powertom 0.2.2

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -0,0 +1,229 @@
1
+ # frozen_string_literal: true
2
+
3
+ require "rexml/document"
4
+ require "zip"
5
+
6
+ module PowerTom
7
+ module Adapter
8
+ # 页面手术:按计划克隆/删除整页(§5.2)。
9
+ #
10
+ # ruby_pptx 0.2.0 尚无 duplicate / delete(Slides 只有 add / add_table_pages,
11
+ # 设计文档 §5.2 已核对),本类在适配层内基于 OOXML 包结构直接完成,
12
+ # XML 操作不外泄出适配层;待上游补齐能力后应收缩为对 ruby_pptx 的调用。
13
+ #
14
+ # 策略:输出页按顺序整体重新编号(slide1..slideM / notesSlide1..M),
15
+ # 彻底规避上游 next_slide_partname 以 sldIdLst.size+1 命名的撞名问题。
16
+ # 媒体 / 布局 / 母版 / tags 等部件被多个页面共享引用,无需复制。
17
+ class PageSurgery
18
+ SLIDE_CT = "application/vnd.openxmlformats-officedocument.presentationml.slide+xml"
19
+ NOTES_CT = "application/vnd.openxmlformats-officedocument.presentationml.notesSlide+xml"
20
+
21
+ # plan: [{source: 原页号, iteration: nil|k}] 按输出顺序(克隆页连续出现)。
22
+ # 返回 [{source:, iteration:}](与输出页一一对应,供渲染阶段做作用域绑定)。
23
+ def self.execute(template_path, out_path, plan)
24
+ new(template_path, out_path, plan).run
25
+ end
26
+
27
+ def initialize(template_path, out_path, plan)
28
+ @template_path = template_path
29
+ @out_path = out_path
30
+ @plan = plan
31
+ end
32
+
33
+ def run
34
+ parts = read_parts
35
+ order = slide_partnames_in_order(parts)
36
+
37
+ # 输出映射:每个输出页对应哪张原始页(含其 notes 部件)
38
+ outputs = @plan.map do |p|
39
+ src = order[p[:source] - 1]
40
+ raise Error, "页码越界:#{p[:source]}(共 #{order.size} 页)" unless src
41
+ { src_slide: src, src_notes: notes_part_for(parts, p[:source]) }
42
+ end
43
+
44
+ rewritten = build_rewritten_parts(parts, outputs)
45
+ write_zip(parts, rewritten)
46
+
47
+ @plan.map { |p| { source: p[:source], iteration: p[:iteration] } }
48
+ end
49
+
50
+ private
51
+
52
+ def slide_name(index) = "slide#{index + 1}.xml"
53
+
54
+ def notes_name(index) = "notesSlide#{index + 1}.xml"
55
+
56
+ # ---- 读取 ----
57
+
58
+ def read_parts
59
+ parts = {}
60
+ Zip::File.open(@template_path) do |zip|
61
+ zip.entries.each do |entry|
62
+ next if entry.directory?
63
+ parts[entry.name] = entry.get_input_stream.read
64
+ end
65
+ end
66
+ parts
67
+ end
68
+
69
+ def slide_partnames_in_order(parts)
70
+ pres = REXML::Document.new(parts.fetch("ppt/presentation.xml"))
71
+ rels = REXML::Document.new(parts.fetch("ppt/_rels/presentation.xml.rels"))
72
+ rid_target = {}
73
+ REXML::XPath.match(rels, "//Relationship").each do |rel|
74
+ target = rel.attributes["Target"].to_s
75
+ rid_target[rel.attributes["Id"]] = target if target.match?(%r{\Aslides/slide\d+\.xml\z})
76
+ end
77
+ REXML::XPath.match(pres, "//p:sldIdLst/p:sldId").map do |sld_id|
78
+ rid_attr = sld_id.attributes.get_attribute("r:id")
79
+ target = rid_target[rid_attr&.value]
80
+ raise Error, "presentation.xml 的 sldId 引用了未知关系" unless target
81
+ "ppt/#{target}"
82
+ end
83
+ end
84
+
85
+ def notes_part_for(parts, old_slide_no)
86
+ rels_name = "ppt/slides/_rels/slide#{old_slide_no}.xml.rels"
87
+ return nil unless parts[rels_name]
88
+ rel = REXML::XPath.match(REXML::Document.new(parts[rels_name]), "//Relationship").find do |r|
89
+ r.attributes["Type"].to_s.end_with?("/notesSlide")
90
+ end
91
+ return nil unless rel
92
