1 | #Region "Microsoft.VisualBasic::992ffb00fb8ac813b61f7e02a3637e6e, Microsoft.VisualBasic.Core\Text\Xml\Linq\Linq.vb" |
2 | |
3 | ' Author: |
4 | ' |
5 | ' asuka (amethyst.asuka@gcmodeller.org) |
6 | ' xie (genetics@smrucc.org) |
7 | ' xieguigang (xie.guigang@live.com) |
8 | ' |
9 | ' Copyright (c) 2018 GPL3 Licensed |
10 | ' |
11 | ' |
12 | ' GNU GENERAL PUBLIC LICENSE (GPL3) |
13 | ' |
14 | ' |
15 | ' This program is free software: you can redistribute it and/or modify |
16 | ' it under the terms of the GNU General Public License as published by |
17 | ' the Free Software Foundation, either version 3 of the License, or |
18 | ' (at your option) any later version. |
19 | ' |
20 | ' This program is distributed in the hope that it will be useful, |
21 | ' but WITHOUT ANY WARRANTY; without even the implied warranty of |
22 | ' MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
23 | ' GNU General Public License for more details. |
24 | ' |
25 | ' You should have received a copy of the GNU General Public License |
26 | ' along with this program. If not, see <http://www.gnu.org/licenses/>. |
27 | |
28 | |
29 | |
30 | ' /********************************************************************************/ |
31 | |
32 | ' Summaries: |
33 | |
34 | ' Module Data |
35 | ' |
36 | ' Function: GetNodeNameDefine, GetTypeName, InternalIterates, LoadUltraLargeXMLDataSet, LoadXmlDataSet |
37 | ' LoadXmlDocument, NodeInstanceBuilder, PopulateXmlElementText, UltraLargeXmlNodesIterator |
38 | ' |
39 | ' |
40 | ' /********************************************************************************/ |
41 | |
42 | #End Region |
43 | |
44 | Imports System.Runtime.CompilerServices |
45 | Imports System.Text |
46 | Imports System.Xml |
47 | Imports System.Xml.Serialization |
48 | Imports Microsoft.VisualBasic.Language |
49 | |
50 | Namespace Text.Xml.Linq |
51 | |
52 | ''' <summary> |
53 | ''' Using large xml file as Linq data source |
54 | ''' </summary> |
55 | Public Module Data |
56 | |
57 | ''' <summary> |
58 | ''' Load a specific xml file from a file location <paramref name="pathOrDoc"/>/ or |
59 | ''' a xml text document data into a <see cref="XmlDocument"/> object. |
60 | ''' </summary> |
61 | ''' <param name="pathOrDoc"></param> |
62 | ''' <returns></returns> |
63 | ''' <remarks> |
64 | ''' using internally <see cref="XDocument.Load"/> to parse whole XML at once |
65 | ''' </remarks> |
66 | <Extension> |
67 | Public Function LoadXmlDocument(pathOrDoc$, Optional preprocess As Func(Of String, String) = Nothing) As XmlDocument |
68 | Dim xmlDoc As New XmlDocument() |
69 | Dim doc$ |
70 | |
71 | If pathOrDoc.FileExists Then |
72 | If Not preprocess Is Nothing Then |
73 | doc = preprocess(pathOrDoc.ReadAllText) |
74 | xmlDoc.LoadXml(doc) |
75 | Else |
76 | Call xmlDoc.Load(pathOrDoc) |
77 | End If |
78 | Else |
79 | If Not preprocess Is Nothing Then |
80 | doc = preprocess(pathOrDoc) |
81 | Else |
82 | doc = pathOrDoc |
83 | End If |
84 | |
85 | Call xmlDoc.LoadXml(doc) |
86 | End If |
87 | |
88 | Return xmlDoc |
89 | End Function |
90 | |
91 | ''' <summary> |
92 | ''' Using <paramref name="default"/> string name or <see cref="Type.Name"/> |
