hwp_script_to_latex 1.3.5 → 1.4.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/lib/hwp_script_to_latex/converter.rb +33 -32
- data/lib/hwp_script_to_latex/version.rb +1 -1
- data/rules.json +12 -8
- metadata +1 -1
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 8a1f7260a320ad6b3be51ded4f0a61b2da7d0143ed340715e37a0ebdcf89f2e1
|
4
|
+
data.tar.gz: 045d8409baaeff28e2803f13f7c5fe152e8f7100749d40a253fef6f2f21f6467
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 020d0e2857c38dbf1aa4ae5051150631bcb4033f075fef06c2229923fce601ca74c9d3058ccaa5e106b3eb2c923528af7b3993802681997a4869d29087333de5
|
7
|
+
data.tar.gz: 5addae0db39cea048c66807b8f1141a31c8206f63dd3788a995bedb3bfb59331096463924eca23f3e2c1c8ba934fab1df9cb007fa5457c185c9e7f0369627b07
|
@@ -14,7 +14,7 @@ module HwpScriptToLatex
|
|
14
14
|
|
15
15
|
@rules= JSON.parse(file_content, symbolize_names: true)
|
16
16
|
@simple_commands = @rules[:command][:simple].sort_by { |hash| -1 * hash[:script].length }
|
17
|
-
@
|
17
|
+
@block_commands = @rules[:command][:block]
|
18
18
|
@meta_keywords = @rules[:keyword][:meta].sort_by { |hash| -1 * hash[:regex].inspect.length }
|
19
19
|
@command_keywords = @rules[:keyword][:command].sort_by { |hash| -1 * hash[:regex].inspect.length }
|
20
20
|
@symbol_keywords = @rules[:keyword][:symbol].sort_by { |hash| -1 * hash[:regex].inspect.length }
|
@@ -24,10 +24,17 @@ module HwpScriptToLatex
|
|
24
24
|
def convert(script, math_mode: false, display_mode: false)
|
25
25
|
# Sanitize for starting convert
|
26
26
|
result = pre_sanitize(script)
|
27
|
-
#
|
27
|
+
# 단순 치환 키워드
|
28
28
|
result = replace_keywords(result)
|
29
|
-
#
|
29
|
+
# 1개의 우항을 가지는 간단한 명령어
|
30
30
|
result = replace_simple_commands(result)
|
31
|
+
# 행렬, 케이스등 블록 명령어
|
32
|
+
# cases {...} => \begin{cases}...\end{cases}
|
33
|
+
# dmatrix {...} => \begin{vmatrix}...\end{vmatrix}
|
34
|
+
# bmatrix {...} => \begin{Bmatrix}...\end{Bmatrix}
|
35
|
+
# pmatrix {...} => \begin{pmatrix}...\end{pmatrix}
|
36
|
+
# matrix {...} => \begin{matrix}...\end{matrix}
|
37
|
+
result = replace_block_commands(result)
|
31
38
|
|
32
39
|
# 특수 형태의 명령어 치환
|
33
40
|
# Case 1. 루트 변환
|
@@ -35,14 +42,8 @@ module HwpScriptToLatex
|
|
35
42
|
# sqrt A of B => \sqrt [A]{B}
|
36
43
|
# Case 2. 분수 변환
|
37
44
|
# A over B => \dfrac {A}{B}
|
38
|
-
# Case 3. 행렬 변환
|
39
|
-
# dmatrix {...} => \begin{vmatrix}...\end{vmatrix}
|
40
|
-
# bmatrix {...} => \begin{Bmatrix}...\end{Bmatrix}
|
41
|
-
# pmatrix {...} => \begin{pmatrix}...\end{pmatrix}
|
42
|
-
# matrix {...} => \begin{matrix}...\end{matrix}
|
43
45
|
result = replace_sqrt(result) # Case 1
|
44
46
|
result = replace_fractions(result) # Case 2
|
45
|
-
result = replace_matrix(result) # Case 3
|
46
47
|
|
47
48
|
# 전체 수식에 디스플레이 스타일 적용
|
48
49
|
result = decorate_displaystyle(result) if display_mode
|
@@ -65,8 +66,8 @@ module HwpScriptToLatex
|
|
65
66
|
script = script.gsub(jokbo_regex, "")
|
66
67
|
# 2개 이상의 공백을 하나의 공백으로 치환
|
67
68
|
script = script.gsub(/\s+/, " ").strip
|
68
|
-
#
|
69
|
-
script = script.gsub(/\\/, "
|
69
|
+
# 백슬래시(로만체)를 삭제
|
70
|
+
script = script.gsub(/\\/, "")
|
70
71
|
# 꺽쇠 치환
|
71
72
|
script = script.gsub(/</, "<")
|
72
73
|
script = script.gsub(/>/, ">")
|
@@ -103,6 +104,24 @@ module HwpScriptToLatex
|
|
103
104
|
return script
|
104
105
|
end
|
105
106
|
|
107
|
+
def replace_block_commands(script)
|
108
|
+
@block_commands.each do |command|