+ "ppt/notesSlides/#{File.basename(rel.attributes['Target'])}"
93
+ end
94
+
95
+ # ---- 重写 ----
96
+
97
+ def build_rewritten_parts(parts, outputs)
98
+ count = outputs.size
99
+ new_rids = allocate_slide_rids(parts, count)
100
+
101
+ slide_parts = outputs.each_with_index.flat_map do |out, i|
102
+ {
103
+ "ppt/slides/#{slide_name(i)}" => parts.fetch(out[:src_slide]),
104
+ "ppt/slides/_rels/#{slide_name(i)}.rels" =>
105
+ rewrite_slide_rels(parts, out[:src_slide], out[:src_notes] ? notes_name(i) : nil)
106
+ }
107
+ end
108
+
109
+ notes_parts = outputs.each_with_index.filter_map do |out, i|
110
+ next nil unless out[:src_notes]
111
+ {
112
+ "ppt/notesSlides/#{notes_name(i)}" => parts.fetch(out[:src_notes]),
113
+ "ppt/notesSlides/_rels/#{notes_name(i)}.rels" =>
114
+ rewrite_notes_rels(parts, out[:src_notes], slide_name(i))
115
+ }
116
+ end
117
+
118
+ {
119
+ "ppt/presentation.xml" => rewrite_presentation(parts, count, new_rids),
120
+ "ppt/_rels/presentation.xml.rels" => rewrite_presentation_rels(parts, count, new_rids),
121
+ "[Content_Types].xml" => rewrite_content_types(parts, count),
122
+ "slide_parts" => slide_parts,
123
+ "notes_parts" => notes_parts
124
+ }
125
+ end
126
+
127
+ # presentation.xml.rels 里非 slide 关系保留原 rId,新 slide 关系从 max+1 起分配
128
+ def allocate_slide_rids(parts, count)
129
+ rels = REXML::Document.new(parts.fetch("ppt/_rels/presentation.xml.rels"))
130
+ max = REXML::XPath.match(rels, "//Relationship")
131
+ .filter_map { |r| r.attributes["Id"].to_s[/\ArId(\d+)\z/, 1] }
132
+ .map(&:to_i).max || 0
133
+ (1..count).map { |i| "rId#{max + i}" }
134
+ end
135
+
136
+ def rewrite_presentation(parts, count, new_rids)
137
+ xml = parts.fetch("ppt/presentation.xml").dup.force_encoding(Encoding::UTF_8)
138
+ entries = (0...count).map { |i| %(<p:sldId id="#{256 + i}" r:id="#{new_rids[i]}"/>) }.join
139
+ unless xml.sub!(%r{<p:sldIdLst>.*?</p:sldIdLst>}m, "<p:sldIdLst>#{entries}</p:sldIdLst>")
140
+ raise Error, "presentation.xml 中未找到 sldIdLst"
141
+ end
142
+ xml.force_encoding(Encoding::BINARY)
143
+ end
144
+
145
+ def rewrite_presentation_rels(parts, count, new_rids)
146
+ xml = parts.fetch("ppt/_rels/presentation.xml.rels").dup.force_encoding(Encoding::UTF_8)
147
+ xml = xml.gsub(%r{<Relationship\b[^>]*Target="slides/slide\d+\.xml"[^>]*/>}, "")
148
+ new_rels = (0...count).map do |i|
149
+ %(<Relationship Id="#{new_rids[i]}" ) +
150
+ %(Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/slide" ) +
151
+ %(Target="slides/#{slide_name(i)}"/>)
152
+ end
153
+ xml.sub!(%r{</Relationships>}, "#{new_rels.join}</Relationships>")
154
+ xml.force_encoding(Encoding::BINARY)
155
+ end
156
+
157
+ def rewrite_content_types(parts, count)
158
+ xml = parts.fetch("[Content_Types].xml").dup.force_encoding(Encoding::UTF_8)
159
+ xml = xml.gsub(%r{<Override [^>]*PartName="/ppt/slides/slide\d+\.xml"[^>]*/>}, "")
160
+ xml = xml.gsub(%r{<Override [^>]*PartName="/ppt/notesSlides/notesSlide\d+\.xml"[^>]*/>}, "")
161
+ overrides = (0...count).flat_map do |i|
162
+ [%(<Override PartName="/ppt/slides/slide#{i + 1}.xml" ContentType="#{SLIDE_CT}"/>),
163