93 | ''' </summary> |
94 | ''' <param name="type"></param> |
95 | ''' <param name="default$"> |
96 | ''' If this parameter value is <see cref="StringEmpty"/>, then <see cref="Type.Name"/> will be use as the xml node name. |
97 | ''' </param> |
98 | ''' <returns></returns> |
99 | <Extension> |
100 | Public Function GetTypeName(type As Type, default$) As String |
101 | If [default].StringEmpty Then |
102 | Return type.Name |
103 | Else |
104 | Return [default] |
105 | End If |
106 | End Function |
107 | |
108 | ''' <summary> |
109 | ''' 分别解析<see cref="XmlTypeAttribute"/>,<see cref="XmlRootAttribute"/>,如果这两个定义都不存在的话就返回<see cref="Type.Name"/> |
110 | ''' </summary> |
111 | ''' <param name="type"></param> |
112 | ''' <returns></returns> |
113 | <Extension> |
114 | Public Function GetNodeNameDefine(type As Type) As String |
115 | With type.GetCustomAttributes(GetType(XmlTypeAttribute), inherit:=True) |
116 | If Not .IsNullOrEmpty Then |
117 | With DirectCast(.First, XmlTypeAttribute).TypeName |
118 | If Not .StringEmpty Then |
119 | Return .ByRef |
120 | End If |
121 | End With |
122 | End If |
123 | End With |
124 | |
125 | With type.GetCustomAttributes(GetType(XmlRootAttribute), inherit:=True) |
126 | If Not .IsNullOrEmpty Then |
127 | With DirectCast(.First, XmlRootAttribute).ElementName |
128 | If Not .StringEmpty Then |
129 | Return .ByRef |
130 | End If |
131 | End With |
132 | End If |
133 | End With |
134 | |
135 | Return type.Name |
136 | End Function |
137 | |
138 | Private Iterator Function InternalIterates(XML$, nodeName$) As IEnumerable(Of String) |
139 | Dim XmlNodeList As XmlNodeList = XML _ |
140 | .LoadXmlDocument _ |
141 | .GetElementsByTagName(nodeName) |
142 | Dim sb As New StringBuilder |
143 | |
144 | For Each xmlNode As XmlNode In XmlNodeList |
145 | Call sb.Clear() |
146 | Call sb.Append($"<{nodeName} xmlns:xsd=""http://www.w3.org/2001/XMLSchema"" xmlns:xsi=""http://www.w3.org/2001/XMLSchema-instance""") |
147 | Call sb.Append(" ") |
148 | |
149 | For Each attr As XmlAttribute In xmlNode.Attributes |
150 | Call sb.Append($"{attr.Name}=""{attr.Value}""") |
151 | Call sb.Append(" ") |
152 | Next |
153 | |
154 | Call sb.AppendLine(">") |
155 | Call sb.AppendLine(xmlNode.InnerXml) |
156 | Call sb.AppendLine($"</{nodeName}>") |
157 | |
158 | Yield sb.ToString |
159 | Next |
160 | End Function |
161 | |
162 | ''' <summary> |
163 | ''' Only works for the xml file that contains a list or array of xml element, and then this function using this list element as linq data source. |
164 | ''' (这个函数只建议在读取比较小的XML文件的时候使用,并且这个XML文件仅仅是一个数组或者列表的序列化结果) |
165 | ''' </summary> |
166 | ''' <typeparam name="T"></typeparam> |
167 | ''' <param name="XML$">超大的XML文件的文件路径</param> |
168 | ''' <param name="typeName"> |
169 | ''' 列表之中的节点在XML之中的tag标记名称,假若这个参数值为空的话,则会默认使用目标类型名称<see cref="Type.Name"/> |
170 | ''' </param> |
171 | ''' <param name="xmlns"> |
172 | ''' Using for the namespace replacement. |
173 | ''' (当这个参数存在的时候,目标命名空间申明将会被替换为空字符串,数据对象才会被正确的加载) |
174 | ''' </param> |
175 | ''' <returns></returns> |
176 | <Extension> |
