update

cropper.py

@@ -0,0 +1,141 @@
+import os
+import fitz
+import numpy as np
+import argparse
+import traceback
+import io
+from PIL import Image
+
+
+def clip(value, min_value, max_value):
+    if min_value is not None:
+        if value < min_value:
+            value = min_value
+    if max_value is not None:
+        if value > max_value:
+            value = max_value
+    return value
+
+def crop(args):
+    infile = args.input
+    # process filepath
+    infile = os.path.join(os.getcwd(), infile)
+    infile = '\\'.join(infile.split('/'))
+    inbase = infile.rsplit('.', maxsplit=1)[0]
+
+    outfile = args.output
+    if outfile is None:
+        outfile = inbase + '_crop.pdf'
+    outbase = outfile.rsplit('.', maxsplit=1)[0]
+    names = None
+    thresh = args.thresh
+
+    # parse working dir.
+    workdir = '/'.join('/'.join(infile.split('\\')).split('/')[:-1])
+    if workdir == '':
+        workdir = '.'
+
+    if infile.split('.')[-1].lower() in ['ppt', 'pptx']:
+        # lets process pptx.
+        import comtypes.client
+        from pptx import Presentation
+        # convert to pdf.
+        powerpoint = comtypes.client.CreateObject("Powerpoint.Application")
+        powerpoint.Visible = 1
+        slides = powerpoint.Presentations.Open(infile)
+        print('convert {0} to {1}'.format(infile, inbase + '.pdf'))
+        slides.SaveAs(inbase + '.pdf', 32)
+        slides.close()
+        powerpoint.Quit()
+        # read comments.
+        ppt = Presentation(infile)
+        slides = ppt.slides
+        names = []
+        for i, slide in enumerate(slides):
+            anno = slide.notes_slide.notes_text_frame.text.strip()
+            if anno == '':
+                anno = '{0}_{1}.pdf'.format(outbase, i+1)
+            names.append(anno)
+        names = [n + '.pdf' if n[:-4] != '.pdf' else n for n in names]
+        infile = inbase + '.pdf'
+
+    if infile.split('.')[-1].lower() != 'pdf':
+        infile += '.pdf'
+
+
+    pdf = fitz.open(infile)
+    for i, page in enumerate(pdf):
+        try:
+            # recover crop box
+            page.setCropBox(page.MediaBox)
+            # render page as img
+            zoom = args.zoom
+            mat = fitz.Matrix(zoom, zoom)
+            img = page.getPixmap(matrix=mat)
+
+            # convert the img to nparray.
+            # img = [[img.pixel(j, i) for j in range(img.w)] for i in range(img.h)]
+            img = Image.open(io.BytesIO(img.tobytes()))
+            img = np.asarray(img, dtype=np.float32)
+
+            # get ltrb
+            background = np.broadcast_to(args.background_color, (img.shape))
+            diff = np.abs(img - background).mean(axis=2)
+            points = np.argwhere(diff > thresh)
+            x = points[:, 1]
+            y = points[:, 0]
+            left, right = min(x), max(x)
+            top, bottom = min(y), max(y)
+            right += 1
+            bottom += 1
+
+            # calculate real crop coordinate.
+            ltrb = left, top, right, bottom
+            ltrb = [coord/zoom for coord in ltrb]
+
+            # reserve border.
+            border = args.border
+            left, top, right, bottom = [x+y for x, y in zip(ltrb, [-border,-border,border,border])]
+
+            # clip ltrb
+            left = clip(left, 0, img.shape[1]/zoom)
+            top = clip(top, 0, img.shape[0]/zoom)
+            right = clip(right, 0, img.shape[1]/zoom)
+            bottom = clip(bottom, 0, img.shape[0]/zoom)
+            if args.visual:
+                print('cropbox =', (left, top), (right, bottom))
+            page.setCropBox(fitz.Rect(left, top, right, bottom))
+        except KeyboardInterrupt:
+            print('Operation Cancled.')