+ %(<Override PartName="/ppt/notesSlides/notesSlide#{i + 1}.xml" ContentType="#{NOTES_CT}"/>)]
164
+ end
165
+ xml.sub!(%r{</Types>}, "#{overrides.join}</Types>")
166
+ xml.force_encoding(Encoding::BINARY)
167
+ end
168
+
169
+ # slide rels:notesSlide 指向重编号后的新 notesSlide;其余(媒体/布局/tags)原样共享
170
+ def rewrite_slide_rels(parts, src_slide_part, new_notes_name)
171
+ rels_name = "ppt/slides/_rels/#{File.basename(src_slide_part)}.rels"
172
+ xml = parts.fetch(rels_name).dup.force_encoding(Encoding::UTF_8)
173
+ if new_notes_name
174
+ xml = xml.gsub(%r{Target="\.\./notesSlides/notesSlide\d+\.xml"},
175
+ %(Target="../notesSlides/#{new_notes_name}"))
176
+ end
177
+ xml.force_encoding(Encoding::BINARY)
178
+ end
179
+
180
+ def rewrite_notes_rels(parts, src_notes_part, new_slide_name)
181
+ rels_name = "ppt/notesSlides/_rels/#{File.basename(src_notes_part)}.rels"
182
+ xml = parts.fetch(rels_name).dup.force_encoding(Encoding::UTF_8)
183
+ xml = xml.gsub(%r{Target="\.\./slides/slide\d+\.xml"}, %(Target="../slides/#{new_slide_name}"))
184
+ xml.force_encoding(Encoding::BINARY)
185
+ end
186
+
187
+ # ---- 写出 ----
188
+
189
+ # 这些部件不照抄:重编号后的 slides/notesSlides、重写后的 presentation 层
190
+ SKIP_PATTERNS = [
191
+ %r{\Appt/slides/slide\d+\.xml\z},
192
+ %r{\Appt/slides/_rels/slide\d+\.xml\.rels\z},
193
+ %r{\Appt/notesSlides/notesSlide\d+\.xml\z},
194
+ %r{\Appt/notesSlides/_rels/notesSlide\d+\.xml\.rels\z},
195
+ %r{\Appt/presentation\.xml\z},
196
+ %r{\Appt/_rels/presentation\.xml\.rels\z}
197
+ ].freeze
198
+
199
+ def write_zip(parts, rewritten)
200
+ Zip::OutputStream.open(@out_path) do |zos|
201
+ zos.put_next_entry("[Content_Types].xml")
202
+ zos.write(rewritten.fetch("[Content_Types].xml"))
203
+
204
+ parts.each do |name, data|
205
+ next if name == "[Content_Types].xml"
206
+ next if SKIP_PATTERNS.any? { |re| name.match?(re) }
207
+ zos.put_next_entry(name)
208
+ zos.write(data)
209
+ end
210
+
211
+ { "ppt/presentation.xml" => "ppt/presentation.xml",
212
+ "ppt/_rels/presentation.xml.rels" => "ppt/_rels/presentation.xml.rels" }.each do |key, name|
213
+ zos.put_next_entry(name)
214
+ zos.write(rewritten.fetch(key))
215
+ end
216
+
217
+ %w[slide_parts notes_parts].each do |group|
218
+ rewritten.fetch(group).each do |entries|
219
+ entries.each do |name, data|
220
+ zos.put_next_entry(name)
221
+ zos.write(data)
222
+ end
223
+ end
224
+ end
225
+ end
226
+ end
227
+ end
228
+ end
229
+ end
@@ -0,0 +1,403 @@
1
+ # frozen_string_literal: true
2
+
3
+ require "ruby_pptx"
4
+
5
+ module PowerTom
6
+ module Adapter
7
+ # 设计文档 §4:适配层是唯一 import ruby_pptx 的模块。
8
+ # 上游 0.x API 变动时,改动被封在这个文件里。
9
+ class RubyPptxAdapter
10
+ def open(path)
11
+ Pptx::Presentation.open(path)
12
+ rescue StandardError => e
13
+ raise InvalidTemplateError, "模板无法打开(#{e.class}: #{e.message})"
14
+ end
15
+
16
+ # 遍历全部幻灯片的文本段落(组合形状递归展开)。
17
+ # 返回 [{slide_no:, shape:, paragraph:, run_texts:}]
18
+ def paragraph_contexts(prs)
19
+ contexts = []
20
+ prs.slides.each_with_index do |slide, i|
21
+ slide_no = i + 1
22
+ each_text_shape(slide.shapes) do |shape|
23
+ shape.text_frame.paragraphs.each do |paragraph|
24
+ runs = paragraph.runs
25
+ next if runs.empty?