177 | Public Function LoadXmlDataSet(Of T As Class)(XML$, Optional typeName$ = Nothing, Optional xmlns$ = Nothing, Optional forceLargeMode As Boolean = False) As IEnumerable(Of T) |
178 | Dim nodeName$ = GetType(T).GetTypeName([default]:=typeName) |
179 | Dim source As IEnumerable(Of String) |
180 | |
181 | If forceLargeMode OrElse XML.FileLength > 1024 * 1024 * 128 Then |
182 | ' 这是一个超大的XML文档 |
183 | source = NodeIterator.IterateArrayNodes(XML, nodeName) |
184 | xmlns = Nothing |
185 | Else |
186 | source = InternalIterates(XML, nodeName) |
187 | End If |
188 | |
189 | Return source.NodeInstanceBuilder(Of T)(xmlns, xmlNode:=nodeName) |
190 | End Function |
191 | |
192 | ''' <summary> |
193 | ''' 从给定的文本之中利用反序列化从XML字符串构建出.NET对象 |
194 | ''' </summary> |
195 | ''' <typeparam name="T"></typeparam> |
196 | ''' <param name="nodes"></param> |
197 | ''' <param name="replaceXmlns$"></param> |
198 | ''' <param name="xmlNode$">文件之中的节点名称</param> |
199 | ''' <returns></returns> |
200 | <Extension> |
201 | Private Iterator Function NodeInstanceBuilder(Of T As Class)(nodes As IEnumerable(Of String), replaceXmlns$, xmlNode$) As IEnumerable(Of T) |
202 | Dim handle As New DeserializeHandler(Of T)(xmlNode) With { |
203 | .ReplaceXmlns = replaceXmlns |
204 | } |
205 | |
206 | For Each xml As String In nodes |
207 | Yield handle.LoadXml(xml) |
208 | Next |
209 | End Function |
210 | |
211 | ''' <summary> |
212 | ''' Apply on a ultra large size XML database, which its data size is greater than 1GB to 100GB or even more. |
213 | ''' </summary> |
214 | ''' <typeparam name="T"></typeparam> |
215 | ''' <param name="path$"></param> |
216 | ''' <param name="typeName$"></param> |
217 | ''' <param name="xmlns$"></param> |
218 | ''' <returns></returns> |
219 | <Extension> |
220 | Public Function LoadUltraLargeXMLDataSet(Of T As Class)(path$, Optional typeName$ = Nothing, Optional xmlns$ = Nothing) As IEnumerable(Of T) |
221 | Dim nodeName$ = GetType(T).GetTypeName([default]:=typeName) |
222 | Return nodeName _ |
223 | .UltraLargeXmlNodesIterator(path) _ |
224 | .NodeInstanceBuilder(Of T)(xmlns, xmlNode:=nodeName) |
225 | End Function |
226 | |
227 | <Extension> |
228 | Private Iterator Function UltraLargeXmlNodesIterator(nodeName$, path$) As IEnumerable(Of String) |
229 | Dim el As New Value(Of XElement) |
230 | Dim XML$ |
231 | |
232 | Using reader As XmlReader = XmlReader.Create(path) |
233 | |
234 | Call reader.MoveToContent() |
235 | |
236 | Do While (reader.Read()) |
237 | ' Parse the file And return each of the child_node |
238 | If (reader.NodeType = XmlNodeType.Element AndAlso reader.Name = nodeName) Then |
239 | If (Not (el = XNode.ReadFrom(reader)) Is Nothing) Then |
240 | XML = el.Value.ToString |
241 | Yield XML |
242 | End If |
243 | End If |
244 | Loop |
245 | End Using |
246 | End Function |
247 | |
248 | <MethodImpl(MethodImplOptions.AggressiveInlining)> |
249 | <Extension> |
250 | Public Function PopulateXmlElementText(Of T As Class)(path$, Optional typeName$ = Nothing) As IEnumerable(Of String) |
251 | Return GetType(T) _ |
252 | .GetTypeName([default]:=typeName) _ |
253 | .UltraLargeXmlNodesIterator(path) |
254 | End Function |
255 | End Module |
256 | End Namespace |