation matrix
         and applies it using `transform`.
 
         Args:

OCR/ocr/services/alignment/image_alignment.py

@@ -1,4 +1,5 @@
 """Module for aligning images using a specified image alignment backend."""
+
 import numpy as np
 
 from ocr.services.alignment.backends import ImageHomography
@@ -8,9 +9,10 @@ class ImageAligner:
     """Class for aligning images using a specified image alignment backend.
 
     Attributes:
-        aligner: An alignment backend class or instance that provides an `align` method. 
+        aligner: An alignment backend class or instance that provides an `align` method.
                  Default is the ImageHomography backend from the ocr.services.alignment module.
     """
+
     def __init__(self, aligner=ImageHomography):
         """Initializes an ImageAligner instance with the specified image alignment backend.
 

OCR/ocr/services/batch_metrics.py

@@ -1,4 +1,5 @@
 """Module for batch processing OCR and ground truth files and calculating metrics."""
+
 from ocr.services.metrics_analysis import OCRMetrics
 import os
 import csv
@@ -16,6 +17,7 @@ class BatchMetricsAnalysis:
         ground_truth_folder (str): Path to the folder containing ground truth files.
         csv_output_folder (str): Path to the folder where CSV output files will be saved.
     """
+
     def __init__(self, ocr_folder: str, ground_truth_folder: str, csv_output_folder: str) -> None:
         """Initializes the BatchMetricsAnalysis class with paths to OCR, ground truth, and output folders.
 

OCR/ocr/services/batch_segmentation.py

@@ -1,4 +1,5 @@
 """Process and segment a batch of images and perform OCR on the results."""
+
 import os
 import json
 import time
@@ -18,7 +19,10 @@ class BatchSegmentationOCR:
         output_folder (str): Path to the folder where OCR results and timing information will be saved.
         model (ImageOCR): An optional pre-defined OCR model; if None, a default instance of ImageOCR is used.
     """
-    def __init__(self, image_folder: str, segmentation_template: str, labels_path: str, output_folder: str, model=None) -> None:
+
+    def __init__(
+        self, image_folder: str, segmentation_template: str, labels_path: str, output_folder: str, model=None
+    ) -> None:
         """Initializes the BatchSegmentationOCR instance with the specified paths and an optional OCR model.
 
         Args:
@@ -76,7 +80,9 @@ def process_images(self) -> list[dict]:
         print("Processing complete.")
         return results
 
-    def segment_ocr_image(self, segmenter: ImageSegmenter, ocr, image_path: str, image_file: str) -> tuple[dict[str, tuple[str, float]], float]:
+    def segment_ocr_image(
+        self, segmenter: ImageSegmenter, ocr, image_path: str, image_file: str
+    ) -> tuple[dict[str, tuple[str, float]], float]:
         """Segments the image and runs OCR, returning results and time taken.
 
         Args:

OCR/ocr/services/image_ocr.py

@@ -1,4 +1,5 @@
 """Module for OCR using a transformers-based OCR model."""
+
 from collections.abc import Iterator
 
 from transformers import TrOCRProcessor, VisionEncoderDecoderModel
@@ -16,6 +17,7 @@ class ImageOCR:
         processor (TrOCRProcessor): Processor for TrOCR model that prepares images for OCR.
         model (VisionEncoderDecoderModel): Pre-trained TrOCR model for extracting text from images.
     """
+
     def __init__(self, model="microsoft/trocr-large-printed"):
         """Initializes the ImageOCR class with the specified OCR model.
 
@@ -32,7 +34,7 @@ def __init__(self, model="microsoft/trocr-large-printed"):
     def compute_line_angle(lines: list) -> Iterator[float]:
         """Computes the angle in degrees of the lines detected by the Hough transform, based on their endpoints.
 
-        This method processes the output of `cv.HoughLinesP` (lines in (x1, y1, x2, y2) format) and computes the angle 
+        This method processes the output of `cv.HoughLinesP` (lines in (x1, y1, x2, y2) format) and computes the angle
         between each line and the horizontal axis.
 
         Args:
@@ -186,12 +188,12 @@ def image_to_text(self, segments: dict[str, np.ndarray]) -> dict[str, tuple[str,
         For each segment, it extracts the text and the average confidence score.
 
         Args:
-            segments (dict[str, np.ndarray]): A dictionary where keys are segment labels (e.g., 'header', 'body'), 
+            segments (dict[str, np.ndarray]): A dictionary where keys are segment labels (e.g., 'header', 'body'),
                                               and values are NumPy arrays representing the corresponding image segments.
 
