Mirror
/
Umi-OCR
mirror of https://github.com/hiroi-sora/Umi-OCR.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
Browse Source

重构部分代码，添加新功能：自动保存配置、生成md文件等

release/1.2.6
hiroi-sora 3 years ago
parent
90a553cf5a
commit
9a8dacd161
1 changed files with 392 additions and 241 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 633
      main.py

633
main.py
View File

@ -1,6 +1,7 @@
from selectAreaWin import SelectAreaWin # 子窗口
from asset import iconPngBase64, getHelpText # 资源
from asset import IconPngBase64, GetHelpText # 资源
from callingOCR import CallingOCR # OCR调用接口
from config import Config
import os
import time
@ -9,12 +10,12 @@ import threading # 线程
from PIL import Image
import tkinter as tk
import tkinter.filedialog
from tkinter import ttk
from tkinter import Variable, ttk
from windnd import hook_dropfiles # 文件拖拽
from pyperclip import copy as pyperclipCopy # 剪贴板
from webbrowser import open as webOpen # “关于”面板打开项目网址
ProjectVer = "1.1" # 版本号
ProjectVer = "1.2.0" # 版本号
ProjectName = f"Umi-OCR 批量图片转文字 v{ProjectVer}" # 名称
ProjectWeb = "https://github.com/hiroi-sora/Umi-OCR"
@ -22,14 +23,15 @@ ProjectWeb = "https://github.com/hiroi-sora/Umi-OCR"
class Win:
def __init__(self):
self.imgDict = {} # 当前载入的图片信息字典，key为表格组件id。必须为有序字典，python3.6以上默认是。
self.areaInfo = None # 特殊处理区域数据
# self.areaInfo = None # 特殊处理区域数据
self.isRunning = 0 # 0未在运行，1正在运行，2正在停止
def initWin(): # 初始化主窗口
# 1.初始化主窗口
def initWin():
self.win = tk.Tk()
self.win.title(ProjectName)
# 窗口大小与位置
w, h = 360, 520 # 窗口初始大小与最小大小
w, h = 360, 500 # 窗口初始大小与最小大小
ws, hs = self.win.winfo_screenwidth(), self.win.winfo_screenheight()
x, y = round(ws/2 - w/2), round(hs/2 - h/2) # 初始位置，屏幕正中
self.win.minsize(w, h) # 最小大小
@ -37,20 +39,57 @@ class Win:
self.win.protocol("WM_DELETE_WINDOW", self.onClose) # 窗口关闭
# 图标
self.iconImg = tkinter.PhotoImage(
data=iconPngBase64) # 载入图标，base64转
data=IconPngBase64) # 载入图标，base64转
self.win.iconphoto(False, self.iconImg) # 设置窗口图标
initWin()
# 2.初始化变量、配置项
def initVar():
self.cfgVar = { # 设置项tk变量
# 输出文件设置
"isOutputFile": tk.BooleanVar(), # T时输出内容写入本地文件
"outputFilePath": tk.StringVar(), # 输出文件目录
"outputFileName": tk.StringVar(), # 输出文件名称
# 输出格式设置
"isOutputDebug": tk.BooleanVar(), # T时输出调试信息
"isIgnoreNoText": tk.BooleanVar(), # T时忽略(不输出)没有文字的图片信息
"outputStyle": tk.IntVar(), # 1：纯文本，2：Markdown
# 识别器设置
"ocrToolPath": tk.StringVar(), # 识别器路径
"imageSuffix": tk.StringVar(), # 图片后缀
}
Config.initValue(self.cfgVar) # 初始化设置项
# 面板值改变时，更新到配置值，并写入本地
self.saveTimer = None # 计时器，改变面板值一段时间后写入本地
def configSave(): # 保存值的事件
Config.save()
self.saveTimer = None
def valueChange(key): # 值改变的事件
Config.update(key) # 更新配置项
if Config.isSaveItem(key):
if self.saveTimer: # 计时器已存在，则停止已存在的
self.win.after_cancel(self.saveTimer)
self.saveTimer = None
self.saveTimer = self.win.after(200, configSave)
for key in self.cfgVar:
self.cfgVar[key].trace( # 跟踪值改变事件
"w", lambda *e, key=key: valueChange(key))
initVar()
# 3. 初始化组件
def initTop(): # 顶部按钮
tk.Frame(self.win, height=5).pack(side='top')
vFrame1 = tk.Frame(self.win)
vFrame1.pack(side='top', fill="x", padx=5)
fr = tk.Frame(self.win)
fr.pack(side='top', fill="x", padx=5)
# 右侧按钮
self.btnRun = tk.Button(
vFrame1, text='开始任务', width=12, height=2, command=self.run)
fr, text='开始任务', width=12, height=2, command=self.run)
self.btnRun.pack(side='right', padx=5)
# 左侧文本和进度条
vFrame2 = tk.Frame(vFrame1)
vFrame2 = tk.Frame(fr)
vFrame2.pack(side='top', fill='x')
labelUse = tk.Label(vFrame2, text="使用说明",
fg="gray", cursor="hand2")
@ -62,35 +101,33 @@ class Win:
self.labelFractions.pack(side='right')
