jdysya/Family-Tree
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

增加一些预处理的脚本

Browse Source
This commit is contained in:
高手 2025-03-27 18:37:31 +08:00
parent f5ef7e92ec
commit 866fa22461
4 changed files with 786 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

217
script/draw_relation.py Normal file
View File

@ -0,0 +1,217 @@
from graphviz import Digraph
"""
绘制家族树
"""
# 创建有向图
dot = Digraph(comment='家族树', node_attr={'fontname': 'WenQuanYi Zen Hei'},
edge_attr={'fontname': 'WenQuanYi Zen Hei'})
dot.attr(rankdir='TB', dpi='300')
# 重名
dot.node('1', "发昌")
dot.node('2', "发昌")
dot.node('3', "长松")
dot.node('4', "长松")
# 女性人物
female_nodes = ["德萍", "长香", "发双", "发娜", "发雅", "发桢", "发雯", "发莲", "发花", "发宇", "发洵", "其晞", "其园",
"其楠", "其晴", "其菲","长婷","长欢","长娟","长怡","发兰"]
# 定义父子关系
parent_child_relations = [
('可观', '光友'),
('光友', '世法'),
('世法', '道胜'), ('世法', '道启'), ('世法', '道佑'),
('道佑', '尚铎'), ('道佑', '尚助'), ('道佑', '尚达'), ('道佑', '尚朝'), ('道佑', '尚国'),
('尚铎', '士进'), ('尚铎', '士忠'), ('尚铎', '士富'),
('尚助', '士元'), ('尚助', '士怀'), ('尚助', '士纲'), ('尚助', '士位'), ('尚助', '士恺'),
('尚达', '士贵'),
('尚朝', '士恭'),
('尚国', '士宽'), ('尚国', '士信'), ('尚国', '士敏'), ('尚国', '士惠'),
('士进', '邦龙'), ('士进', '邦乐'),
('士忠', '邦志'), ('士忠', '邦正'),
('士富', '邦秀'),
('士怀', '邦朝'), ('士怀', '邦田'), ('士怀', '邦奇'), ('士怀', '邦学'), ('士怀', '邦强'),
('士纲', '邦彦'), ('士纲', '邦清'),
('士位', '邦宗'),
('士贵', '邦国'), ('士贵', '邦胜'),
('士恭', '邦应'), ('士恭', '邦全'),
('士信', '邦和'), ('士信', '邦云'),
('士敏', '邦元'),
('士惠', '邦霖'),
('邦乐', '载万'), ('邦乐', '载有'), ('邦乐', '载彦'),
('邦秀', '载禄'), ('邦秀', '载厚'), ('邦秀', '载元'),
('邦田', '载祥'), ('邦田', '载春'),
('邦彦', '载阳'),
('邦国', '载兴'), ('邦国', '载清'), ('邦国', '载荣'), ('邦国', '载祜'),
('邦胜', '载赓'), ('邦胜', '载歌'),
('邦应', '载福'), ('邦应', '载寿'), ('邦应', '载恩'),
('邦全', '载仁'),
('邦和', '载道'), ('邦和', '载富'),
('邦元', '载德'), ('邦元', '载义'), ('邦元', '载礼'),
('邦霖', '载让'), ('邦霖', '载成'), ('邦霖', '载宝'),
('载彦', '元年'), ('载彦', '元春'), ('载彦', '元长'),
('载阳', '元泰'), ('载阳', '元唱'), ('载阳', '元庆'), ('载阳', '元魁'), ('载阳', '元丰'),
('载清', '元炳'), ('载清', '元济'), ('载清', '元勋'), ('载清', '元功'), ('载清', '元修'),
('载荣', '元鼎'),
('载荣', '元钟'),
('载荣', '元彛'),
('载荣', '元琛'),
('载祜', '元抡'), ('载祜', '元会'), ('载祜', '元松'),
('载赓', '元良'),
('载歌', '元恺'), ('载歌', '元选'), ('载歌', '元举'), ('载歌', '元宪'),
('载福', '元臣'), ('载福', '元定'), ('载福', '元顺'),
('载恩', '元凤'), ('载恩', '元龙'),
('载仁', '元勤'), ('载仁', '元俭'), ('载仁', '元学'),
('载道', '元端'), ('载道', '元章'),
('载让', '元甫'),
('元春', '善全'),
('元唱', '善怀'), ('元唱', '善元'),
('元庆', '善良'), ('元庆', '善和'),
('元丰', '善琪'),
('元炳', '善辅'), ('元炳', '善举'), ('元炳', '善传'), ('元炳', '善富'),
('元勋', '善宏'),
('元修', '善行'), ('元修', '善言'),
('元抡', '善彰'), ('元抡', '善扬'),
('元会', '善藏'), ('元会', '善身'), ('元会', '善维'), ('元会', '善宝'),
('元良', '善述'),
('元恺', '善荫'),
('元举', '善炽'), ('元举', '善祯'), ('元举', '善进'), ('元举', '善从'), ('元举', '善持'),
('元臣', '善道'), ('元臣', '善为'), ('元臣', '善庆'), ('元臣', '善保'), ('元臣', '善守'),
