| 1 | <?xml version="1.0" encoding="UTF-8"?> |
|---|
| 2 | <!-- |
|---|
| 3 | Licensed to the Apache Software Foundation (ASF) under one or more |
|---|
| 4 | contributor license agreements. See the NOTICE file distributed with |
|---|
| 5 | this work for additional information regarding copyright ownership. |
|---|
| 6 | The ASF licenses this file to You under the Apache License, Version 2.0 |
|---|
| 7 | (the "License"); you may not use this file except in compliance with |
|---|
| 8 | the License. You may obtain a copy of the License at |
|---|
| 9 | |
|---|
| 10 | http://www.apache.org/licenses/LICENSE-2.0 |
|---|
| 11 | |
|---|
| 12 | Unless required by applicable law or agreed to in writing, software |
|---|
| 13 | distributed under the License is distributed on an "AS IS" BASIS, |
|---|
| 14 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
|---|
| 15 | See the License for the specific language governing permissions and |
|---|
| 16 | limitations under the License. |
|---|
| 17 | |
|---|
| 18 | Description: This xml file defines the valid mime types used by Tika. |
|---|
| 19 | The mime types within this file are based on the types in the mime-types.xml |
|---|
| 20 | file available in Apache Nutch. |
|---|
| 21 | --> |
|---|
| 22 | |
|---|
| 23 | <mime-info> |
|---|
| 24 | |
|---|
| 25 | <mime-type type="text/plain"> |
|---|
| 26 | <magic priority="50"> |
|---|
| 27 | <match value="This is TeX," type="string" offset="0" /> |
|---|
| 28 | <match value="This is METAFONT," type="string" offset="0" /> |
|---|
| 29 | </magic> |
|---|
| 30 | <glob pattern="*.txt" /> |
|---|
| 31 | <glob pattern="*.asc" /> |
|---|
| 32 | </mime-type> |
|---|
| 33 | |
|---|
| 34 | <mime-type type="text/html"> |
|---|
| 35 | <magic priority="50"> |
|---|
| 36 | <match value="<!DOCTYPE HTML" type="string" |
|---|
| 37 | offset="0:64" /> |
|---|
| 38 | <match value="<!doctype html" type="string" |
|---|
| 39 | offset="0:64" /> |
|---|
| 40 | <match value="<HEAD" type="string" offset="0:64" /> |
|---|
| 41 | <match value="<head" type="string" offset="0:64" /> |
|---|
| 42 | <match value="<TITLE" type="string" offset="0:64" /> |
|---|
| 43 | <match value="<title" type="string" offset="0:64" /> |
|---|
| 44 | <match value="<html" type="string" offset="0:64" /> |
|---|
| 45 | <match value="<HTML" type="string" offset="0:64" /> |
|---|
| 46 | <match value="<BODY" type="string" offset="0" /> |
|---|
| 47 | <match value="<body" type="string" offset="0" /> |
|---|
| 48 | <match value="<TITLE" type="string" offset="0" /> |
|---|
| 49 | <match value="<title" type="string" offset="0" /> |
|---|
| 50 | <match value="<!--" type="string" offset="0" /> |
|---|
| 51 | <match value="<h1" type="string" offset="0" /> |
|---|
| 52 | <match value="<H1" type="string" offset="0" /> |
|---|
| 53 | <match value="<!doctype HTML" type="string" offset="0" /> |
|---|
| 54 | <match value="<!DOCTYPE html" type="string" offset="0" /> |
|---|
| 55 | </magic> |
|---|
| 56 | <glob pattern="*.html" /> |
|---|
| 57 | <glob pattern="*.htm" /> |
|---|
| 58 | </mime-type> |
|---|
| 59 | |
|---|
| 60 | <mime-type type="application/xhtml+xml"> |
|---|
| 61 | <glob pattern="*.xhtml" /> |
|---|
| 62 | <root-XML namespaceURI='http://www.w3.org/1999/xhtml' |
|---|
| 63 | localName='html' /> |
|---|
| 64 | </mime-type> |
|---|
| 65 | |
|---|
| 66 | <mime-type type="application/vnd.ms-powerpoint"> |
|---|
| 67 | <glob pattern="*.ppz" /> |
|---|
| 68 | <glob pattern="*.ppt" /> |
|---|
| 69 | <glob pattern="*.pps" /> |
|---|
| 70 | <glob pattern="*.pot" /> |
|---|