self.labelTime = tk.Label(vFrame2, text="0s") # 已用时间 12.3s
self.labelTime.pack(side='right', padx=5)
self.progressbar = ttk.Progressbar(vFrame1)
self.progressbar = ttk.Progressbar(fr)
self.progressbar.pack(side='top', padx=5, fill="x")
initTop()
# 初始化选项卡组件
self.notebook = ttk.Notebook(self.win)
self.notebook = ttk.Notebook(self.win) # 初始化选项卡组件
self.notebook.pack(expand=True, fill=tk.BOTH) # 填满父组件
def initTab1(): # 表格卡
tabFrame = tk.Frame(self.notebook) # 选项卡主容器
self.notebook.add(tabFrame, text=f'{"处理列表": ^10s}')
vFrame1 = tk.Frame(tabFrame)
vFrame1.pack(side='top', fill='x', pady=2)
tk.Button(vFrame1, text=' 浏览 ', command=self.openFileWin).pack(
# 顶栏
fr1 = tk.Frame(tabFrame)
fr1.pack(side='top', fill='x', pady=2)
tk.Button(fr1, text=' 浏览 ', command=self.openFileWin).pack(
side='left', padx=5)
tk.Label(vFrame1, text="或直接拖入").pack(side='left')
tk.Button(vFrame1, text='清空表格', width=12, command=self.clearTable).pack(
tk.Label(fr1, text="或直接拖入").pack(side='left')
tk.Button(fr1, text='清空表格', width=12, command=self.clearTable).pack(
side='right')
tk.Button(vFrame1, text='移除选中图片', width=12, command=self.delImgList).pack(
tk.Button(fr1, text='移除选中图片', width=12, command=self.delImgList).pack(
side='right', padx=5)
vFrame2 = tk.Frame(tabFrame)
vFrame2.pack(side='top', fill='both')
columns = ['name', 'time', 'score']
# 表格主体
fr2 = tk.Frame(tabFrame)
fr2.pack(side='top', fill='both')
self.table = ttk.Treeview(
master=vFrame2, # 父容器
master=fr2, # 父容器
height=50, # 表格显示的行数,height行
columns=columns, # 显示的列
columns=['name', 'time', 'score'], # 显示的列
show='headings', # 隐藏首列
)
hook_dropfiles(self.table, func=self.draggedImages) # 注册文件拖入
@ -101,144 +138,180 @@ class Win:
self.table.column('name', minwidth=40)
self.table.column('time', width=20, minwidth=20)
self.table.column('score', width=30, minwidth=30)
scroll = tk.Scrollbar( # 绑定滚动条
vFrame2, orient='vertical', command=self.table.yview)
scroll.pack(side="left", fill='y')
self.table["yscrollcommand"] = scroll.set
vbar = tk.Scrollbar( # 绑定滚动条
fr2, orient='vertical', command=self.table.yview)
vbar.pack(side="left", fill='y')
self.table["yscrollcommand"] = vbar.set
initTab1()
def initTab2(): # 输出卡
self.tabFrameOutput = tk.Frame(self.notebook) # 选项卡主容器
self.notebook.add(self.tabFrameOutput, text=f'{"识别内容": ^10s}')
vFrame1 = tk.Frame(self.tabFrameOutput)
vFrame1.pack(side='top', fill='x', pady=2)
fr1 = tk.Frame(self.tabFrameOutput)
fr1.pack(side='top', fill='x', pady=2)
self.isAutoRoll = tk.IntVar()
self.isAutoRoll.set(1)
tk.Checkbutton(vFrame1, variable=self.isAutoRoll, text="自动滚动到底部").pack(
tk.Checkbutton(fr1, variable=self.isAutoRoll, text="自动滚动到底部").pack(
side='left')
tk.Button(vFrame1, text='清空版面', width=12,
tk.Button(fr1, text='清空版面', width=12,
command=lambda: self.textOutput.delete('1.0', tk.END)).pack(side='right')
tk.Button(vFrame1, text='复制到剪贴板', width=12,
tk.Button(fr1, text='复制到剪贴板', width=12,
command=lambda: pyperclipCopy(self.textOutput.get("1.0", tk.END))).pack(side='right', padx=5)
vFrame2 = tk.Frame(self.tabFrameOutput)
vFrame2.pack(side='top', fill='both')
scroll = tk.Scrollbar( # 滚动条
vFrame2, orient='vertical')
scroll.pack(side="right", fill='y')
self.textOutput = tk.Text(vFrame2, height=500, width=500)
fr2 = tk.Frame(self.tabFrameOutput)
fr2.pack(side='top', fill='both')
vbar = tk.Scrollbar(fr2, orient='vertical') # 滚动条
vbar.pack(side="right", fill='y')
self.textOutput = tk.Text(fr2, height=500, width=500)
self.textOutput.pack(fill='both', side="left")
scroll["command"] = self.textOutput.yview
self.textOutput["yscrollcommand"] = scroll.set
