def parse_multicolumn(cls, s):
r"""找出s中第一次出现的满足模式的multicolumn，返回3个关键值
:returns:
第1个参数是该合并单元格的尺寸，固定格式： (行数, 列数)，只有一行是也会写'1'
>>> TexTabular.parse_multicolumn(r'\multicolumn{2}{|c|}{aa\multirow{3}*{特点}bb}')
((3, 2), '|c|', 'aa特点bb')
"""
# 1 基本的模式匹配抓取
m = re.search(r'\\multicolumn' + grp_bracket(3, inner=True) * 2
+ grp_bracket(5, inner=True), s) # 最后层多套下，我怕不够用
if not m: return None
# 2 取出参数值
m, col_align, text = m.groups()
m = int(m)
# 3 如果有 multirow
if 'multirow' in text:
n, bigstructs, width, fixup, text = cls.parse_multirow(text, brace_text_only=False)
else:
n = 1
# if isinstance(text, str) and re.match(r'\d+$', text): text = int(text) # 如果可以，直接识别为数字
return (n, m), col_align, text
def grp_figure(cnt_groups=0, parpic=False):
"""生成跟图片匹配相关的表达式
D:\2017LaTeX\D招培试卷\高中地理，用过 \captionfig{3-3.eps}{图~3}
奕本从2018秋季教材开始使用多种图片格式
191224周二18:20 更新：匹配到的图片名不带花括号
"""
ibrace3 = grp_bracket(3, inner=True)
if cnt_groups == 0: # 不分组
s = r'\\(?:includegraphics|figt|figc|figr|fig).*?' + grp_bracket(
3) # 注意第1组fig要放最后面
elif cnt_groups == 1: # 只分1组，那么只对图片括号内的内容分组
s = r'\\(?:includegraphics|figt|figc|figr|fig).*?' + ibrace3
elif cnt_groups == 2: # 只分2组，那么只对插图命令和图片分组
s = r'\\(includegraphics|figt|figc|figr|fig).*?' + ibrace3
elif cnt_groups == 3:
s = r'\\(includegraphics|figt|figc|figr|fig)(.*?)' + ibrace3
else:
s = None
if s and parpic:
s = r'{?\\parpic(?:\[.\])?{' + s + r'}*'
return s
def parse_align(cls, s):
r"""解析latex表头的列对齐格式
latex表头的规则很复杂，这里目前只处理一些较常用的功能点
:param s: 内容文本
:return: 不考虑竖线和一些高级对齐格式，暂时返回一个str
长度是表格列数，每个元素是一个字母存储对齐信息（后续可以扩展更细致的对齐格式信息）
>>> TexTabular.parse_align('{|c|c|c|c|c|c|c|c|c<{}|c|}')
'cccccccccc'
>>> TexTabular.parse_align('{|c|w{6em}|w{23mm}|w{47mm}|w{22mm}|}')
'cwwww'
>>> TexTabular.parse_align('cc*{8}{l}')
'ccllllllll'
>>> TexTabular.parse_align('|c|')
'c'
>>> TexTabular.parse_align('|c|*{2}{m{38mm}<{\\centering}|}')
'cmm'
"""
# 展开 *{n}{列格式} 模式
s = re.sub(r'\*(\d+)', r'*{\1}', s) # 给*数字加上花括号，不然我的匹配会错
s = re.sub(r'\*{(\d+)}' + grp_bracket(3, inner=True), lambda m: m.group(2) * int(m.group(1)), s)
# 删除其他干扰字符
if s[0] == '{' and s[-1] == '}': s = s[1:-1] # 去掉头尾 { }
s = re.sub(r'{.*?}', '', s)
for char in '|<>!':
s = s.replace(char, '')
return s
def core(s):
if inner: # 只取\lewis{}花括号里内容的定位
raise ValueError(
r"lewis模式没有inner模式，如果需要可以使用NestEnv(s).inner(r'\lewis{')")
lewis = r'\\(l|L)ewis' + grp_bracket(5, inner=True) # 基本匹配模式
ms = re.finditer(rf'(H?~*{lewis}\s*|~*H)*(~*{lewis}|~*H)',
s) # 有一定的延展
return [m.span(0) for m in ms if 'lewis' in m.group().lower()]
def brieftexstr(s):
"""对比两段tex文本
"""
# 1 删除百分注
s = re.sub(r'%' + grp_bracket(2, '<', '>'), r'', s)
# 2 删除所有空白字符
# debuglib.dprint(debuglib.typename(s))
s = re.sub(r'\s+', '', s)
# 3 转小写字符
s = s.casefold()
return s
def core(s):
right, parts = 0, []
while True:
m0 = re.search(r'\\(' + name + r')(?![a-zA-Z])', s[right:])
if not m0: break
left, right = m0.start() + right, m0.end() + right
if star:
m1 = re.match(r'(\s*)(\*)', s[right:])
if m1 and m1.group(1).count('\n') <= linefeed and m1.group(
2):
right += m1.end()
if optional:
m2 = re.match(r'(\s*)(' + grp_bracket(5, '[') + ')',
s[right:])
if m2 and m2.group(1).count('\n') <= linefeed and m2.group(
2):
right += m2.end()
cur_cnt, pattern = 0, r'(\s*)(' + grp_bracket(5) + ')'
max_bracket_ = max_bracket
if max_bracket == float('inf'):
if m0.group(1) in ('begin', 'end'):
max_bracket_ = 1 # 有些命令只能匹配一个花括号
if m0.group(1) in ('hfil', 'hfill'):
max_bracket_ = 0 # 有些命令不能匹配花括号
while cur_cnt < max_bracket_:
m3 = re.match(pattern, s[right:])
if m3 and m3.group(1).count('\n') <= linefeed and m3.group(
2):
right += m3.end()
cur_cnt += 1
else:
break
if cur_cnt >= min_bracket:
parts.append([left, right])
return parts
def parse_multirow(cls, s, brace_text_only=True):
r"""
:param brace_text_only: 只取花括号里面的内容
如果为False，会把multirow外部的内容做拼接
multirow 和 multicolumn 的不同是，第1、2个花括号后面可以有可选参数。
第2个花括号如果内容是*，可以省略。
两个[]的内容先省略，不做分析处理
注意：这里会取出前后缀内容！业务需要，防止bug，不过这种概率很小
>>> TexTabular.parse_multirow(r'\multirow{2}*{特点}')
(2, None, '*', None, '特点')
>>> TexTabular.parse_multirow(r'\multirow{2}{*}{特点}')
(2, None, '*', None, '特点')
>>> TexTabular.parse_multirow(r'aa\multirow{2}[bla1]{*}[bla2]{特点}bb', brace_text_only=False)
(2, 'bla1', '*', 'bla2', 'aa特点bb')
TODO multirow第一个数字是可以负值的，代表向上合并单元格数，
"""
square = r'(?:\[(.*?)\])?' # 可选参数
m = re.search(r'\\multirow' + grp_bracket(3, inner=True) + square +
r'(?:{(.*?)}|(\*))' + square + grp_bracket(5, inner=True), s)
if not m: return None
n, bigstructs, width1, width2, fixup, text = m.groups()
width = width1 or width2
if not brace_text_only: text = s[:m.start()] + text + s[m.end():]
# if re.match(r'\d+$', text): text = int(text) # 如果可以，直接识别为数字
n = int(n)
if -1 <= n <= 1:
n = 1
elif n > 1:
pass
else:
raise ValueError(f'{s} 不支持解析multirow第一个值为负数，向上合并单元格的情况')
return n, bigstructs, width, fixup, text
def init_from_latex(ws, content):
""" 注意没有取名为from_latex，因为ws是事先创建好的，这里只是能输入latex代码进行初始化而已 """
from openpyxl.styles import Border, Alignment, Side
from pyxllib.text.pupil import grp_bracket
from pyxllib.text.latex import TexTabular
BRACE2 = grp_bracket(2, inner=True)
BRACE5 = grp_bracket(5, inner=True)
# 暂时统一边框线的样式 borders。不做细化解析
double = Side(border_style='thin', color='000000')
# 处理表头
data_tex = re.search(r'\\begin{tabular}\s*(?:\[.*?\])?\s*' + BRACE5,
content).group(1)
col_pos = TexTabular.parse_align(data_tex) # 每列的格式控制
# dprint(self.data_tex, col_pos)
total_col = len(col_pos)
# 删除头尾标记
s = re.sub(r'\\begin{tabular}(?:\[.*?\])?' + BRACE5, '',
re.sub(r'\\end{tabular}', '', content))
row, col = 1, 1
