owocr - pypi Package Compare versions

Comparing version 1.26.0 to 1.26.1

ndlocr_lite/deim.py

@@ -6,2 +6,3 @@ from PIL import Image, ImageDraw
 import numpy as np
+import cv2
 from typing import Tuple, List
@@ -57,15 +58,13 @@ import xml.etree.ElementTree as ET
         max_wh=max(img.shape[0],img.shape[1])
-        paddedimg=np.zeros((max_wh,max_wh,3)).astype(np.uint8)
-        paddedimg[:img.shape[0],:img.shape[1],:]=img.copy()
-        pil_image = Image.fromarray(paddedimg)
-        self.image_width,self.image_height = pil_image.size
-        pil_resized = pil_image.resize((self.input_width, self.input_height))
-        resized=np.array(pil_resized)
-        #resized=resized[:,:,::-1]
-        
-        # Scale input pixel value to 0 to 1
-        resized = resized / 255.0
+        paddedimg=np.zeros((max_wh,max_wh,3),dtype=np.uint8)
+        paddedimg[:img.shape[0],:img.shape[1],:]=img
+        self.image_width=max_wh
+        self.image_height=max_wh
+        resized=cv2.resize(paddedimg,(self.input_width, self.input_height),interpolation=cv2.INTER_CUBIC)
+        input_image=resized.astype(np.float32)
+        input_image/=255.0
         mean = np.array([0.485, 0.456, 0.406], dtype=np.float32)
         std = np.array([0.229, 0.224, 0.225], dtype=np.float32)
-        input_image= (resized-mean) / std
+        input_image-=mean
+        input_image/=std
         input_image = input_image.transpose(2,0,1)
@@ -72,0 +71,0 @@ input_tensor = input_image[np.newaxis, :, :, :].astype(np.float32)

ndlocr_lite/ocr.py

@@ -26,3 +26,3 @@ import sys
         self.pred_str = pred_str
-    def __lt__(self, other):  
+    def __lt__(self, other):
         return self.idx < other.idx
@@ -117,7 +117,7 @@
         os.makedirs(outputpath,exist_ok=True)
-        output_path = os.path.join(outputpath,f"viz_{Path(inputname).name}")
-        if output_path.split(".")[-1]=="jp2":
-            output_path=output_path.split(".")[:-4]+".jpg"
-        print(f"[INFO] Saving result on {output_path}")
-        pil_image.save(output_path)
+        output_filepath = os.path.join(outputpath,f"viz_{Path(inputname).name}")
+        if output_filepath.split(".")[-1]=="jp2":
+            output_filepath=output_filepath[:-4]+".jpg"
+        print(f"[INFO] Saving result on {output_filepath}")
+        pil_image.save(output_filepath)
     return detections,classeslist
@@ -132,7 +132,7 @@
         os.makedirs(outputpath,exist_ok=True)
-        output_path = os.path.join(outputpath,f"viz_{Path(inputname).name}")
-        if output_path.split(".")[-1]=="jp2":
-            output_path=output_path.split(".")[:-4]+".jpg"
-        print(f"[INFO] Saving result on {output_path}")
-        pil_image.save(output_path)
+        output_filepath = os.path.join(outputpath,f"viz_{Path(inputname).name}")
+        if output_filepath.split(".")[-1]=="jp2":
+            output_filepath=output_filepath[:-4]+".jpg"
+        print(f"[INFO] Saving result on {output_filepath}")
+        pil_image.save(output_filepath)
     return detections,classeslist
@@ -151,3 +151,3 @@
         ext=inputpath.split(".")[-1]
-        if ext in ["jpg","png","tiff","jp2","tif","jpeg","bmp"]:
+        if ext.lower() in ["jpg","png","tiff","jp2","tif","jpeg","bmp"]:
             inputpathlist.append(inputpath)
@@ -161,2 +161,3 @@ if len(inputpathlist)==0:
     