vbar["command"] = self.textOutput.yview
self.textOutput["yscrollcommand"] = vbar.set
initTab2()
def initTab3(): # 设置卡
tabFrame = tk.Frame(self.notebook) # 选项卡主容器
self.notebook.add(tabFrame, text=f'{"设置": ^10s}')
self.labelOptionTips = tk.Label(tabFrame, fg="red") # 提示
self.labelOptionTips.pack()
# 输出文件设置
vFrameOutFile = tk.LabelFrame(tabFrame, text="输出设置")
vFrameOutFile.pack(side='top', fill='x', pady=2, padx=5)
vFrameO1 = tk.Frame(vFrameOutFile)
vFrameO1.pack(side='top', fill='x', pady=2)
self.isOutputFile = tk.IntVar()
self.isOutputFile.set(1)
tk.Checkbutton(vFrameO1, variable=self.isOutputFile, text="将识别内容写入txt文件").pack(
side='left')
self.isOutputDebug = tk.IntVar()
self.isOutputDebug.set(0)
tk.Checkbutton(vFrameO1, variable=self.isOutputDebug, text="输出调试信息").pack(
side='left', padx=23)
tk.Label(vFrameOutFile, fg="gray", text="下面两项为空时，默认输出到第一张图片所在的文件夹").pack(
side='top', padx=5)
vFrameO2 = tk.Frame(vFrameOutFile)
vFrameO2.pack(side='top', fill='x', pady=2)
tk.Label(vFrameO2, text="输出目录： ").pack(
side='left', padx=5)
self.enOutPath = tk.Entry(vFrameO2)
self.enOutPath.pack(side='top', fill="x", padx=5)
vFrameO3 = tk.Frame(vFrameOutFile)
vFrameO3.pack(side='top', fill='x', pady=2)
tk.Label(vFrameO3, text="输出文件名：").pack(side='left', padx=5)
self.enOutName = tk.Entry(vFrameO3)
self.enOutName.pack(side='top', fill="x", padx=5)
# 忽略区域
vFrameArea = tk.LabelFrame(tabFrame, text="忽略图片中某些区域内的文字")
vFrameArea.pack(side='top', fill='x', pady=2, padx=5)
vFrameA1 = tk.Frame(vFrameArea)
vFrameA1.pack(side='top', fill='x', pady=2)
tk.Button(vFrameA1, text='添加忽略区域',
command=self.openSelectArea).pack(side="left", padx=5)
tk.Button(vFrameA1, text='清空所有区域',
command=self.clearArea).pack(side="left")
self.areaLabel = tk.Label(vFrameA1, text="待添加", padx=5)
self.areaLabel.pack(side="right")
self.canvasHeight = 140 # 画板高度不变，宽度根据选区回传数据调整
self.canvas = tk.Canvas(vFrameArea, width=249, height=self.canvasHeight,
bg="black")
self.canvas.pack(side='top')
# 识别器exe与图片后缀设置
vFrameEXE = tk.LabelFrame(tabFrame, text="识别器设置 [切换多国语言和不同格式图片]")
vFrameEXE.pack(side='top', fill='x', pady=2, padx=5)
vFrame5 = tk.Frame(vFrameEXE)
vFrame5.pack(side='top', fill='x', pady=2)
tk.Label(vFrame5, text="识别器路径：").pack(
side='left', padx=5)
self.enEXE = tk.Entry(vFrame5)
self.enEXE.pack(side='top', fill="x", padx=5)
self.enEXE.insert(0, "PaddleOCR-json\PaddleOCR_json.exe")
vFrame4 = tk.Frame(vFrameEXE)
vFrame4.pack(side='top', fill='x', pady=2)
tk.Label(vFrame4, text="图片后缀： ").pack(
side='left', padx=5)
self.enInSuffix = tk.Entry(vFrame4)
self.enInSuffix.pack(side='top', fill="x", padx=5)
self.enInSuffix.insert(
0, ".jpg .jpe .jpeg .jfif .png .webp .bmp .tif .tiff")
# 关于面板，隐藏在设置选项卡默认大小的下方外面
tk.Frame(tabFrame, height=20).pack()
vFrameAbout = tk.LabelFrame(
tabFrame, text="关于")
vFrameAbout.pack(side='top', fill='both', padx=5, ipady=10)
tk.Label(vFrameAbout, image=self.iconImg).pack() # 图标
tk.Label(vFrameAbout, text=ProjectName, fg="gray").pack()
labelWeb = tk.Label(vFrameAbout, text=ProjectWeb, cursor="hand2",
fg="deeppink")
labelWeb.pack() # 文字
labelWeb.bind( # 绑定鼠标左键点击，打开网页
'<Button-1>', self.openProjectWeb)
tk.Frame(tabFrame, height=150).pack()
tk.Label(tabFrame, text="真没有啦！", fg="gray").pack()
def initOptFrame(): # 初始化可滚动画布 及 内嵌框架
optVbar = tk.Scrollbar(
tabFrame, orient="vertical") # 创建滚动条
optVbar.pack(side="right", fill="y")
self.optCanvas = tk.Canvas(
tabFrame, highlightthickness=0) # 创建画布，用于承载框架。highlightthickness取消高亮边框
self.optCanvas.pack(side="left", fill="both",
expand="yes") # 填满父窗口
self.optCanvas["yscrollcommand"] = optVbar.set # 绑定滚动条