('元学', '善祥'), ('元学', '善士'),
('善全', '德昌'),
('善怀', '德大'), ('善怀', '德恒'),
('善元', '德安'), ('善元', '德溥'),
('善良', '德太'), ('善良', '德定'), ('善良', '德友'), ('善良', '德仁'),
('善和', '德文'), ('善和', '德武'),
('善琪', '德金'), ('善琪', '德艳'), ('善琪', '德映'),
('善辅', '德远'), ('善辅', '德麟'),
('善传', '德长'),
('善富', '德兴'), ('善富', '德海'), ('善富', '德怀'), ('善富', '德周'), ('善富', '德齐'),
('善宏', '德政'),
('善行', '德明'), ('善行', '德辅'),
('善言', '德庆'), ('善言', '德广'), ('善言', '德林'),
('善扬', '德敬'), ('善扬', '德钰'),
('善身', '德备'), ('善身', '德胜'),
('善维', '德峰'), ('善维', '德萍'),
('善宝', '德忠'), ('善宝', '德成'),
('善述', '德聘'), ('善述', '德炎'), ('善述', '德种'),
('善进', '德元'), ('善进', '德利'),
('善持', '德福'), ('善持', '德泳'),
('德昌', '长辉'), ('德昌', '长怡'),
('德安', '长江'), ('德安', '长恩'), ('德安', '长有'),
('德定', '长泳'),
('德仁', '长明'), ('德仁', '长伦'),
('德文', '长元'), ('德文', '长绵'), ('德文', '长新'), ('德文', '长生'),
('德艳', '长益'),
('德映', '长威'), ('德映', '长翔'),
('德远', '3'), ('德远', '长青'), ('德远', '长学'),
('德兴', '长存'), ('德兴', '长友'), ('德兴', '长胜'),
('德海', '长文'), ('德海', '长峰'),
('德庆', '长润'), ('德庆', '长柏'),
('德广', '长华'), ('德广', '长国'),
('德林', '长旺'), ('德林', '长平'),
('德敬', '4'), ('德敬', '长征'),
('德胜', '长宏'), ('德胜', '长香'),
('德峰', '长婷'),
('德萍', '长伟'),
('德忠', '长磊'), ('德忠', '长欢'),
('德成', '长乐'), ('德成', '长娟'),
('德炎', '长福'), ('德炎', '长贵'), ('德炎', '长安'), ('德炎', '长寿'),
('德种', '长林'),
('德元', '长刚'), ('德元', '长龙'),
('长江', '发双'),
('长有', '发逸'), ('长有', '发娜'),
('长泳', '发瑞'), ('长泳', '发雅'),
('长明', '发东'),
('长绵', '发旺'), ('长绵', '发辉'),
('长新', '发洋'),
('长生', '发桢'),
('长威', '发雯'),
('长青', '发良'), ('长青', '发莲'),
('长友', '发家'), ('长友', '发花'),
('长胜', '发传'), ('长胜', '发涛'),
('长润', '发翔'),
('长柏', '发波'),
('长华', '发松'),
('长旺', '1'),
('长平', '发鹏'), ('长平', '发磊'),
('长香', '2'), ('长香', '发广'), ('长香', '发宇'),
('长伟', '发洵'),
('长福', '发启'), ('长福', '发明'),
('长贵', '发文'), ('长贵', '发祥'),
('长寿', '发平'),
('长林', '发義'), ('长林', '发兰'),
('发良', '其非'), ('发良', '其凡'),
('发翔', '其晞'),
('发文', '其旺'), ('发文', '其园'),
('发祥', '其达'), ('发祥', '其楠'),
('发平', '其红'), ('发平', '其晴'), ('发平', '其菲'),
('发義', '其昊')
]
# 定义过继关系
adoption_relations = [
('士宽', '邦云'),
('邦朝', '载祥'),
('邦奇', '载祥'),
('邦学', '载春'),
('邦强', '载春'),
('邦清', '载阳'),
('邦宗', '载阳'),
('邦宗', '元魁'),
('邦清', '元庆'),
('邦云', '载富'),
('载有', '元俭'),
('载禄', '元俭'),
('载元', '元俭'),
('载祥', '元丰'),
('载春', '元丰'),
('载兴', '元济'),
('载富', '元章'),
('元俭', '善和'),
('元济', '善传'),
('元鼎', '善言'),
('元钟', '善言'),
('元宪', '善祯'),
('元顺', '善为'),
('元松', '善宝'),
('善举', '德兴'),
('善彰', '德敬'),
('善荫', '德种'),
('德武', '长新'),
('德政', '长青'),
('德齐', '长胜')
]
# 添加节点和父子关系边
# for parent, child in parent_child_relations:
# dot.node(parent)
# dot.node(child)
# dot.edge(parent, child)
# 添加节点和父子关系边
for parent, child in parent_child_relations:
if parent in female_nodes:
dot.node(parent, shape='ellipse', style='filled', color='pink')
else:
dot.node(parent, shape='box', style='filled', color='lightblue')
if child in female_nodes:
dot.node(child, shape='ellipse', style='filled', color='pink')
else:
dot.node(child, shape='box', style='filled', color='lightblue')
dot.edge(parent, child)
# 添加过继关系边
for adoptive_parent, adoptee in adoption_relations:
dot.node(adoptive_parent)
# dot.node(adoptee)
dot.node(adoptee, shape='diamond', style='filled', color='lightgreen')
dot.edge(adoptive_parent, adoptee, style='dashed', label='继', color="blue")
# 渲染图形
dot.render('family_tree.gv', view=True)

44
script/extract_images.py Normal file
View File

@ -0,0 +1,44 @@
import fitz # PyMuPDF
import os
from PIL import Image
import io
"""
提取PDF中的图片
"""
def extract_images_from_pdf(pdf_path, output_folder):
# 打开PDF文件
document = fitz.open(pdf_path)
# 确保输出文件夹存在
os.makedirs(output_folder, exist_ok=True)
# 遍历每一页
for page_number in range(len(document)):
page = document.load_page(page_number)
# 获取页面中的图片
images = page.get_images(full=True)
for image_index, img in enumerate(images):
xref = img[0]
# 提取图片
base_image = document.extract_image(xref)
image_bytes = base_image["image"]
image_ext = base_image["ext"]
# 使用Pillow处理图片
image = Image.open(io.BytesIO(image_bytes))
# 构建图片文件名
image_filename = f"page_{page_number + 1}_img_{image_index + 1}.{image_ext}"
image_path = os.path.join(output_folder, image_filename)
# 保存图片
image.save(image_path)
print(f"保存图片: {image_path}")
# 关闭文档
document.close()
# 使用示例
pdf_path = "/Users/xiangyu/Documents/余氏宗谱（新洲区等支族）/第一册/（5）余氏彩页 P17-40.pdf" # 替换为你的PDF文件路径
output_folder = "extracted_images" # 替换为你想要保存图片的文件夹
extract_images_from_pdf(pdf_path, output_folder)

482
script/pdf_text_extract.py Normal file
View File

@ -0,0 +1,482 @@
import cv2
import numpy as np
import fitz # PyMuPDF
import tkinter as tk
from tkinter import ttk
from PIL import Image, ImageTk
import tkinter.messagebox as messagebox
import datetime
"""
提取PDF中的竖排文本(针对分隔区域)
"""
class PDFProcessor:
def __init__(self):
self.image = None
self.regions = []
def convert_pdf_to_image(self, pdf_path, page_num=0):
# 打开PDF文件
doc = fitz.open(pdf_path)
page = doc[page_num]
# 将PDF页面转换为图片
pix = page.get_pixmap(matrix=fitz.Matrix(2, 2)) # 2倍缩放以获得更好的质量
img = Image.frombytes("RGB", [pix.width, pix.height], pix.samples)
# 转换为OpenCV格式
self.image = cv2.cvtColor(np.array(img), cv2.COLOR_RGB2BGR)
return self.image
def detect_regions(self):
if self.image is None:
return []
# 转换为灰度图
gray = cv2.cvtColor(self.image, cv2.COLOR_BGR2GRAY)
# 使用高斯模糊减少噪声
blurred = cv2.GaussianBlur(gray, (5, 5), 0)
# 自适应阈值处理
binary = cv2.adaptiveThreshold(
blurred,
255,
cv2.ADAPTIVE_THRESH_GAUSSIAN_C,
cv2.THRESH_BINARY_INV,
11,
2
)
# 膨胀操作
kernel = np.ones((3,3), np.uint8)
dilated = cv2.dilate(binary, kernel, iterations=1)
# 查找所有轮廓
contours, hierarchy = cv2.findContours(
dilated,
cv2.RETR_TREE,
cv2.CHAIN_APPROX_SIMPLE