+    detector=get_detector(args)
     recognizer100=get_recognizer(args=args)
@@ -177,3 +178,3 @@ recognizer30=get_recognizer(args=args,weights_path=args.rec_weights30)
         img_h,img_w=img.shape[:2]
-        detections,classeslist=inference_on_detector(args=args,inputname=imgname,npimage=img,outputpath=args.output,issaveimg=args.viz)
+        detections,classeslist=process_detector(detector,inputname=imgname,npimage=img,outputpath=args.output,issaveimg=args.viz)
         e1=time.time()
@@ -262,5 +263,6 @@ resultobj=[dict(),dict()]
             alltextlist=alltextlist[::-1]
-        with open(os.path.join(args.output,os.path.basename(inputpath).split(".")[0]+".xml"),"w",encoding="utf-8") as wf:
+        output_stem = os.path.splitext(os.path.basename(inputpath))[0]
+        with open(os.path.join(args.output,output_stem+".xml"),"w",encoding="utf-8") as wf:
             wf.write(allxmlstr)
-        with open(os.path.join(args.output,os.path.basename(inputpath).split(".")[0]+".json"),"w",encoding="utf-8") as wf:
+        with open(os.path.join(args.output,output_stem+".json"),"w",encoding="utf-8") as wf:
             alljsonobj={
@@ -277,3 +279,3 @@ "contents":[resjsonarray],
             wf.write(alljsonstr)
-        with open(os.path.join(args.output,os.path.basename(inputpath).split(".")[0]+".txt"),"w",encoding="utf-8") as wtf:
+        with open(os.path.join(args.output,output_stem+".txt"),"w",encoding="utf-8") as wtf:
             wtf.write("\n".join(alltextlist))
@@ -280,0 +282,0 @@ print("Total calculation time (Detection + Recognition):",time.time()-start)

ndlocr_lite/parseq.py

@@ -6,2 +6,3 @@ from PIL import Image
 import numpy as np
+import cv2
 from typing import Tuple, List
@@ -51,21 +52,18 @@
     def preprocess(self, img: np.ndarray) -> np.ndarray:
-        pil_image = Image.fromarray(img)
-        if pil_image.height>pil_image.width:
-            pil_image =pil_image.transpose(Image.ROTATE_90)
-        pil_resized = pil_image.resize((self.input_width, self.input_height))
-        
-        resized = np.array(pil_resized, dtype=np.float32)
-        resized = resized[:,:,::-1]
-        input_image = resized / 255.0
-        input_image = 2.0*(input_image-0.5)
+        h,w=img.shape[:2]
+        if h>w:
+            img=cv2.rotate(img,cv2.ROTATE_90_COUNTERCLOCKWISE)
+        resized=cv2.resize(img,(self.input_width, self.input_height),interpolation=cv2.INTER_LINEAR)
+        input_image=np.ascontiguousarray(resized[:,:,::-1]).astype(np.float32)
+        input_image/=127.5
+        input_image-=1.0
         input_image = input_image.transpose(2,0,1)
-        input_tensor = input_image[np.newaxis, :, :, :].astype(np.float32)
-        return input_tensor
+        return input_image[np.newaxis, :, :, :]
     
     def read(self, img: np.ndarray) -> List:
-        if img is None:
-            return None
+        if img is None or img.size == 0:
+            return ""
         input_tensor = self.preprocess(img)
         outputs = self.session.run(self.output_names, {self.input_names[0]: input_tensor})[0]
-        indices = np.argmax(outputs, axis=2)[0]
+        indices = np.argmax(outputs[0], axis=1)
         stop_idx = np.where(indices == 0)[0]
@@ -72,0 +70,0 @@ end_pos = stop_idx[0] if stop_idx.size > 0 else len(indices)