optVbar["command"] = self.optCanvas.yview
self.optFrame = tk.Frame(self.optCanvas) # 容纳设置项的框架
self.optFrame.pack()
self.optCanvas.create_window( # 框架塞进画布
(0, 0), window=self.optFrame, anchor="nw")
self.labelOptionTips = tk.Label(self.optFrame, fg="red")
self.labelOptionTips.pack()
initOptFrame()
LabelFramePadY = 3 # 每个区域上下间距
def initArea(): # 忽略区域设置
self.areaLabel = tk.LabelFrame(
self.optFrame, text="忽略图片中某些区域内的文字")
self.areaLabel.pack(side='top', fill='x',
ipady=2, pady=LabelFramePadY, padx=4)
self.areaLabel.grid_columnconfigure(0, minsize=4)
tk.Button(self.areaLabel, text='添加区域',
command=self.openSelectArea).grid(column=1, row=0, sticky="w")
tk.Button(self.areaLabel, text='清空区域',
command=self.clearArea).grid(column=1, row=1, sticky="w")
self.areaLabel.grid_rowconfigure(2, minsize=10)
# tk.Button(self.areaLabel, text='读取预设',
# command=self.showTest).grid(column=1, row=3, sticky="w")
# tk.Button(self.areaLabel, text='保存预设',
# command=self.showTest).grid(column=1, row=4, sticky="w")
self.areaLabel.grid_columnconfigure(2, minsize=4)
self.canvasHeight = 140 # 画板高度不变，宽度根据选区回传数据调整
self.canvas = tk.Canvas(self.areaLabel, width=249, height=self.canvasHeight,
bg="black")
self.canvas.grid(column=3, row=0, rowspan=10)
initArea()
def initOutFile(): # 输出文件设置
frameOutFile = tk.LabelFrame(self.optFrame, text="输出设置")
frameOutFile.pack(side='top', fill='x',
ipady=2, pady=LabelFramePadY, padx=4)
fr1 = tk.Frame(frameOutFile)
fr1.pack(side='top', fill='x', pady=2, padx=5)
tk.Checkbutton(fr1, variable=self.cfgVar["isOutputFile"],
text="将识别内容写入本地文件").grid(column=0, row=0, columnspan=2, sticky="w")
tk.Checkbutton(fr1, text="输出调试信息", variable=self.cfgVar["isOutputDebug"]
).grid(column=0, row=1, sticky="w")
tk.Checkbutton(fr1, text="忽略无文字的图片", variable=self.cfgVar["isIgnoreNoText"],
).grid(column=1, row=1, sticky="w")
tk.Radiobutton(fr1, text='纯文本.txt文件', value=1, variable=self.cfgVar["outputStyle"],
).grid(column=0, row=2, sticky="w")
tk.Radiobutton(fr1, text='Markdown风格.md文件', value=2, variable=self.cfgVar["outputStyle"],
).grid(column=1, row=2, sticky="w")
tk.Label(fr1, fg="gray",
text="下面两项为空时，默认输出到第一张图片所在的文件夹"
).grid(column=0, row=3, columnspan=2, sticky="nsew")
fr2 = tk.Frame(frameOutFile)
fr2.pack(side='top', fill='x', pady=2, padx=5)
tk.Label(fr2, text="输出目录：").grid(column=0, row=3, sticky="w")
enOutPath = tk.Entry(
fr2, textvariable=self.cfgVar["outputFilePath"])
enOutPath.grid(column=1, row=3, sticky="nsew")
fr2.grid_rowconfigure(4, minsize=2) # 第二行拉开间距
tk.Label(fr2, text="输出文件名：").grid(column=0, row=5, sticky="w")
enOutName = tk.Entry(
fr2, textvariable=self.cfgVar["outputFileName"])
enOutName.grid(column=1, row=5, sticky="nsew")
fr2.grid_columnconfigure(1, weight=1) # 第二列自动扩充
initOutFile()
def initOcrUI(): # 识别器exe与图片后缀设置
frameOCR = tk.LabelFrame(
self.optFrame, text="识别器设置 [切换多国语言和不同格式图片]")
frameOCR.pack(side='top', fill='x', ipady=2,
pady=LabelFramePadY, padx=4)
fr1 = tk.Frame(frameOCR)
fr1.pack(side='top', fill='x', pady=2, padx=5)
tk.Label(fr1, text="识别器路径：").grid(column=0, row=0, sticky="w")
enEXE = tk.Entry(fr1, textvariable=self.cfgVar["ocrToolPath"])
enEXE.grid(column=1, row=0, sticky="nsew")
tk.Label(fr1, text="图片后缀：").grid(column=0, row=2, sticky="w")
enInSuffix = tk.Entry(
fr1, textvariable=self.cfgVar["imageSuffix"])
enInSuffix.grid(column=1, row=2, sticky="nsew")
fr1.grid_columnconfigure(1, weight=1)
fr1.grid_rowconfigure(1, minsize=2)
initOcrUI()
def initAbout(): # 关于面板
frameAbout = tk.LabelFrame(
self.optFrame, text="关于")
frameAbout.pack(side='top', fill='x', ipady=2,