# 先用简单不严谨的规则确定用全网格，还是无网格
# if '\\hline' not in s and '\\midrule' not in s:
# border = 0
# 用 \\ 分割处理每一行
for line in re.split(r'\\\\(?!{)', s)[:-1]:
# dprint(line)
# 1 处理当前行的所有列元素
cur_line = line
# dprint(line)
# 清除特殊格式数据
cur_line = re.sub(r'\\cmidrule' + BRACE2, '', cur_line)
cur_line = re.sub(r'\\cline' + BRACE2, '', cur_line)
for t in (r'\midrule', r'\toprule', r'\bottomrule', r'\hline',
'\n'):
cur_line = cur_line.replace(t, '')
# 遍历每列
# dprint(cur_line)
for item in cur_line.strip().split('&'):
item = item.strip()
# dprint(item)
cur_loc = excel_addr(row, col)
# dprint(row, col)
if 'multicolumn' in item:
size, align, text = TexTabular.parse_multicolumn(item)
align = TexTabular.parse_align(
align) if align else col_pos[col -
1] # 如果没有写对齐，用默认列的格式
n, m = size
# 左右对齐，默认是left
align = {
'l': 'left',
'c': 'center',
'r': 'right'
}.get(align, 'left')
cell = ws[cur_loc].mcell()
if cell.value:
cell.value += '\n' + text
else:
cell.value = text
ws[cur_loc].alignment = Alignment(horizontal=align,
vertical='center')
merge_loc = excel_addr(row + n - 1, col + m - 1)
ws.merge_cells(f'{cur_loc}:{merge_loc}')
col += m
elif 'multirow' in item:
n, bigstructs, width, fixup, text = TexTabular.parse_multirow(
item, brace_text_only=False)
try:
ws[cur_loc] = text
except AttributeError:
# 遇到合并单元格重叠问题，就修改旧的合并单元格，然后添加新单元格
# 例如原来 A1:A3 是一个合并单元格，现在要独立一个A3，则原来的部分重置为A1:A2
rng = ws[cur_loc].in_range()
ws.unmerge_cells(rng.coord) # 解除旧的合并单元格
ws.merge_cells(re.sub(r'\d+$', f'{row - 1}',
rng.coord))
ws[cur_loc] = text
align = {
'l': 'left',
'c': 'center',
'r': 'right'
}.get(col_pos[col - 1], 'left')
ws[cur_loc].alignment = Alignment(horizontal=align,
vertical='center')
# dprint(item, row, n)
merge_loc = excel_addr(row + n - 1, col)
ws.merge_cells(f'{cur_loc}:{merge_loc}')
col += 1
else:
if ws[cur_loc].celltype() == 0:
ws[cur_loc].value = item
# dprint(item, col_pos, col)
align = {
'l': 'left',
'c': 'center',
'r': 'right'
}.get(col_pos[col - 1], 'left')
ws[cur_loc].alignment = Alignment(horizontal=align,
vertical='center')
col += 1
# 2 其他border等格式控制
if r'\midrule' in line or r'\toprule' in line or r'\bottomrule' in line or r'\hline' in line:
# 该行画整条线
loc_1 = excel_addr(row, 1)
loc_2 = excel_addr(row, total_col)
comb_loc = f'{loc_1}:{loc_2}'
for cell in ws[comb_loc][0]:
cell.border = Border(top=double)
if r'\cmidrule' in line:
for match in re.findall(r'\\cmidrule{([0-9]+)-([0-9]+)}',
line):
loc_1 = excel_addr(row, match[0])
loc_2 = excel_addr(row, match[1])
comb_loc = f'{loc_1}:{loc_2}'
for cell in ws[comb_loc][0]:
cell.border = Border(top=double)
if r'\cline' in line:
for match in re.findall(r'\\cline{([0-9]+)-([0-9]+)}', line):
loc_1 = excel_addr(row, match[0])
loc_2 = excel_addr(row, match[1])
comb_loc = f'{loc_1}:{loc_2}'
for cell in ws[comb_loc][0]:
cell.border = Border(top=double)
row, col = row + 1, 1