)
# 收集所有可能的矩形区域
candidates = []
min_area = 1000
for cnt in contours:
peri = cv2.arcLength(cnt, True)
approx = cv2.approxPolyDP(cnt, 0.02 * peri, True)
x, y, w, h = cv2.boundingRect(cnt)
area = w * h
aspect_ratio = w / float(h) if h != 0 else 0
if (area > min_area and
w > 30 and h > 30 and
0.1 < aspect_ratio < 10):
if len(approx) >= 4:
candidates.append((x, y, w, h))
# 按宽度从大到小排序
candidates.sort(key=lambda x: x[2], reverse=True)
# 查找宽高一致的五个区域
target_regions = []
for i, (x1, y1, w1, h1) in enumerate(candidates):
similar_regions = [(x1, y1, w1, h1)]
# 在剩余区域中查找宽高相似的区域
for x2, y2, w2, h2 in candidates[i+1:]:
# 检查宽高是否相似（允许2%的误差）
if (abs(w2 - w1) / w1 < 0.02 and
abs(h2 - h1) / h1 < 0.02):
similar_regions.append((x2, y2, w2, h2))
# 如果找到至少5个相似区域
if len(similar_regions) >= 5:
# 按y坐标排序（从上到下）
similar_regions.sort(key=lambda x: x[1])
target_regions = similar_regions[:5]
break
self.regions = target_regions
return self.regions
class PDFViewer:
def __init__(self, root):
self.root = root
self.root.title("PDF区域检测器")
# 设置窗口初始大小
self.root.geometry("1024x768")
# 创建主框架
self.main_frame = ttk.Frame(self.root)
self.main_frame.pack(fill=tk.BOTH, expand=True)
# 创建工具栏
self.toolbar = ttk.Frame(self.main_frame)
self.toolbar.pack(side=tk.TOP, fill=tk.X)
# 添加缩放控制
self.zoom_label = ttk.Label(self.toolbar, text="缩放: ")
self.zoom_label.pack(side=tk.LEFT, padx=5)
self.zoom_scale = ttk.Scale(
self.toolbar,
from_=10,
to=200,
orient=tk.HORIZONTAL,
length=200,
command=self._on_zoom_change
)
self.zoom_scale.set(30) # 设置初始缩放为30%
self.zoom_scale.pack(side=tk.LEFT, padx=5)
# 创建框架来容纳画布和滚动条
self.frame = ttk.Frame(self.main_frame)
self.frame.pack(fill=tk.BOTH, expand=True)
# 创建水平和垂直滚动条
self.v_scrollbar = ttk.Scrollbar(self.frame, orient="vertical")
self.h_scrollbar = ttk.Scrollbar(self.frame, orient="horizontal")
self.v_scrollbar.pack(side=tk.RIGHT, fill=tk.Y)
self.h_scrollbar.pack(side=tk.BOTTOM, fill=tk.X)
# 创建画布并配置滚动
self.canvas = tk.Canvas(
self.frame,
yscrollcommand=self.v_scrollbar.set,
xscrollcommand=self.h_scrollbar.set,
bg='gray90' # 添加背景色以便于区分
)
self.canvas.pack(side=tk.LEFT, fill=tk.BOTH, expand=True)
# 配置滚动条
self.v_scrollbar.config(command=self.canvas.yview)
self.h_scrollbar.config(command=self.canvas.xview)
# 绑定鼠标事件
self.canvas.bind('<MouseWheel>', self._on_mousewheel_y)
self.canvas.bind('<Shift-MouseWheel>', self._on_mousewheel_x)
self.canvas.bind('<Button-4>', self._on_mousewheel_y)
self.canvas.bind('<Button-5>', self._on_mousewheel_y)
self.canvas.bind('<Shift-Button-4>', self._on_mousewheel_x)
self.canvas.bind('<Shift-Button-5>', self._on_mousewheel_x)
self.canvas.bind('<ButtonPress-1>', self._start_drag)
self.canvas.bind('<B1-Motion>', self._drag)
# 添加Ctrl+鼠标滚轮缩放
self.canvas.bind('<Control-MouseWheel>', self._on_zoom_wheel)