26
+ contexts << {
27
+ slide_no: slide_no,
28
+ shape: shape.name,
29
+ paragraph: paragraph,
30
+ run_texts: runs.map(&:text)
31
+ }
32
+ end
33
+ end
34
+ end
35
+ contexts
36
+ end
37
+
38
+ # §5.1 写回:按下标设置每个 run 的文本。文本数组长度必须与现有 run 数一致,
39
+ # run 节点本身不动(保留 rPr 样式引用)。
40
+ def write_paragraph_runs(paragraph, texts)
41
+ runs = paragraph.runs
42
+ if texts.size != runs.size
43
+ raise Error, "写回失败:run 数量不匹配(现有 #{runs.size},传入 #{texts.size})"
44
+ end
45
+ runs.each_with_index { |run, i| run.text = texts[i] }
46
+ end
47
+
48
+ # 备注文本(§3.3:v0 不渲染备注,只用于发现其中的标签并记 warning)。
49
+ # 返回 [{slide_no:, text:}];无备注页返回空串。
50
+ def notes_texts(prs)
51
+ prs.slides.each_with_index.map do |slide, i|
52
+ { slide_no: i + 1, text: slide.notes.to_s }
53
+ end
54
+ rescue StandardError
55
+ # 备注读取异常不影响正文渲染(v0 备注只做检查)
56
+ []
57
+ end
58
+
59
+ # 图片占位上下文:形状 name 或 alt text 符合 powertom:image:字段名 约定(§5.3)
60
+ # 返回 [{slide_no:, shape:, field:, slide:, picture:}]
61
+ def picture_contexts(prs)
62
+ contexts = []
63
+ prs.slides.each_with_index do |slide, i|
64
+ walk_shapes(slide.shapes) do |shape|
65
+ next unless shape.is_a?(Pptx::Picture)
66
+ field = image_field(shape)
67
+ next unless field
68
+ contexts << { slide_no: i + 1, shape: shape.name, field: field,
69
+ slide: slide, picture: shape }
70
+ end
71
+ end
72
+ contexts
73
+ end
74
+
75
+ # §5.3 图片替换:get_or_add 新 image part(按内容 sha1 去重,扩展名/content type
76
+ # 由 ruby_pptx 按真实图片类型维护),repoint blip 引用,再按"填满形状、居中裁切"
77
+ # 设置 srcRect。SVG 等矢量图不做裁切。
78
+ def replace_image(slide, picture, image_path)
79
+ _image_part, r_id = slide.part.get_or_add_image_part(image_path)
80
+ pic_el = picture.element
81
+ pic_el.blip.embed = r_id
82
+ image = Pptx::Image.from_file(image_path)
83
+ unless image.vector?
84
+ view = [picture.width, picture.height]
85
+ if view[0] && view[1]
86
+ # crop_to_fit 只关心比例;Length 没有 to_f,传 EMU 裸值
87
+ native = image.native_size
88
+ pic_el.crop_to_fit([native[0].emu, native[1].emu],
89
+ [view[0].emu, view[1].emu])
90
+ end
91
+ end
92
+ true
93
+ rescue StandardError => e
94
+ raise PowerTom::Error,
95
+ "图片替换失败(#{File.basename(image_path.to_s)}):#{e.class} #{e.message}"
96
+ end
97
+
98
+ # 逐页形状名(含组合内、含无文本形状),用于模板结构自检。
99
+ # 返回 [{slide_no:, names: [String]}]
100
+ def shape_names_by_slide(prs)
101
+ prs.slides.each_with_index.map do |slide, i|
102
+ names = []
103
+ walk_shapes(slide.shapes) { |shape| names << shape.name }
104
+ { slide_no: i + 1, names: names }
105
+ end
106
+ end
107
+
108
+ # 母版与版式上的文本形状(用于发现写在母版/版式上的标签)。
109
+ # 返回 [{container:, shape:, text:}]
110
+ def container_texts(prs)
111
+ entries = []
112
+ prs.slide_masters.each do |master|
113
+ collect_container_texts(entries, master, "母版")
114
+ master.slide_layouts.each do |layout|
115
+ collect_container_texts(entries, layout, "版式")
116
+ end
117
+ end
118
+ entries
119
+ end
120
+
121
+ # 表格结构(表格行循环用):逐页每张表的行与单元格文本,以及合并单元格分布
122
+ # 返回 [{slide_no:, shape:, columns:, element:, frame:, rows: [[cell_text, ...], ...],
123
+ # merged_rows: [bool, ...]}]
124
+ def table_contexts(prs)
125
+ list = []
126
+ prs.slides.each_with_index do |slide, i|
127
+ walk_shapes(slide.shapes) do |shape|
128
+ next unless shape.respond_to?(:table?) && shape.table?