owocr.egg-info/PKG-INFO

 Metadata-Version: 2.4
 Name: owocr
-Version: 1.26.0
-Summary: Optical character recognition for Japanese text
+Version: 1.26.1
+Summary: Multi-service, multi-platform optical character recognition
 Author-email: AuroraWright <[e-mail]>
@@ -19,2 +19,3 @@ License-Expression: GPL-3.0-only
 Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
 Classifier: Programming Language :: Python :: 3
@@ -77,3 +78,3 @@ Classifier: Programming Language :: Python
 Provides-Extra: azure
-Requires-Dist: azure-ai-vision-imageanalysis; extra == "azure"
+Requires-Dist: azure-ai-documentintelligence; extra == "azure"
 Provides-Extra: ndlocrlite
@@ -85,2 +86,3 @@ Requires-Dist: lxml; extra == "ndlocrlite"
 Requires-Dist: tqdm; extra == "ndlocrlite"
+Requires-Dist: opencv-python-headless; extra == "ndlocrlite"
 Dynamic: license-file
@@ -117,3 +119,3 @@
 
-OwOCR has been tested on Python 3.11, 3.12 and 3.13. It can be installed with `pip install owocr` after you install Python. You also need to have one or more OCR engines, check the list below for instructions. I recommend installing at least Google Lens on any operating system, and OneOCR if you are on Windows. Bing is pre-installed, Apple Vision and Live Text come pre-installed on macOS.
+OwOCR has been tested on Python 3.11 to 3.14. It can be installed with `pip install owocr` after you install Python. You also need to have one or more OCR engines, check the list below for instructions. I recommend installing at least Google Lens on any operating system, and OneOCR if you are on Windows. Bing is pre-installed, Apple Vision and Live Text come pre-installed on macOS.
 
@@ -190,3 +192,3 @@ ## Usage (terminal)
 - Google Vision - You need a service account .json file named google_vision.json in `user directory/.config/` → Terminal: install with `pip install "owocr[gvision]"`, key: `g`
-- Azure Image Analysis - You need to specify an api key and an endpoint in the config file → Terminal: install with `pip install "owocr[azure]"`, key: `v`
+- Azure Document Intelligence - You need to specify an api key and an endpoint in the config file → Terminal: install with `pip install "owocr[azure]"`, key: `v`
 - OCRSpace - You need to specify an api key in the config file. → Terminal key: `o`
@@ -193,0 +195,0 @@

owocr.egg-info/requires.txt

@@ -27,3 +27,3 @@ jaconv
 [azure]
-azure-ai-vision-imageanalysis
+azure-ai-documentintelligence
 
@@ -61,2 +61,3 @@ [easyocr]
 tqdm
+opencv-python-headless
 
@@ -63,0 +64,0 @@ [oneocr]

owocr/__init__.py

@@ -1,2 +0,2 @@
-__version__ = (1, 26, 0)
+__version__ = (1, 26, 1)
 __version_string__ = '.'.join(map(str, __version__))

owocr/config_editor.py

@@ -210,3 +210,3 @@ import configparser
 
-        width, height = 700, 700
+        width, height = 750, 750
         window_scale = 1.0
@@ -249,2 +249,3 @@ if sys.platform == 'win32':
                 ('delete_images', 'bool', 'Delete images from the folder after processing'),
+                ('skip_existing_images', 'bool', 'Ignore images already in the folder when owocr starts'),
                 ('pause_at_startup', 'bool', 'Pause when owocr starts'),
@@ -747,19 +748,9 @@ ('notifications', 'bool', 'Show OS notifications with the detected text'),
     def _open_picker(self, option, var):
-        if option in ['read_from', 'read_from_secondary']:
-            folder_path = filedialog.askdirectory(
-                title=f'Select directory for {option}',
-                mustexist=True
-            )
-            if folder_path:
-                var.set(folder_path)
-                self._update_general_state()
-        elif option == 'write_to':
-            file_path = filedialog.asksaveasfilename(
-                title='Select output text file',
-                defaultextension='.txt',
-                filetypes=[('Text files', '*.txt'), ('All files', '*.*')]
-            )
-            if file_path:
-                var.set(file_path)
-                self._update_general_state()
+        folder_path = filedialog.askdirectory(
+            title=f'Select directory for {option}',
+            mustexist=True
+        )
+        if folder_path:
+            var.set(folder_path)
+            self._update_general_state()
 