|
109
|
+
command_regex = %r((?<![a-zA-Z])(?i:#{command[:script]})\s*(?<block_content>{(?>[^{}]+|(?:\g<block_content>))*}))
|
110
|
+
|
111
|
+
match_data = script.match(command_regex)
|
112
|
+
while match_data
|
113
|
+
# 시작, 끝 중괄호 제거
|
114
|
+
block_content = remove_curly_brackets(match_data['block_content'])
|
115
|
+
# sub 메서드에 블록 문법을 사용한 이유:
|
116
|
+
# gsub, sub에서 백슬래시 문제
|
117
|
+
script = script.sub(command_regex) { " \\begin{#{command[:latex]}} %s \\end{#{command[:latex]}} " % block_content }
|
118
|
+
match_data = script.match(command_regex)
|
119
|
+
end
|
120
|
+
end
|
121
|
+
|
122
|
+
return script
|
123
|
+
end
|
124
|
+
|
106
125
|
def replace_keywords(script)
|
107
126
|
keywords = @meta_keywords + @command_keywords + @symbol_keywords + @reserved_keywords
|
108
127
|
matched_count = 0
|
@@ -150,24 +169,6 @@ module HwpScriptToLatex
|
|
150
169
|
return script
|
151
170
|
end
|
152
171
|
|
153
|
-
def replace_matrix(script)
|
154
|
-
@matrix_commands.each do |command|
|
155
|
-
matrix_regex = %r((?<![a-zA-Z])(?i:#{command[:script]})\s*(?<matrix_content>{(?>[^{}]+|(?:\\g<matrix_content>))*}))
|
156
|
-
|
157
|
-
match_data = script.match(matrix_regex)
|
158
|
-
while match_data
|
159
|
-
# 시작, 끝 중괄호 제거
|
160
|
-
matrix_content = remove_curly_brackets(match_data['matrix_content'])
|
161
|
-
# sub 메서드에 블록 문법을 사용한 이유:
|
162
|
-
# gsub, sub에서 백슬래시 문제
|
163
|
-
script = script.sub(matrix_regex) { " \\begin{#{command[:latex]}} %s \\end{#{command[:latex]}} " % matrix_content }
|
164
|
-
match_data = script.match(matrix_regex)
|
165
|
-
end
|
166
|
-
end
|
167
|
-
|
168
|
-
return script
|
169
|
-
end
|
170
|
-
|
171
172
|
def replace_sqrt(script)
|
172
173
|
right_term_name1 = 'rt1'
|
173
174
|
right_term_name2 = 'rt2'
|
@@ -217,9 +218,9 @@ module HwpScriptToLatex
|
|
217
218
|
end
|
218
219
|
|
219
220
|
def decorate_displaystyle(script)
|
220
|
-
script =
|
221
|
-
script = script.gsub(
|
222
|
-
script = script.gsub(
|
221
|
+
script = script.gsub(/\\sum/, "\\displaystyle \\sum")
|
222
|
+
script = script.gsub(/\\int/, "\\displaystyle \\int")
|
223
|
+
script = script.gsub(/\\oint/, "\\displaystyle \\oint")
|
223
224
|
|
224
225
|
return script
|
225
226
|
end
|
data/rules.json
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
{
|
2
2
|
"_comments": [
|
3
3
|
"command.simple: 1개의 우항을 가지는 간단한 명령어",
|
4
|
-
"command.
|
4
|
+
"command.block: begin/end 블록을 가지는 명령어(행렬, 케이스)",
|
5
5
|
"keyword.meta: 공백, 줄바꿈, left right 등 특수문자",
|
6
6
|
"keyword.command: 항이 없는 단순 치환 명령어",
|
7
7
|
"keyword.symbol: 기호",
|
@@ -82,7 +82,11 @@
|
|
82
82
|
"latex": "\\underline"
|
83
83
|
}
|
84
84
|
],
|
85
|
-
"
|
85
|
+
"block": [
|
86
|
+
{
|
87
|
+
"script": "cases",
|
88
|
+
"latex": "cases"
|
89
|
+
},
|
86
90
|
{
|
87
91
|
"script": "dmatrix",
|
88
92
|
"latex": "vmatrix"
|
@@ -1098,16 +1102,16 @@
|
|
1098
1102
|
],
|
1099
1103
|
"command": [
|
1100
1104
|
{
|
1101
|
-
"regex": ["
|
1102
|
-
"latex": "
|
1105
|
+
"regex": ["eqalign", "Eqalign", "EQALIGN"],
|
1106
|
+
"latex": ""
|
1103
1107
|
},
|
1104
1108
|
{
|
1105
|
-
"regex": ["
|
1106
|
-
"latex": "
|
1109
|
+
"regex": ["rm", "Rm", "RM"],
|
1110
|
+
"latex": ""
|
1107
1111
|
},
|
1108
1112
|
{
|
1109
|
-
"regex": ["
|
1110
|
-
"latex": "
|
1113
|
+
"regex": ["it", "It", "IT"],
|
1114
|
+
"latex": ""
|
1111
1115
|
},
|
1112
1116
|
{
|
1113
1117
|
"regex": ["sum", "Sum", "SUM"],
|