129
+
130
+ table = shape.table
131
+ list << {
132
+ slide_no: i + 1,
133
+ shape: shape.name,
134
+ columns: table.columns.size,
135
+ element: table_element(shape),
136
+ frame: shape.element,
137
+ rows: table.rows.map { |row| row.cells.map { |cell| cell.text.to_s } },
138
+ merged_rows: table.rows.map do |row|
139
+ row.cells.any? { |cell| cell.merge_origin? || cell.spanned? }
140
+ end
141
+ }
142
+ end
143
+ end
144
+ list
145
+ end
146
+
147
+ # 克隆第 row_index 行为 count 份(插在原位置),再删除模板行;
148
+ # 同步 graphicFrame 高度,避免行数变化后被外框裁剪。
149
+ def expand_table_row(table_element, frame_element, row_index, count)
150
+ template = table_element.tr_list[row_index]
151
+ count.times { template.add_previous_sibling(table_element.import(template)) }
152
+ table_element.remove(template)
153
+ sync_table_height(table_element, frame_element)
154
+ true
155
+ end
156
+
157
+ # 两级嵌套展开:每个父项先输出一行父内容行,随后输出它的子内容行。
158
+ # child_counts: [每个父项的子数组长度]
159
+ # 模板中的父/子标记行与内容行在展开后全部删除。
160
+ def expand_table_rows_nested(table_element, frame_element, plan, child_counts)
161
+ trs = table_element.tr_list
162
+ parent_marker = trs[plan[:parent][:marker]]
163
+ parent_content = trs[plan[:parent][:content]]
164
+ child_marker = trs[plan[:child][:marker]]
165
+ child_content = trs[plan[:child][:content]]
166
+
167
+ child_counts.each do |count|
168
+ parent_marker.add_previous_sibling(table_element.import(parent_content))
169
+ count.times { parent_marker.add_previous_sibling(table_element.import(child_content)) }
170
+ end
171
+
172
+ [parent_marker, parent_content, child_marker, child_content].each do |tr|
173
+ table_element.remove(tr)
174
+ end
175
+ sync_table_height(table_element, frame_element)
176
+ true
177
+ end
178
+
179
+ # 动态列(#grid):首列固定,第二列是列模板;按 column_count 克隆
180
+ # a:gridCol 与每一行的第二个 a:tc,再删除模板列。
181
+ def expand_grid_columns(table_element, frame_element, column_count)
182
+ grid = table_element.xpath("./a:tblGrid").first
183
+ template_col = grid.gridCol_list[1]
184
+ column_count.times { template_col.add_previous_sibling(grid.import(template_col)) }
185
+ grid.remove(template_col)
186
+
187
+ table_element.tr_list.each do |tr|
188
+ template_cell = tr.tc_list[1]
189
+ column_count.times { template_cell.add_previous_sibling(tr.import(template_cell)) }
190
+ tr.remove(template_cell)
191
+ end
192
+ sync_table_height(table_element, frame_element)
193
+ true
194
+ end
195
+
196
+ # grid 标记只用于声明,渲染前从表头首格移除(保留同格的其他文字与首 run 样式)
197
+ def strip_grid_marker(table_element, row_index, column_index)
198
+ cell = table_element.tr_list[row_index].tc_list[column_index]
199
+ strip_tags_in_cell(cell, /\{\{\s*#grid\b[^{}]*\}\}/)
200
+ true
201
+ end
202
+
203
+ # 列带标记({{#cols}} / {{/cols}})在列克隆前移除,避免被复制到每一列
204
+ def strip_column_markers(table_element, row_index, first_col, last_col)
205
+ cells = table_element.tr_list[row_index].tc_list
206
+ (first_col..last_col).each do |c|
207
+ strip_tags_in_cell(cells[c], /\{\{\s*(?:#cols\b[^{}]*|\/cols)\s*\}\}/)