         Returns:
-            dict[str, tuple[str, float]]: A dictionary where each key corresponds to a segment label, and each value is 
-                                          a tuple containing the recognized text (as a string) and the confidence score 
+            dict[str, tuple[str, float]]: A dictionary where each key corresponds to a segment label, and each value is
+                                          a tuple containing the recognized text (as a string) and the confidence score
                                           (as a float) for the recognition.
         """
         digitized: dict[str, tuple[str, float]] = {}

OCR/ocr/services/image_segmenter.py

@@ -1,4 +1,5 @@
 """Module to segment images based on a segmentation template and a set of labels."""
+
 import cv2 as cv
 import numpy as np
 import json
@@ -31,7 +32,9 @@ def crop_zeros(image: np.ndarray) -> np.ndarray:
     ]  # inclusive
 
 
-def segment_by_mask_then_crop(raw_image: np.ndarray, segmentation_template: np.ndarray, labels: list[dict[str, str]], debug: bool) -> dict[str, np.ndarray]:
+def segment_by_mask_then_crop(
+    raw_image: np.ndarray, segmentation_template: np.ndarray, labels: list[dict[str, str]], debug: bool
+) -> dict[str, np.ndarray]:
     """Segments a raw image based on a color mask in the segmentation template, and then crops the resulting regions to remove zero (black) areas.
 
     Args:
@@ -75,7 +78,9 @@ def segment_by_mask_then_crop(raw_image: np.ndarray, segmentation_template: np.n
     return segments
 
 
-def segment_by_color_bounding_box(raw_image: np.ndarray, segmentation_template: np.ndarray, labels: list[dict[str, str]], debug: bool) -> dict[str, np.ndarray]:
+def segment_by_color_bounding_box(
+    raw_image: np.ndarray, segmentation_template: np.ndarray, labels: list[dict[str, str]], debug: bool
+) -> dict[str, np.ndarray]:
     """Segments a raw image by detecting colored boundary boxes in the segmentation template.
 
     Args:
@@ -119,6 +124,7 @@ class ImageSegmenter:
                                           `segment_by_mask_then_crop` or `segment_by_color_bounding_box`.
         debug (bool): If `True`, saves debug images and prints additional information.
     """
+
     def __init__(
         self,
         segmentation_function=segment_by_mask_then_crop,
@@ -152,7 +158,9 @@ def segment(
         """
         return self.segmentation_function(raw_image, segmentation_template, labels, self.debug)
 
-    def load_and_segment(self, raw_image_path: str, segmentation_template_path: str, labels_path: str) -> dict[str, np.ndarray]:
+    def load_and_segment(
+        self, raw_image_path: str, segmentation_template_path: str, labels_path: str
+    ) -> dict[str, np.ndarray]:
         """Loads image files and labels from specified paths, and then segments the image.
 
         Args:

OCR/ocr/services/metrics_analysis.py

@@ -1,4 +1,5 @@
 """Module to calculate OCR metrics and compare results to ground truth data."""
+
 import json
 import csv
 import Levenshtein
@@ -15,6 +16,7 @@ class OCRMetrics:
         ocr_json (dict): A dict from JSON data containing OCR results.
         ground_truth_json (dict): A dict from JSON data containing the ground truth values.
     """
+
     def __init__(self, ocr_json_path=None, ground_truth_json_path=None, ocr_json=None, ground_truth_json=None):
         """Initializes the OCRMetrics object with OCR and ground truth data, either loaded from files or provided as dictionaries.
 

OCR/ocr/services/tesseract_ocr.py

@@ -22,6 +22,7 @@ class TesseractOCR:
     * https://github.com/sirfz/tesserocr/blob/bbe0fb8edabdcc990f1e6fa9334c0747c2ac76ee/tesserocr/__init__.pyi#L47
     * https://tesseract-ocr.github.io/tessdoc/tess3/ControlParams.html
     """
+
     def __init__(self, psm=PSM.AUTO, variables=dict()):
         """Initializes the TesseractOCR object with the specified page segmentation mode and internal variables.
 
@@ -88,12 +89,12 @@ def image_to_text(self, segments: dict[str, np.ndarray]) -> dict[str, tuple[str,
         For each segment, it extracts the text and the average confidence score returned from the Tesseract API.
 
         Args:
-            segments (dict[str, np.ndarray]): A dictionary where keys are segment labels (e.g., 'header', 'body'), 
+            segments (dict[str, np.ndarray]): A dictionary where keys are segment labels (e.g., 'header', 'body'),
                                               and values are NumPy arrays representing the corresponding image segments.
 
         Returns:
-            dict[str, tuple[str, float]]: A dictionary where each key corresponds to a segment label, and each value is 
-                                          a tuple containing the recognized text (as a string) and the confidence score 
+            dict[str, tuple[str, float]]: A dictionary where each key corresponds to a segment label, and each value is
+                                          a tuple containing the recognized text (as a string) and the confidence score
                                           (as a float) for the recognition.
         """
         digitized: dict[str, tuple[str, float]] = {}