self.processor = PDFProcessor()
self.current_image = None
self.current_regions = None
self.zoom_factor = 0.3 # 改为30%的初始缩放
# 添加适应窗口大小的按钮
self.fit_button = ttk.Button(
self.toolbar,
text="适应窗口",
command=self._fit_to_window
)
self.fit_button.pack(side=tk.LEFT, padx=5)
# 绑定窗口大小改变事件
self.root.bind('<Configure>', self._on_window_resize)
# 添加页面控制
self.page_frame = ttk.Frame(self.toolbar)
self.page_frame.pack(side=tk.LEFT, padx=5)
self.page_label = ttk.Label(self.page_frame, text="页码：")
self.page_label.pack(side=tk.LEFT)
self.current_page = tk.StringVar(value="1")
self.total_pages = 1
self.page_entry = ttk.Entry(self.page_frame, textvariable=self.current_page, width=5)
self.page_entry.pack(side=tk.LEFT, padx=2)
self.total_pages_label = ttk.Label(self.page_frame, text="/1")
self.total_pages_label.pack(side=tk.LEFT)
self.prev_button = ttk.Button(self.page_frame, text="上一页", command=self._prev_page)
self.prev_button.pack(side=tk.LEFT, padx=2)
self.next_button = ttk.Button(self.page_frame, text="下一页", command=self._next_page)
self.next_button.pack(side=tk.LEFT, padx=2)
# 修改确认按钮文本
self.confirm_button = ttk.Button(
self.toolbar,
text="提取所有页面文字",
command=self._extract_text
)
self.confirm_button.pack(side=tk.LEFT, padx=5)
self.pdf_path = None
self.doc = None
# 加载默认PDF文件
default_pdf = "test.pdf"
try:
self.load_pdf(default_pdf)
except Exception as e:
messagebox.showerror("错误", f"无法加载默认PDF文件：{str(e)}")
def load_pdf(self, pdf_path):
"""加载PDF并显示"""
self.pdf_path = pdf_path
self.doc = fitz.open(pdf_path)
self.total_pages = len(self.doc)
self.total_pages_label.configure(text=f"/{self.total_pages}")
self._load_current_page()
def _load_current_page(self):
"""加载当前页面"""
try:
page_num = int(self.current_page.get()) - 1
if 0 <= page_num < self.total_pages:
image = self.processor.convert_pdf_to_image(self.pdf_path, page_num)
regions = self.processor.detect_regions()
self.current_image = image
self.current_regions = regions
self._update_display()
except ValueError:
messagebox.showerror("错误", "请输入有效的页码")
def _prev_page(self):
"""显示上一页"""
try:
current = int(self.current_page.get())
if current > 1:
self.current_page.set(str(current - 1))
self._load_current_page()
except ValueError:
pass
def _next_page(self):
"""显示下一页"""
try:
current = int(self.current_page.get())
if current < self.total_pages:
self.current_page.set(str(current + 1))
self._load_current_page()
except ValueError:
pass
def _update_display(self):
"""更新显示"""
if self.current_image is None:
return
# 获取原始尺寸
height, width = self.current_image.shape[:2]
# 计算缩放后的尺寸
new_width = int(width * self.zoom_factor)
new_height = int(height * self.zoom_factor)
# 缩放图片
resized_image = cv2.resize(self.current_image, (new_width, new_height))
# 转换并显示图片
image = cv2.cvtColor(resized_image, cv2.COLOR_BGR2RGB)