208
+ end
209
+ true
210
+ end
211
+
212
+ def table_row_element(table_element, row_index)
213
+ table_element.tr_list[row_index]
214
+ end
215
+
216
+ # 把 [open_element, close_element] 之间的模板行整体克隆 count 份。
217
+ # 返回每份克隆的行元素数组(按输出顺序),并删除开闭标记与模板行。
218
+ def clone_table_row_range(table_element, open_element, close_element, count)
219
+ trs = table_element.tr_list
220
+ open_index = trs.index { |tr| tr.node == open_element.node }
221
+ close_index = trs.index { |tr| tr.node == close_element.node }
222
+ unless open_index && close_index && close_index > open_index
223
+ raise PowerTom::Error, "行区块标记定位失败(区块可能已被展开)"
224
+ end
225
+
226
+ templates = (open_index + 1...close_index).map { |i| trs[i] }
227
+ groups = Array.new(count) do
228
+ templates.map { |template| open_element.add_previous_sibling(table_element.import(template)) }
229
+ end
230
+ ([open_element, close_element] + templates).each { |tr| table_element.remove(tr) }
231
+ groups
232
+ end
233
+
234
+ # 条件区块:标记行总是删除;keep_content=false 时内容行一并删除
235
+ def prune_table_row_range(table_element, open_element, close_element, keep_content:)
236
+ trs = table_element.tr_list
237
+ open_index = trs.index { |tr| tr.node == open_element.node }
238
+ close_index = trs.index { |tr| tr.node == close_element.node }
239
+ unless open_index && close_index && close_index > open_index
240
+ raise PowerTom::Error, "条件区块标记定位失败(区块可能已被展开)"
241
+ end
242
+
243
+ content = (open_index + 1...close_index).map { |i| trs[i] }
244
+ content.each { |tr| table_element.remove(tr) } unless keep_content
245
+ table_element.remove(open_element)
246
+ table_element.remove(close_element)
247
+ true
248
+ end
249
+
250
+ # 区块展开/裁剪后统一刷新表格高度
251
+ def refresh_table_height(table_element, frame_element)
252
+ sync_table_height(table_element, frame_element)
253
+ true
254
+ end
255
+
256
+ # 动态列带:模板列 [first_col..last_col] 按 count 克隆(表格网格 + 每一行)
257
+ def expand_column_band(table_element, frame_element, first_col, last_col, count)
258
+ grid = table_element.xpath("./a:tblGrid").first
259
+ grid_cols = grid.gridCol_list
260
+ template_cols = (first_col..last_col).map { |i| grid_cols[i] }
261
+ anchor = grid_cols[last_col]
262
+ count.times do
263
+ template_cols.each { |col| anchor.add_previous_sibling(grid.import(col)) }
264
+ end
265
+ template_cols.each { |col| grid.remove(col) }
266
+
267
+ table_element.tr_list.each do |tr|
268
+ cells = tr.tc_list
269
+ template_cells = (first_col..last_col).map { |i| cells[i] }
270
+ cell_anchor = cells[last_col]
271
+ count.times do
272
+ template_cells.each { |cell| cell_anchor.add_previous_sibling(tr.import(cell)) }
273
+ end
274
+ template_cells.each { |cell| tr.remove(cell) }
275
+ end
276
+ sync_table_height(table_element, frame_element)
277
+ true
278
+ end
279
+
280
+ # 把单元格段落里的指定标签移除(保留同格其他文字与首 run 样式)
281
+ def strip_tags_in_cell(cell, pattern)
282
+ cell.xpath("./a:txBody/a:p").each do |paragraph|
283
+ runs = paragraph.xpath("./a:r")
284
+ next if runs.empty?