pady=LabelFramePadY, padx=4)
tk.Label(frameAbout, image=self.iconImg).pack() # 图标
tk.Label(frameAbout, text=ProjectName, fg="gray").pack()
labelWeb = tk.Label(frameAbout, text=ProjectWeb, cursor="hand2",
fg="deeppink")
labelWeb.pack() # 文字
labelWeb.bind( # 绑定鼠标左键点击，打开网页
'<Button-1>', self.openProjectWeb)
initAbout()
def initOptFrameWH(): # 初始化框架的宽高
self.optFrame.update() # 强制刷新
rH = self.optFrame.winfo_height() # 由组件撑起的 框架高度
self.optCanvas.config(scrollregion=(0, 0, 0, rH)) # 画布内高度为框架高度
self.optFrame.pack_propagate(False) # 禁用框架自动宽高调整
self.optFrame["height"] = rH # 手动还原高度。一次性设置，之后无需再管。
self.optCanvasWidth = 1 # 宽度则是随窗口大小而改变。
def onCanvasResize(event): # 绑定画布大小改变事件
cW = event.width-3 # 当前 画布宽度
if not cW == self.optCanvasWidth: # 若与上次不同：
self.optFrame["width"] = cW # 修改设置页 框架宽度
self.optCanvasWidth = cW
self.optCanvas.bind( # 绑定画布大小改变事件。只有画布组件前台显示时才会触发，减少性能占用
'<Configure>', onCanvasResize)
def onCanvasMouseWheel(event): # 绑定画布中滚轮滚动事件
self.optCanvas.yview_scroll(
1 if event.delta < 0 else -1, "units")
self.optCanvas.bind_all("<MouseWheel>", onCanvasMouseWheel)
initOptFrameWH()
# self.notebook.select(tabFrame)
initTab3()
self.win.mainloop()
def openSelectArea(self): # 打开选择区域
if not self.isRunning == 0:
return
defaultPath = ""
if self.imgDict:
defaultPath = next(iter(self.imgDict.values()))["path"]
self.win.attributes("-disabled", 1) # 禁用父窗口
SelectAreaWin(self.closeSelectArea, defaultPath)
def closeSelectArea(self, info=None): # 关闭选择区域，获取选择区域数据
self.win.attributes("-disabled", 0) # 启用父窗口
if not info:
return
self.areaInfo = info
self.areaLabel["text"] = f"生效分辨率：{info[0][0]}x{info[0][1]}"
self.canvas.delete(tk.ALL) # 清除画布
scale = self.canvasHeight / info[0][1] # 显示缩放比例
width = int(self.canvasHeight * (info[0][0] / info[0][1]))
self.canvas["width"] = width
areaColor = ["red", "green", "gold1"]
for i in range(3):
for a in info[1][i]:
x0, y0 = a[0][0]*scale, a[0][1]*scale,
x1, y1 = a[1][0]*scale, a[1][1]*scale,
self.canvas.create_rectangle(
x0, y0, x1, y1, fill=areaColor[i]) # 绘制新图
# 加载图片 ===============================================
def draggedImages(self, paths): # 拖入图片
if not self.isRunning == 0:
@ -251,51 +324,87 @@ class Win:
def openFileWin(self): # 打开选择文件窗
if not self.isRunning == 0:
return
suf = self.enInSuffix.get() # 许可后缀
suf = Config.get("imageSuffix") # 许可后缀
paths = tk.filedialog.askopenfilenames(
title='选择图片', filetypes=[('图片', suf)])
self.addImagesList(paths)
def addImagesList(self, paths):
suf = self.enInSuffix.get().split() # 许可后缀
def addImagesList(self, paths): # 添加一批图片列表
suf = Config.get("imageSuffix").split() # 许可后缀列表
def addImage(path): # 添加一张图片。传入路径，许可后缀。
path = path.replace("/", "\\") # 浏览是左斜杠，拖入是右斜杠；需要统一
if suf and os.path.splitext(path)[1].lower() not in suf:
return # 需要判别许可后缀 且 文件后缀不在许可内，不添加。
# 检测是否重复
for key, value in self.imgDict.items():
if value["path"] == path:
return
# 检测是否可用
try:
s = Image.open(path).size
except Exception as e:
tk.messagebox.showwarning(
"遇到了一点小问题", f"图片载入失败。图片地址：\n{path}\n\n错误信息：\n{e}")
return
# 计算路径
p = os.path.abspath(os.path.join(path, os.pardir)) # 父文件夹
if not Config.get("outputFilePath"): # 初始化输出路径
Config.set("outputFilePath", p)
if not Config.get("outputFileName"): # 初始化输出文件名
n = f"[转文字]_{os.path.basename(p)}"
Config.set("outputFileName", n)
# 加入待处理列表
name = os.path.basename(path) # 带后缀的文件名
tableInfo = (name, "", "")
id = self.table.insert('', 'end', values=tableInfo) # 添加到表格组件中
dictInfo = {"name": name, "path": path, "size": s}
self.imgDict[id] = (dictInfo) # 添加到字典中
for path in paths: # 遍历拖入的所有路径
if os.path.isdir(path): # 若是目录