@@ -802,6 +793,8 @@ def _update_ui_state(self):
 
-        widget_info = self.widgets['delete_images']
-        frame = widget_info.get('frame')
-        show_delete = self._is_folder_file_selected('read_from') or self._is_folder_file_selected('read_from_secondary')
-        frame.grid() if show_delete else frame.grid_remove()
+        show_folder_options = self._is_folder_file_selected('read_from') or self._is_folder_file_selected('read_from_secondary')
+        folder_options = ['delete_images', 'skip_existing_images']
+        for option in folder_options: 
+            widget_info = self.widgets[option]
+            frame = widget_info.get('frame')
+            frame.grid() if show_folder_options else frame.grid_remove()
 
@@ -808,0 +801,0 @@ dropdown_options = ['read_from', 'read_from_secondary', 'write_to']

owocr/config.py

@@ -18,3 +18,3 @@ import os
     It can read images copied to the system clipboard or placed in a directory, images sent via a websocket or a Unix domain socket, or directly capture a screen (or a portion of it) or a window.
-    Recognized text can be either saved to system clipboard, appended to a text file or sent via a websocket.
+    Recognized text can be either saved to system clipboard, written to a text/json file or sent via a websocket.
 '''))
@@ -27,3 +27,3 @@
 parser.add_argument('-w', '--write_to', type=str, default=argparse.SUPPRESS,
-                    help='Where to save recognized texts to. Can be either "clipboard", "websocket", or a path to a text file.')
+                    help='Where to save recognized texts to. Can be either "clipboard", "websocket", or a path to a directory to write text/json files to.')
 parser.add_argument('-e', '--engine', type=str, default=argparse.SUPPRESS,
@@ -37,2 +37,4 @@ help='OCR engine to use. Available: "mangaocr", "mangaocrs", "glens", "bing", "gvision", "screenai", "avision", "alivetext", "azure", "winrtocr", "oneocr", "ndlocrlite", "easyocr", "rapidocr", "ocrspace".')
                     help='Delete image files after processing when reading from a directory.')
+parser.add_argument('-se', '--skip_existing_images', type=str2bool, nargs='?', const=True, default=argparse.SUPPRESS,
+                    help='Skip existing images at startup when reading from a directory.')
 parser.add_argument('-n', '--notifications', type=str2bool, nargs='?', const=True, default=argparse.SUPPRESS,
@@ -115,2 +117,3 @@ help='Show an operating system notification with the detected text. Will be ignored when reading with screen capture and periodic screenshots.')
         'delete_images': False,
+        'skip_existing_images': False,
         'engines': [],
@@ -117,0 +120,0 @@ 'delay_seconds': 0.5,

PKG-INFO

 Metadata-Version: 2.4
 Name: owocr
-Version: 1.26.0
-Summary: Optical character recognition for Japanese text
+Version: 1.26.1
+Summary: Multi-service, multi-platform optical character recognition
 Author-email: AuroraWright <[e-mail]>
@@ -19,2 +19,3 @@ License-Expression: GPL-3.0-only
 Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
 Classifier: Programming Language :: Python :: 3
@@ -77,3 +78,3 @@ Classifier: Programming Language :: Python
 Provides-Extra: azure
-Requires-Dist: azure-ai-vision-imageanalysis; extra == "azure"
+Requires-Dist: azure-ai-documentintelligence; extra == "azure"
 Provides-Extra: ndlocrlite
@@ -85,2 +86,3 @@ Requires-Dist: lxml; extra == "ndlocrlite"
 Requires-Dist: tqdm; extra == "ndlocrlite"
+Requires-Dist: opencv-python-headless; extra == "ndlocrlite"
 Dynamic: license-file
@@ -117,3 +119,3 @@
 