image_pil = Image.fromarray(image)
self.photo = ImageTk.PhotoImage(image=image_pil)
# 清除画布
self.canvas.delete("all")
# 显示图片
self.canvas.create_image(0, 0, image=self.photo, anchor=tk.NW)
# 绘制区域
if self.current_regions:
for i, (x, y, w, h) in enumerate(self.current_regions):
# 缩放坐标和尺寸
scaled_x = int(x * self.zoom_factor)
scaled_y = int(y * self.zoom_factor)
scaled_w = int(w * self.zoom_factor)
scaled_h = int(h * self.zoom_factor)
rect_id = self.canvas.create_rectangle(
scaled_x, scaled_y,
scaled_x + scaled_w,
scaled_y + scaled_h,
outline="red",
width=max(1, int(2 * self.zoom_factor)),
tags=f"region_{i+1}"
)
self.canvas.create_text(
scaled_x + scaled_w//2,
scaled_y - 10 * self.zoom_factor,
text=f"目标区域 {i+1} ({w}x{h})",
fill="red",
tags=f"region_{i+1}"
)
# 绑定鼠标事件
self.canvas.tag_bind(
f"region_{i+1}",
'<Enter>',
lambda e, rid=rect_id: self._highlight_region(rid)
)
self.canvas.tag_bind(
f"region_{i+1}",
'<Leave>',
lambda e, rid=rect_id: self._unhighlight_region(rid)
)
# 更新画布滚动区域
self.canvas.config(scrollregion=self.canvas.bbox(tk.ALL))
def _on_zoom_wheel(self, event):
"""处理Ctrl+滚轮缩放"""
if event.delta > 0:
self.zoom_scale.set(min(200, self.zoom_scale.get() + 10))
else:
self.zoom_scale.set(max(10, self.zoom_scale.get() - 10))
def _on_zoom_change(self, value):
"""处理缩放变化"""
self.zoom_factor = float(value) / 100
self._update_display()
def _on_mousewheel_y(self, event):
"""处理垂直方向的鼠标滚轮事件"""
if event.num == 4 or event.delta > 0:
self.canvas.yview_scroll(-1, "units")
elif event.num == 5 or event.delta < 0:
self.canvas.yview_scroll(1, "units")
def _on_mousewheel_x(self, event):
"""处理水平方向的鼠标滚轮事件"""
if event.num == 4 or event.delta > 0:
self.canvas.xview_scroll(-1, "units")
elif event.num == 5 or event.delta < 0:
self.canvas.xview_scroll(1, "units")
def _start_drag(self, event):
"""开始拖动"""
self.canvas.scan_mark(event.x, event.y)
def _drag(self, event):
"""拖动画布"""
self.canvas.scan_dragto(event.x, event.y, gain=1)
def _highlight_region(self, region_id):
"""高亮显示区域"""
self.canvas.itemconfig(region_id, width=max(2, int(3 * self.zoom_factor)), outline="yellow")
def _unhighlight_region(self, region_id):
"""取消高亮显示"""
self.canvas.itemconfig(region_id, width=max(1, int(2 * self.zoom_factor)), outline="red")
def _fit_to_window(self):
"""调整缩放以适应窗口大小"""
if self.current_image is None:
return
# 获取窗口和图像尺寸
window_width = self.canvas.winfo_width()
window_height = self.canvas.winfo_height()
image_height, image_width = self.current_image.shape[:2]
# 计算合适的缩放比例
width_ratio = window_width / image_width
height_ratio = window_height / image_height
# 选择较小的比例以确保完全显示
new_zoom = min(width_ratio, height_ratio) * 0.9 # 留出一些边距
# 更新缩放
self.zoom_scale.set(new_zoom * 100)
def _on_window_resize(self, event):
