@@ -111,3 +111,76 @@ def __clean_invisible_elements(self, data_json: DataJson) -> str:
111111 for element in elements :
112112 remove_element (element )
113113 return element_to_html (tree )
114+
115+
116+ class TestHTMLFileToDataJsonPreExtractor (HTMLFileFormatFilterPreExtractor ):
117+ """为了方便noclip管线对测试数据进行测试，根据路径读取html文件和main_html文件，然后转换为DataJson格式。"""
118+
119+ def __init__ (self , config : dict , html_parent_dir : str ):
120+ """
121+ 初始化函数
122+ Args:
123+ config:
124+ html_parent_dir:
125+ """
126+ super ().__init__ (config )
127+ self .__html_parent_path = html_parent_dir
128+
129+ @override
130+ def _do_pre_extract (self , data_json : DataJson ) -> DataJson :
131+ """对输入的html和main_html拼装到DataJson中，形成标准输入格式."""
132+ proj_root_dir = get_proj_root_dir ()
133+ html_file_path = os .path .join (proj_root_dir , self .__html_parent_path , data_json .get ('path' ))
134+ main_html_file_path = os .path .join (proj_root_dir , self .__html_parent_path , data_json .get ('main_path' ))
135+
136+ with open (html_file_path , 'r' , encoding = 'utf-8' ) as f :
137+ html = f .read ()
138+ data_json ['html' ] = html
139+ del data_json ['path' ]
140+
141+ with open (main_html_file_path , 'r' , encoding = 'utf-8' ) as f :
142+ main_html = f .read ()
143+ data_json ['main_html' ] = main_html
144+ del data_json ['main_path' ]
145+ return data_json
146+
147+
148+ class HTMLFileFormatNoClipPreExtractor (HTMLFileFormatFilterPreExtractor ):
149+ """noclip管线对main_html预处理."""
150+ def __init__ (self , config : dict ):
151+ super ().__init__ (config )
152+
153+ @override
154+ def _do_pre_extract (self , data_json : DataJson ) -> DataJson :
155+ data_json ['main_html' ] = self .__clean_interactive_elements (data_json )
156+ return data_json
157+
158+ def __clean_interactive_elements (self , data_json : DataJson ) -> str :
159+ """清除main_html中交互式元素."""
160+ html_content = data_json ['main_html' ]
161+ tree = html_to_element (html_content )
162+ interactive_tags = ['input' , 'select' , 'textarea' , 'button' ]
163+ # 删除<body>内的交互标签及关联label
164+ for tag in interactive_tags :
165+ for element in tree .xpath (f'//body//{ tag } ):
166+ # 删除标签本身
167+ parent = element .getparent ()
168+ if parent is not None :
169+ parent .remove (element )
170+
171+ # 删除关联的label（通过for属性匹配）
172+ if 'id' in element .attrib :
173+ for label in tree .xpath (f'//body//label[@for="{ element .attrib ["id" ]} ):
174+ label .getparent ().remove (label )
175+
176+ # 处理<form>内的交互标签及关联label
177+ for form in tree .xpath ('//form' ):
178+ # 删除表单内所有交互标签
179+ form_elements = form .xpath ('.//input | .//select | .//textarea | .//button | .//label | .//img' )
180+ for element in form_elements :
181+ element .getparent ().remove (element )
182+
183+ # 检查表单是否为空（无子元素或仅剩空白文本）
184+ if len (form .getchildren ()) == 0 or not form .text_content ().strip ():
185+ form .getparent ().remove (form )
186+ return element_to_html (tree )
0 commit comments