+            break
+        except:
+            print('Error happened while cropint page', i+1)
+            print(traceback.format_exc())
+
+    if not args.split:
+        pdf.save(outfile)
+    else:
+        if names is None:
+            names = ['{0}_{1}.pdf'.format(outbase, i+1) for i in range(len(pdf))]
+        if args.names is not None:
+            names = args.names
+            namesfile = os.path.join(workdir, names)
+            if os.path.isfile(namesfile) and names.split('.')[-1] == 'txt':
+                with open(namesfile, 'r', encoding='utf-8') as f:
+                    names = f.read()
+                names = names.split('\n')
+                names = [n.strip() for n in names]
+                names = [n for n in names if n != '']
+            else:
+                names = names.split(',')
+            names = [n + '.pdf' if n[:-4] != '.pdf' else n for n in names]
+        if len(names) < len(pdf):
+            names += ['{0}_{1}.pdf'.format(outbase, i+1) for i in range(len(names), len(pdf))]
+        names = [os.path.join(workdir, name) for name in names]
+        for i, (name, page) in enumerate(zip(names, pdf)):
+            page_pdf = fitz.Document()
+            page_pdf.insertPDF(pdf, from_page=i, to_page=i)
+            page_pdf.save(name)
+            if not args.mute:
+                print('saving cropped page to', name)
+    pdf.close()

main.py

@@ -0,0 +1,104 @@
+# import argparse
+import tkinter as tk
+from cropper import crop
+import traceback
+
+def str2bool(string):
+    positive = ['true',
+        't',
+        'y',
+        'yes',
+        '1',
+        'correct',
+        'accept',
+        'positive'
+    ]
+    if string.lower() in positive:
+        return True
+    else:
+        return False
+
+def str2list(string):
+    try:
+        string = string.split(',')
+        l = []
+        for sp in string:
+            if sp != '':
+                l.append(float(sp))
+        l = l[:3]
+        if len(l) < 3:
+            raise ValueError
+        return l
+    except:
+        print('parse list failed.')
+        raise ValueError
+
+def empty_as_none(string):
+    if string == '':
+        return None
+    else:
+        return string
+
+class Configuration():
+    def __init__(self) -> None:
+        pass
+
+def parse_and_crop(conf):
+    handlers = {
+        "output": empty_as_none,
+        "names": empty_as_none,
+        "background_color": str2list,
+        "border": float,
+        "zoom": float,
+        "thresh": float,
+        "split": str2bool,
+        "visual": str2bool,
+        "mute": str2bool,
+    }
+    for key in conf:
+        if key in handlers:
+            conf[key] = handlers[key](conf[key])
+    print('======= conf =======\n', '\n'.join(['{0}={1}'.format(key, value) for key, value in conf.items()]))
+    if conf['input'] == '':
+        print('Error, input file must be given.')
+    else:
+        args = Configuration()
+        args.__dict__.update(conf)
+        try:
+            crop(args)
+        except Exception:
+            print(traceback.format_exc())
+
+
+
+    
+
+# parser = argparse.ArgumentParser(description="Remove white border in pdf")
+# parser.add_argument('--input', '-i', type=str, help='path to the input pdf.')
+# parser.add_argument('--output', '-o', type=str, default=None, help='path to output file, default=infile_crop.pdf')
+# parser.add_argument('--background_color', '-bgc', type=str2list, default=[255, 255, 255], help='pixels that are considered as background')
+# parser.add_argument('--border', '-b', type=float, default=0.0, help='a value in pixel that specifies the border to given.')
+# parser.add_argument('--zoom', '-z', type=float, default=1.0, help='bigger is better, however also slower.')
+# parser.add_argument('--thresh', '-t', type=float, default=1, help='threshold that a pixel is considered as background')
+# parser.add_argument('--split', '-s', action='store_true', default=False, help='auto split the file, default names as out_1.pdf, ...')
+# parser.add_argument('--names', '-n', type=str, default=None, help='specify the name of the cropped pdf.')
+# parser.add_argument('--visual', '-v', default=False, action='store_true', help='display cropbox.')
+# parser.add_argument('--mute', '-m', default=False, action='store_true', help='do not display output file path.')