subFiles = os.listdir(path) # 遍历子文件
for s in subFiles:
self.addImage(path+"\\"+s, suf) # 添加
addImage(path+"\\"+s) # 添加
elif os.path.isfile(path): # 若是文件：
self.addImage(path, suf) # 直接添加
addImage(path) # 直接添加
def addImage(self, path, okSuf=None): # 添加一张图片。传入路径，许可后缀。
path = path.replace("/", "\\") # 浏览是左斜杠，拖入是右斜杠；需要统一
if okSuf and os.path.splitext(path)[1].lower() not in okSuf:
return # 需要判别许可后缀 且 文件后缀不在许可内，不添加。
# 检测是否重复
for key, value in self.imgDict.items():
if value["path"] == path:
return
# 检测是否可用
try:
s = Image.open(path).size
except Exception as e:
tk.messagebox.showwarning(
"遇到了一点小问题", f"图片载入失败。图片地址：\n{path}\n\n错误信息：\n{e}")
# 忽略区域 ===============================================
def openSelectArea(self): # 打开选择区域
if not self.isRunning == 0:
return
# 计算路径
p = os.path.abspath(os.path.join(path, os.pardir)) # 父文件夹
if not self.enOutPath.get(): # 初始化输出路径
self.enOutPath.delete('0', tk.END)
self.enOutPath.insert(0, p)
if not self.enOutName.get(): # 初始化输出文件名
n = f"[转文字]_{os.path.basename(p)}.txt"
self.enOutName.delete('0', tk.END)
self.enOutName.insert(0, n)
# 加入待处理列表
name = os.path.basename(path) # 带后缀的文件名
tableInfo = (name, "", "")
id = self.table.insert('', 'end', values=tableInfo) # 添加到表格组件中
dictInfo = {"name": name, "path": path, "size": s}
self.imgDict[id] = (dictInfo) # 添加到字典中
defaultPath = ""
if self.imgDict:
defaultPath = next(iter(self.imgDict.values()))["path"]
self.win.attributes("-disabled", 1) # 禁用父窗口
SelectAreaWin(self.closeSelectArea, defaultPath)
def closeSelectArea(self): # 关闭选择区域，获取选择区域数据
self.win.attributes("-disabled", 0) # 启用父窗口
area = Config.get("ignoreArea")
if not area:
return
self.areaLabel["text"] = f"忽略区域 生效分辨率：{area['size'][0]}x{area['size'][1]}"
self.canvas.delete(tk.ALL) # 清除画布
scale = self.canvasHeight / area['size'][1] # 显示缩放比例
width = int(self.canvasHeight * (area['size'][0] / area['size'][1]))
self.canvas["width"] = width
areaColor = ["red", "green", "gold1"]
for i in range(3):
for a in area['area'][i]:
x0, y0 = a[0][0]*scale, a[0][1]*scale,
x1, y1 = a[1][0]*scale, a[1][1]*scale,
self.canvas.create_rectangle(
x0, y0, x1, y1, fill=areaColor[i]) # 绘制新图
def clearArea(self): # 清空忽略区域
Config.set("ignoreArea", None)
self.areaLabel["text"] = "忽略图片中某些区域内的文字"
self.canvas.delete(tk.ALL) # 清除画布
self.canvas["width"] = int(self.canvasHeight * (16/9))
# 表格操作 ===============================================
def clearTable(self): # 清空表格
if not self.isRunning == 0:
@ -304,19 +413,13 @@ class Win:
self.labelPercentage["text"] = "0%"
self.labelFractions["text"] = "0/0"
self.labelTime["text"] = "0s"
self.enOutPath.delete('0', tk.END)
self.enOutName.delete('0', tk.END)
Config.set("outputFilePath", "")
Config.set("outputFileName", "")
self.imgDict = {}
chi = self.table.get_children()
for i in chi:
self.table.delete(i) # 表格组件移除
def clearArea(self): # 清空忽略区域
self.areaInfo = None
self.areaLabel["text"] = "待添加"
self.canvas.delete(tk.ALL) # 清除画布
self.canvas["width"] = int(self.canvasHeight * (16/9))
def delImgList(self): # 图片列表中删除选中
if not self.isRunning == 0:
return
@ -344,14 +447,16 @@ class Win:
if not self.imgDict:
return
# 检测识别器存在
exePath = self.enEXE.get()
if not os.path.exists(exePath):
ocrToolPath = Config.get("ocrToolPath")
if not os.path.exists(ocrToolPath):
tk.messagebox.showerror(
'遇到了一点小问题', f'未在以下地址找到识别器！\n{exePath}')
'遇到了一点小问题', f'未在以下地址找到识别器！\n{ocrToolPath}')
return
# 创建输出文件
if self.isOutputFile.get() == 1:
outPath = self.enOutPath.get() + "\\" + self.enOutName.get()
if Config.get("isOutputFile"):
suffix = ".txt" if Config.get("outputStyle") == 1 else ".md"