285
+
286
+ joined = runs.map { |run| run.xpath("./a:t").first&.text.to_s }.join
287
+ next unless joined.match?(pattern)
288
+
289
+ cleaned = joined.gsub(pattern, "").strip
290
+ runs.each_with_index do |run, i|
291
+ text_node = run.xpath("./a:t").first
292
+ text_node.text = i.zero? ? cleaned : "" if text_node
293
+ end
294
+ end
295
+ true
296
+ end
297
+
298
+ # 表格单元格里的段落(展开行之后采集,用于绑定)
299
+ def table_cell_contexts(prs)
300
+ list = []
301
+ prs.slides.each_with_index do |slide, i|
302
+ walk_shapes(slide.shapes) do |shape|
303
+ next unless shape.respond_to?(:table?) && shape.table?
304
+
305
+ table = shape.table
306
+ table.rows.each_with_index do |row, r|
307
+ row.cells.each_with_index do |cell, c|
308
+ cell.text_frame.paragraphs.each do |paragraph|
309
+ runs = paragraph.runs
310
+ next if runs.empty?
311
+
312
+ list << {
313
+ slide_no: i + 1,
314
+ shape: "#{shape.name}[#{r + 1},#{c + 1}]",
315
+ table_shape: shape.name,
316
+ row_index: r,
317
+ col_index: c,
318
+ in_table: true,
319
+ paragraph: paragraph,
320
+ run_texts: runs.map(&:text)
321
+ }
322
+ end
323
+ end
324
+ end
325
+ end
326
+ end
327
+ list
328
+ end
329
+
330
+ # 按名称查找形状(含组合内);找不到返回 nil
331
+ def find_shape(prs, slide_no, name)
332
+ slide = prs.slides[slide_no - 1]
333
+ return nil unless slide
334
+ found = nil
335
+ walk_shapes(slide.shapes) { |shape| found ||= shape if shape.name == name }
336
+ found
337
+ end
338
+
339
+ def save(prs, path)
340
+ prs.save(path)
341
+ end
342
+
343
+ private
344
+
345
+ def walk_shapes(shapes, &block)
346
+ shapes.each do |shape|
347
+ if shape.is_a?(Pptx::GroupShape)
348
+ walk_shapes(shape.shapes, &block)
349
+ else
350
+ block.call(shape)
351
+ end
352
+ end
353
+ end
354
+
355
+ def each_text_shape(shapes, &block)
356
+ walk_shapes(shapes) { |shape| block.call(shape) if shape.text_frame? }
357
+ end
358
+
359
+ IMAGE_FIELD_PREFIX = "powertom:image:"
360
+
361
+ def image_field(shape)
362
+ name = shape.name.to_s
363
+ return name.delete_prefix(IMAGE_FIELD_PREFIX) if name.start_with?(IMAGE_FIELD_PREFIX)
364
+ descr = alt_text_of(shape)
365
+ return descr.delete_prefix(IMAGE_FIELD_PREFIX) if descr&.start_with?(IMAGE_FIELD_PREFIX)
366
+ nil
367
+ end
368
+
369
+ def alt_text_of(shape)
370
+ cnv_pr = shape.element.xpath("./p:nvPicPr/p:cNvPr").first
371
+ cnv_pr&.get("descr")
372
+ rescue StandardError
373
+ nil
374
+ end
375
+
376
+ def table_element(shape)
377
+ shape.element.xpath("./a:graphic/a:graphicData/a:tbl").first
378
+ end
379
+
380
+ # 表格实际高度 = 各行 h 之和;写回 graphicFrame 的 a:ext/@cy
381
+ def sync_table_height(table_element, frame_element)
382
+ total = table_element.tr_list.sum { |tr| tr.get("h").to_i }
383
+ ext = frame_element.xpath("./p:xfrm/a:ext").first
384
+ ext&.set("cy", total.to_s)
385
+ end
386
+
387
+ def collect_container_texts(entries, container, kind)
388
+ label = begin
389
+ name = container.part.name
390
+ name.nil? || name.to_s.empty? ? kind : "#{kind}『#{name}』"
391
+ rescue StandardError
392
+ kind
393
+ end
394
+ walk_shapes(container.shapes) do |shape|
395
+ next unless shape.text_frame?
396
+ text = shape.text_frame.text.to_s
397
+ next unless text.include?("{{")
398
+ entries << { container: label, shape: shape.name, text: text }
399
+ end
400
+ end
401
+ end
402
+ end
403
+ end