+# args = parser.parse_args()
+
+from ui import Loader
+conf = {
+        "input": {"name": "源文件", "type": "readfile", "extension": ("PDF & PPT", ".pdf .pptx")},
+        "output": {"name": "保存路径", "type": "savefile"},
+        "background_color": {"name": "背景颜色", "type": "color", "default": "255,255,255"},
+        "border": {"name": "留白", "type": "str", "default": "0.0"},
+        "zoom": {"name": "缩放等级", "type": "str", "default": "1.0"},
+        "thresh": {"name": "阈值", "type": "str", "default": "1.0"},
+        "split": {"name": "拆分", "type": "str", "default": "true"},
+        "names": {"name": "页名称", "type": "str"},
+        "visual": {"name": "显示裁切框", "type": "str", "default": "false"},
+        "mute": {"name": "显示保存文件", "type": "str", "default": "false"},
+    }
+root = tk.Tk()
+Loader(master=root, conf=conf, execution=parse_and_crop, title="PDF/PPT自动裁边")
+root.mainloop()

ui.py

@@ -0,0 +1,213 @@
+import tkinter as tk
+from tkinter import Text, ttk
+from tkinter.constants import DISABLED, RIGHT, NORMAL, END
+import tkinter.filedialog
+import tkinter.font as tkFont
+from PIL import Image, ImageTk, ImageFont, ImageDraw
+from tkinter.colorchooser import askcolor
+import sys
+
+class StdSimulator():
+    def __init__(self, box) -> None:
+        self.box = box
+        self.stdout = sys.stdout
+        self.stderr = sys.stderr
+    
+    def write(self, msg):
+        self.box.insert(END, msg)
+        self.box.see(END)
+    
+    def flush(self):
+        pass
+
+def text_on_image(img, text, pos, size=16, color='#000000'):
+    imgdraw = ImageDraw.Draw(img)
+    imgfont = ImageFont.truetype("msyh.ttc",size=size)
+    textw, texth = imgdraw.textsize(text, font=imgfont)
+    left_top = (pos[0]-int(textw/2), pos[1]-int(texth/2))
+    imgdraw.text(left_top, text, fill=color, font=imgfont)
+    return img
+
+class Loader(tk.Frame):
+    def __init__(self, master=None, conf={}, execution=None, title=None):
+        super().__init__(master)
+        self.master = master
+        self.conf = conf
+        self.execution = execution
+        self.title = title
+
+        self.pack(fill='both', expand=1)
+        self.resources = {}
+        self.fontStyle = tkFont.Font(size=10, family='微软雅黑')
+        self.fontStyle_anchor = tkFont.Font(size=16)
+        self.create_widgets()
+        # self.font = tk.Font(family='', size=40,weight='',slant='',underline='',overstrike='')
+        if self.title is not None:
+            self.master.title(self.title)
+    
+    def add_resource(self, obj, name):
+        if name in self.resources:
+            print('Name [{0}] Repeated.'.format(name))
+            print('Current used name:', list(self.resources.keys()))
+            raise ValueError
+        self.resources[name] = obj
+    
+    def add_entry(self, default, width, layout, name, master=None):
+        if master is None:
+            master = self
+        resource = tk.Entry(master, width=width)
+        if type(default) is str:
+            resource.insert(0, default)
+        resource.grid(**layout)
+        self.add_resource(resource, name)
+
+    def add_label(self, text, layout, name, master=None):
+        if master is None:
+            master = self
+        resource = tk.Label(master, text=text, font=self.fontStyle)
+        resource.grid(**layout)
+        self.add_resource(resource, name)
+
+    def add_button(self, text, command, layout, size, name, image=None, master=None):
+        if master is None:
+            master = self
+        resource = {}
+        if image is not None:
+            width, height = size
+            image = image.resize((width,height))
+            if text is not None:
+                image = text_on_image(image, text, (int(width/2), int(height/2)), size=16, color='#000000')
+            img = ImageTk.PhotoImage(image)
+            resource['img'] = img
+            resource['button'] = tk.Button(master, text=text, command=command, borderwidth=0)
+            resource['button'].config(image=img)
+        else:
+            resource['button'] = tk.Button(master, text=text, command=command)
+        resource['button'].grid(**layout)
+        self.add_resource(resource, name)
+    
+    def add_textbox(self, width, layout, name, master=None):
+        if master is None:
+            master = self
+        resource = tk.Text(master, width=width, state=NORMAL)
+        resource.grid(**layout)
+        # vsb = tk.Scrollbar(self, orient="vertical", command=resource.yview)
+        # resource.configure(yscrollcommand=vsb.set)
+        self.add_resource(resource, name)
+        # self.add_resource(vsb, name + '_scroll')
+
+
+    def choose_file(self, dst, choose_type='readfile', extension=None, initial_file=None):