-OwOCR has been tested on Python 3.11, 3.12 and 3.13. It can be installed with `pip install owocr` after you install Python. You also need to have one or more OCR engines, check the list below for instructions. I recommend installing at least Google Lens on any operating system, and OneOCR if you are on Windows. Bing is pre-installed, Apple Vision and Live Text come pre-installed on macOS.
+OwOCR has been tested on Python 3.11 to 3.14. It can be installed with `pip install owocr` after you install Python. You also need to have one or more OCR engines, check the list below for instructions. I recommend installing at least Google Lens on any operating system, and OneOCR if you are on Windows. Bing is pre-installed, Apple Vision and Live Text come pre-installed on macOS.
 
@@ -190,3 +192,3 @@ ## Usage (terminal)
 - Google Vision - You need a service account .json file named google_vision.json in `user directory/.config/` → Terminal: install with `pip install "owocr[gvision]"`, key: `g`
-- Azure Image Analysis - You need to specify an api key and an endpoint in the config file → Terminal: install with `pip install "owocr[azure]"`, key: `v`
+- Azure Document Intelligence - You need to specify an api key and an endpoint in the config file → Terminal: install with `pip install "owocr[azure]"`, key: `v`
 - OCRSpace - You need to specify an api key in the config file. → Terminal key: `o`
@@ -193,0 +195,0 @@

pyproject.toml

@@ -8,5 +8,5 @@ [build-system]
 dynamic = ["version"]
-description = "Optical character recognition for Japanese text"
+description = "Multi-service, multi-platform optical character recognition"
 readme = "README.md"
-requires-python = ">=3.11"  # onnxruntime is not yet available for Python 3.14
+requires-python = ">=3.11"
 authors = [
@@ -32,2 +32,3 @@ {name = "AuroraWright", email = "[e-mail]"}
     "Programming Language :: Python :: 3.13",
+    "Programming Language :: Python :: 3.14",
     "Programming Language :: Python :: 3",
@@ -113,3 +114,3 @@ "Programming Language :: Python",
 azure = [
-    "azure-ai-vision-imageanalysis"
+    "azure-ai-documentintelligence"
 ]
@@ -122,3 +123,4 @@
     "networkx",
-    "tqdm"
+    "tqdm",
+    "opencv-python-headless"
 ]
@@ -125,0 +127,0 @@

README.md

@@ -30,3 +30,3 @@ <div align="center">
 
-OwOCR has been tested on Python 3.11, 3.12 and 3.13. It can be installed with `pip install owocr` after you install Python. You also need to have one or more OCR engines, check the list below for instructions. I recommend installing at least Google Lens on any operating system, and OneOCR if you are on Windows. Bing is pre-installed, Apple Vision and Live Text come pre-installed on macOS.
+OwOCR has been tested on Python 3.11 to 3.14. It can be installed with `pip install owocr` after you install Python. You also need to have one or more OCR engines, check the list below for instructions. I recommend installing at least Google Lens on any operating system, and OneOCR if you are on Windows. Bing is pre-installed, Apple Vision and Live Text come pre-installed on macOS.
 
@@ -103,3 +103,3 @@ ## Usage (terminal)
 - Google Vision - You need a service account .json file named google_vision.json in `user directory/.config/` → Terminal: install with `pip install "owocr[gvision]"`, key: `g`
-- Azure Image Analysis - You need to specify an api key and an endpoint in the config file → Terminal: install with `pip install "owocr[azure]"`, key: `v`
+- Azure Document Intelligence - You need to specify an api key and an endpoint in the config file → Terminal: install with `pip install "owocr[azure]"`, key: `v`
 - OCRSpace - You need to specify an api key in the config file. → Terminal key: `o`
@@ -106,0 +106,0 @@

Alert delta unavailable

Currently unable to show alert delta for PyPI packages.

Improved metrics

Total package byte prevSize

1187044

0.25%

Lines of code

17899

0.28%