"""窗口大小改变时的处理"""
# 仅当事件来自主窗口时才处理
if event.widget == self.root:
# 更新画布滚动区域
self.canvas.config(scrollregion=self.canvas.bbox(tk.ALL))
def _extract_text(self):
"""提取所有页面的文字并分别保存到五个文件"""
if not self.doc:
messagebox.showwarning("警告", "请先加载PDF文件！")
return
try:
# 创建五个输出文件（使用追加模式）
base_path = self.pdf_path.rsplit('.', 1)[0]
output_files = [open(f"{base_path}_region{i+1}.txt", 'a', encoding='utf-8') for i in range(5)]
# 写入分隔符（带时间戳）
timestamp = datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S")
for i, f in enumerate(output_files):
f.write(f"\n\n=== 新的提取任务 {timestamp} ===\n\n")
# 处理每一页
for page_num in range(self.total_pages):
# 更新当前页面以获取区域
self.current_page.set(str(page_num + 1))
self._load_current_page()
if not self.current_regions or len(self.current_regions) < 5:
messagebox.showwarning("警告", f"第 {page_num+1} 页未检测到足够区域，跳过该页")
continue
page = self.doc[page_num]
# 按y坐标排序处理每个区域
sorted_regions = sorted(self.current_regions, key=lambda x: x[1])
for region_idx, region in enumerate(sorted_regions[:5]): # 只取前五个区域
x, y, w, h = region
# 将OpenCV的坐标转换为PDF坐标（除以2因为之前放大了2倍）
pdf_x = x / 2
pdf_y = y / 2
pdf_w = w / 2
pdf_h = h / 2
rect = fitz.Rect(pdf_x, pdf_y, pdf_x + pdf_w, pdf_y + pdf_h)
text = page.get_text(clip=rect, sort=True)
if text.strip():
output_files[region_idx].write(
f"=== 第 {page_num + 1} 页 ===\n"
f"区域坐标: ({x}, {y}) 尺寸: {w}x{h}\n"
f"{text}\n"
"-------------------\n\n"
)
# 关闭所有文件
for f in output_files:
f.close()
messagebox.showinfo("成功", f"文字已分别保存到：\n"
f"{base_path}_region1.txt\n"
f"{base_path}_region2.txt\n"
f"{base_path}_region3.txt\n"
f"{base_path}_region4.txt\n"
f"{base_path}_region5.txt")
except Exception as e:
messagebox.showerror("错误", f"提取文字时发生错误：{str(e)}")
finally:
# 确保文件被关闭
for f in output_files:
if not f.closed:
f.close()
def main():
root = tk.Tk()
app = PDFViewer(root)
root.mainloop()
if __name__ == "__main__":
main()

43
script/pdf_whole_text_extract.py Normal file
View File

@ -0,0 +1,43 @@
import fitz # PyMuPDF
"""
提取PDF中的竖排文本(针对整页文本)
"""
def extract_vertical_text(pdf_path):
"""
提取竖排文本（从右至左阅读顺序）并包含页码信息
"""
doc = fitz.open(pdf_path)
full_text = []
for page in doc:
# 添加页码标识
full_text.append(f"=== 第{page.number + 1}页 ===")
# 旋转页面以适应竖排文本阅读方向
page.set_rotation(270)
# 获取并排序文本块
blocks = page.get_text("blocks", flags=fitz.TEXT_PRESERVE_LIGATURES)
blocks.sort(key=lambda b: (-b[2], b[1]))
# 收集当前页文本
page_text = []
for b in blocks:
text = b[4].strip()
if text:
page_text.append(text)
full_text.append('\n'.join(page_text))
return '\n\n'.join(full_text)
# 修改后的使用示例
if __name__ == "__main__":
import sys
text = extract_vertical_text('/Users/xiangyu/Documents/余氏宗谱（新洲区等支族）/第一册/（6）余氏宗谱 （四諫堂） 卷之首一 （三校）.pdf')
# 写入文件
with open('output.txt', 'w', encoding='utf-8') as f:
f.write(text)
print("文本已保存到 output.txt")