outPath = Config.get("outputFilePath") + \
"\\" + Config.get("outputFileName")+suffix
try:
if os.path.exists(outPath): # 文件存在
os.remove(outPath) # 删除文件
@ -381,19 +486,48 @@ class Win:
async def run_(self): # 异步，执行任务
self.labelPercentage["text"] = "初始化"
isOutputFile = self.isOutputFile.get() # 是否输出文件
isOutputDebug = self.isOutputDebug.get() # 是否输出调试
areaInfo = self.areaInfo
isOutputFile = Config.get("isOutputFile") # 是否输出文件
isOutputDebug = Config.get("isOutputDebug") # 是否输出调试
isIgnoreNoText = Config.get("isIgnoreNoText") # 是否忽略无字图片
outputStyle = Config.get("outputStyle") # 输出风格
areaInfo = Config.get("ignoreArea")
if isOutputFile:
outPath = self.enOutPath.get() + "\\" + self.enOutName.get() # 输出文件路径
suffix = ".txt" if outputStyle == 1 else ".md"
outPath = Config.get("outputFilePath") + \
"\\" + Config.get("outputFileName")+suffix
def output(outStr, type_): # 输出字符串
"""
debug ：调试信息
text ：正文
name ：文件名
none ：不做修改
"""
# 写入输出面板，无需格式
self.textOutput.insert(tk.END, f"\n{outStr}\n")
if self.isAutoRoll.get(): # 需要自动滚动
self.textOutput.see(tk.END)
def output(str_): # 输出字符串
# 写入本地文件，按照格式
if isOutputFile:
if outputStyle == 1: # 纯文本风格
if type_ == "debug":
outStr = f"```\n{outStr}```\n"
elif type_ == "name":
outStr = f"\n\n≦ {outStr} ≧\n"
elif outputStyle == 2: # markdown风格
if type_ == "debug":
outStr = f"```\n{outStr}```\n"
elif type_ == "text":
outList = outStr.split("\n")
outStr = ""
for i in outList:
outStr += f"> {i} \n"
elif type_ == "name":
path = outStr.replace(" ", "%20")
outStr = f"---\n![{outStr}]({path})\n[{outStr}]({path})\n"
with open(outPath, "a", encoding='utf-8') as f: # 追加写入本地文件
f.write(str_)
self.textOutput.insert(tk.END, str_) # 1写入输出面板
if self.isAutoRoll.get(): # 需要自动滚动
self.textOutput.see(tk.END)
f.write(outStr)
def close(): # 关闭所有异步相关的东西
del self.ocr # 关闭OCR进程
@ -401,31 +535,35 @@ class Win:
self.setRunning(0)
self.labelPercentage["text"] = "已终止"
def getText(oget, img): # 分析一张图转出的文字
def analyzeText(oget, img): # 分析一张图转出的文字
def isInBox(aPos0, aPos1, bPos0, bPos1): # 检测框左上、右下角，待测者左上、右下角
return bPos0[0] >= aPos0[0] and bPos0[1] >= aPos0[1] and bPos1[0] <= aPos1[0] and bPos1[1] <= aPos1[1]
def isIden(): # 是否识别区域模式
if areaInfo[1][1]: # 需要检测
if areaInfo["area"][1]: # 需要检测
for o in oget: # 遍历每一个文字块
for a in areaInfo[1][1]: # 遍历每一个检测块
for a in areaInfo["area"][1]: # 遍历每一个检测块
if isInBox(a[0], a[1], (o["box"][0], o["box"][1]), (o["box"][4], o["box"][5])):
return True
text = ""
textDebug = "" # 调试信息
score = 0 # 平均置信度
scoreNum = 0
if not areaInfo or not areaInfo[0][0] == img["size"][0] or not areaInfo[0][1] == img["size"][1]:
# 无需忽略区域
if not areaInfo or not areaInfo["size"][0] == img["size"][0] or not areaInfo["size"][1] == img["size"][1]:
for i in oget:
text += i["text"]+"\n"
score += i["score"]
scoreNum += 1
# 判断，是忽略模式2
# 忽略模式2
elif isIden():
fn = 0 # 记录忽略的数量
for o in oget:
flag = True
for a in areaInfo[1][2]: # 遍历每一个检测块
for a in areaInfo["area"][2]: # 遍历每一个检测块
if isInBox(a[0], a[1], (o["box"][0], o["box"][1]), (o["box"][4], o["box"][5])):
flag = False # 踩到任何一个块，GG
break
@ -436,12 +574,14 @@ class Win:
else:
fn += 1
if isOutputDebug:
textDebug = f"〔忽略模式2：忽略{fn}条〕\n"
else: # 否则，忽略模式1
textDebug = f"忽略模式2：忽略{fn}条\n"
# 忽略模式1
else:
fn = 0 # 记录忽略的数量
for o in oget:
flag = True
for a in areaInfo[1][0]: # 遍历每一个检测块
for a in areaInfo["area"][0]: # 遍历每一个检测块
if isInBox(a[0], a[1], (o["box"][0], o["box"][1]), (o["box"][4], o["box"][5])):
flag = False # 踩到任何一个块，GG
break
@ -452,30 +592,29 @@ class Win:
else:
fn += 1
if isOutputDebug:
textDebug = f"〔忽略模式1：忽略{fn}条〕\n"
if text and not scoreNum == 0:
text = textDebug+text