+        default_extension = ("所有文件", '.*')
+        args = {"filetypes": []}
+        if extension is not None:
+            args['filetypes'].append(extension)
+        args['filetypes'].append(default_extension)
+        if initial_file is not None:
+            args['initialfile'] = initial_file
+
+        # print('args:', args)
+        filename = ''
+        if choose_type == 'readfile':
+            filename = tkinter.filedialog.askopenfilename(**args)
+        elif choose_type == 'savefile':
+            filename = tkinter.filedialog.asksaveasfilename(**args)
+        elif choose_type == 'directory':
+            args = {
+                'initialdir': initial_file
+            }
+            filename = tkinter.filedialog.askdirectory(**args)
+        if filename != '':
+            self.resources[dst].delete(0, 'end')
+            self.resources[dst].insert(0, filename)
+    
+    def choose_color(self, dst):
+        color = askcolor()[0]
+        if color is not None:
+            color = [str(c) for c in color]
+            color = ','.join(color)
+            self.resources[dst].delete(0, 'end')
+            self.resources[dst].insert(0, color)
+    
+    def execute(self):
+        conf = {}
+        for key in self.resources:
+            dtype, name = key.split('_', maxsplit=1)
+            if dtype == 'entry':
+                conf[name] = self.resources[key].get()
+        self.execution(conf)
+    
+    def create_widgets(self):
+        self.rowcount = 0
+        for key, value in self.conf.items():
+            name = value['name']
+            dtype = value['type']
+            default = value.get('default', '')
+            self.add_label(name, {"row": self.rowcount, "column": 0, "padx": 10, "pady": 5}, name="label_" + key)
+            self.add_entry(
+                default=default,
+                width=None,
+                layout={"row": self.rowcount, "column": 1, "padx": 0, "sticky": "we"},
+                name='entry_' + key
+            )
+            destination = 'entry_' + key
+            if dtype in ['readfile', 'savefile', 'directory']:
+                extension = value.get('extension', None)
+                initialfile = default
+                args = {
+                    "dst": destination,
+                    "choose_type": dtype,
+                    "extension": extension,
+                    "initial_file": initialfile
+                }
+                self.add_button(
+                    text="选择文件",
+                    command=lambda args=args:self.choose_file(**args),
+                    size=(16, 16),
+                    layout={"row":self.rowcount, "column": 2, "padx": 5},
+                    name='{0}_choose'.format(name),
+                    image=None
+                )
+                self.conf[key]['type'] = 'str'
+            elif dtype == 'color':
+                self.add_button(
+                    text="选择颜色",
+                    command=lambda dst=destination:self.choose_color(dst),
+                    size=(16, 16),
+                    layout={"row":self.rowcount, "column": 2, "padx": 5},
+                    name='{0}_choose'.format(name),
+                    image=None
+                )
+
+            self.rowcount += 1
+        self.add_button(
+            text="         裁剪           ",
+            command=self.execute,
+            size=(16, 16),
+            layout={"row":self.rowcount, "column": 0, "columnspan": 2, "pady": 5},
+            name='execute_final'
+        )
+        self.add_textbox(
+            width=40, 
+            layout={"row":0, "column": 3, "rowspan": self.rowcount, "padx": 5, "pady": 0},
+            name="textbox_log"
+        )
+        sys.stdout = StdSimulator(self.resources['textbox_log'])
+        sys.stderr = StdSimulator(self.resources['textbox_log'])
+
+if __name__ == '__main__':
+    conf = {
+        "input": {"name": "源文件", "type": "readfile", "extension": ("PDF & PPT", ".pdf .pptx")},
+        "output": {"name": "保存路径", "type": "savefile"},
+        "background_color": {"name": "背景颜色", "type": "color", "default": "255,255,255"},
+        "border": {"name": "留白", "type": "str", "default": "0.0"},
+        "zoom": {"name": "缩放等级", "type": "str", "default": "1.0"},
+        "thresh": {"name": "阈值", "type": "str", "default": "1.0"},
+        "split": {"name": "拆分", "type": "str", "default": "false"},
+        "names": {"name": "页名称", "type": "str"},
+        "visual": {"name": "显示裁切框", "type": "str", "default": "false"},
+        "mute": {"name": "显示保存文件", "type": "str", "default": "false"},
+    }
+    root = tk.Tk()
+    Loader(master=root, conf=conf)
+    root.mainloop()