textDebug = f"忽略模式1：忽略{fn}条\n"
if text and not scoreNum == 0: # 区域内有文本，计算置信度
score /= scoreNum
score = str(score)
score = str(score) # 转文本
else:
text = textDebug+"所有文字在忽略范围内\n"
score = "全部忽略"
return text, score
score = "1" # 区域内没有文本，置信度为1
return text, textDebug, score
# 开始
startStr = f"\n任务开始时间：{time.strftime('%Y-%m-%d %H:%M:%S', time.localtime(time.time()))}\n"
startStr = f"任务开始时间：{time.strftime('%Y-%m-%d %H:%M:%S', time.localtime(time.time()))}\n"
output(startStr, "text")
if isOutputDebug:
if areaInfo:
startStr += f"忽略区域：开启\n适用分辨率：{areaInfo[0]}\n"
startStr += f"忽略区域1：{areaInfo[1][0]}\n"
startStr += f"识别区域：{areaInfo[1][1]}\n"
startStr += f"忽略区域2：{areaInfo[1][2]}\n"
startStr = f'忽略区域：开启\n适用分辨率：{areaInfo["size"]}\n'
startStr += f'忽略区域1：{areaInfo["area"][0]}\n'
startStr += f'识别区域：{areaInfo["area"][1]}\n'
startStr += f'忽略区域2：{areaInfo["area"][2]}\n'
else:
startStr += f"忽略区域：关闭\n"
output(startStr)
startStr = f"忽略区域：关闭\n"
output(startStr, "debug")
# 创建OCR进程
exe = self.enEXE.get()
self.ocr = CallingOCR(exe)
self.ocr = CallingOCR(Config.get("ocrToolPath"))
# 初始化UI
for key in self.imgDict.keys(): # 清空表格参数
self.table.set(key, column='time', value="")
@ -512,9 +651,11 @@ class Win:
value=needTimeStr[:4]) # 时间写入表格
# 分析数据
dataStr = ""
textDebug = ""
if oget['code'] == 100: # 成功
numOK += 1
dataStr, score = getText(oget['data'], value) # 获取文字
dataStr, textDebug, score = analyzeText(
oget['data'], value) # 获取文字
elif oget['code'] == 101: # 无文字
numNON += 1
score = "无文字"
@ -523,28 +664,34 @@ class Win:
dataStr = "识别失败" # 不管开不开输出调试，都要输出报错
dataStr += f"，错误码：{oget['code']}\n错误信息：{str(oget['data'])}\n"
score = "失败"
writeStr = f'\n\n≦ {value["name"]} ≧\n'
if isOutputDebug: # 输出调试
writeStr += f"〔识别耗时：{needTimeStr}s 置信度：{score}〕\n"
writeStr += dataStr # 输出内容
self.table.set(key, column='score',
value=score[:4]) # 写入表格
output(writeStr)
# 写入表格
self.table.set(key, column='score', value=score[:4])
# 格式化输出
if isIgnoreNoText and not dataStr:
continue # 忽略无字图片
output(value["name"], "name")
if isOutputDebug:
output(
f"识别耗时：{needTimeStr}s 置信度：{score}\n{textDebug}", "debug")
output(dataStr, "text")
except Exception as e:
tk.messagebox.showerror(
'遇到了亿点小问题', f'图片识别异常：\n{value["name"]}\n异常信息：\n{e}')
# print(f'出问题了：{value["name"]}\n{e}')
# 结束
endTime = time.time()
endStr = f"\n\n\n任务结束时间：{time.strftime('%Y-%m-%d %H:%M:%S', time.localtime(endTime))}\n"
output("\n\n---\n", "none")
endStr = f"任务结束时间：{time.strftime('%Y-%m-%d %H:%M:%S', time.localtime(endTime))}\n"
output(endStr, "text")
if isOutputDebug:
endStr += f"任务耗时（秒）： {endTime-startTime}\n"
endStr += f"单张平均耗时： {(endTime-startTime)/allNum}\n"
endStr = f"任务耗时（秒）： {endTime-startTime}\n"
if not allNum == 0:
endStr += f"单张平均耗时： {(endTime-startTime)/allNum}\n"
endStr += f"共计图片数量： {numOK+numNON+numERR}\n"
endStr += f"识别正常 的图片数量： {numOK}\n"
endStr += f"未识别到文字 的图片数量：{numNON}\n"
endStr += f"识别失败 的图片数量： {numERR}\n\n"
output(endStr)
endStr += f"识别失败 的图片数量： {numERR}\n"
output(endStr, "debug")
close() # 完成后关闭
self.labelPercentage["text"] = "完成！"
@ -559,7 +706,7 @@ class Win:
if not tkinter.messagebox.askokcancel('提示', '将清空输出面板。要继续吗？'):
return
self.textOutput.delete('1.0', tk.END)
self.textOutput.insert(tk.END, getHelpText(ProjectWeb))
self.textOutput.insert(tk.END, GetHelpText(ProjectWeb))
def openProjectWeb(self, e=None): # 打开项目网页
webOpen(ProjectWeb)
@ -579,6 +726,10 @@ class Win:
else:
self.win.after(50, self.waitClose) # 等待关闭，50ms轮询一次是否已结束子进程
# def showTest(self):
# tk.messagebox.showwarning(
# "警告", "此功能尚未完工。")
if __name__ == "__main__":
Win()

Write Preview
Loading…
Cancel
Save