| 71 | <magic priority="50"> |
|---|
| 72 | <match value="0xcfd0e011" type="little32" offset="0" /> |
|---|
| 73 | </magic> |
|---|
| 74 | </mime-type> |
|---|
| 75 | |
|---|
| 76 | <mime-type type="application/vnd.ms-excel"> |
|---|
| 77 | <magic priority="50"> |
|---|
| 78 | <match value="Microsoft Excel 5.0 Worksheet" type="string" |
|---|
| 79 | offset="2080" /> |
|---|
| 80 | </magic> |
|---|
| 81 | <glob pattern="*.xls" /> |
|---|
| 82 | <glob pattern="*.xlc" /> |
|---|
| 83 | <glob pattern="*.xll" /> |
|---|
| 84 | <glob pattern="*.xlm" /> |
|---|
| 85 | <glob pattern="*.xlw" /> |
|---|
| 86 | <glob pattern="*.xla" /> |
|---|
| 87 | <glob pattern="*.xlt" /> |
|---|
| 88 | <glob pattern="*.xld" /> |
|---|
| 89 | <alias type="application/msexcel" /> |
|---|
| 90 | </mime-type> |
|---|
| 91 | |
|---|
| 92 | <mime-type type="application/vnd.oasis.opendocument.text"> |
|---|
| 93 | <glob pattern="*.odt" /> |
|---|
| 94 | </mime-type> |
|---|
| 95 | |
|---|
| 96 | |
|---|
| 97 | <mime-type type="application/zip"> |
|---|
| 98 | <alias type="application/x-zip-compressed" /> |
|---|
| 99 | <magic priority="40"> |
|---|
| 100 | <match value="PK\003\004" type="string" offset="0" /> |
|---|
| 101 | </magic> |
|---|
| 102 | <glob pattern="*.zip" /> |
|---|
| 103 | </mime-type> |
|---|
| 104 | |
|---|
| 105 | <mime-type type="application/vnd.oasis.opendocument.text"> |
|---|
| 106 | <glob pattern="*.oth" /> |
|---|
| 107 | </mime-type> |
|---|
| 108 | |
|---|
| 109 | <mime-type type="application/msword"> |
|---|
| 110 | <magic priority="50"> |
|---|
| 111 | <match value="\x31\xbe\x00\x00" type="string" offset="0" /> |
|---|
| 112 | <match value="PO^Q`" type="string" offset="0" /> |
|---|
| 113 | <match value="\376\067\0\043" type="string" offset="0" /> |
|---|
| 114 | <match value="\333\245-\0\0\0" type="string" offset="0" /> |
|---|
| 115 | <match value="Microsoft Word 6.0 Document" type="string" |
|---|
| 116 | offset="2080" /> |
|---|
| 117 | <match value="Microsoft Word document data" type="string" |
|---|
| 118 | offset="2112" /> |
|---|
| 119 | </magic> |
|---|
| 120 | <glob pattern="*.doc" /> |
|---|
| 121 | <alias type="application/vnd.ms-word" /> |
|---|
| 122 | </mime-type> |
|---|
| 123 | |
|---|
| 124 | <mime-type type="application/octet-stream"> |
|---|
| 125 | <magic priority="50"> |
|---|
| 126 | <match value="\037\036" type="string" offset="0" /> |
|---|
| 127 | <match value="017437" type="host16" offset="0" /> |
|---|
| 128 | <match value="0x1fff" type="host16" offset="0" /> |
|---|
| 129 | <match value="\377\037" type="string" offset="0" /> |
|---|
| 130 | <match value="0145405" type="host16" offset="0" /> |
|---|
| 131 | </magic> |
|---|
| 132 | <glob pattern="*.bin" /> |
|---|
| 133 | </mime-type> |
|---|
| 134 | |
|---|
| 135 | <mime-type type="application/pdf"> |
|---|
| 136 | <magic priority="50"> |
|---|
| 137 | <match value="%PDF-" type="string" offset="0" /> |
|---|
| 138 | </magic> |
|---|
| 139 | <glob pattern="*.pdf" /> |
|---|
| 140 | <alias type="application/x-pdf" /> |
|---|
| 141 | </mime-type> |
|---|
| 142 | |
|---|
| 143 | <mime-type type="application/atom+xml"> |
|---|
| 144 | <root-XML localName="feed" |
|---|
| 145 | namespaceURI="http://purl.org/atom/ns#" /> |
|---|
| 146 | </mime-type> |
|---|
| 147 | |
|---|
| 148 | <mime-type type="application/mac-binhex40"> |
|---|
| 149 | <glob pattern="*.hqx" /> |
|---|
| 150 | </mime-type> |
|---|
| 151 | |
|---|
| 152 | <mime-type type="application/mac-compactpro"> |
|---|
| 153 | <glob pattern="*.cpt" /> |
|---|
| 154 | </mime-type> |
|---|
| 155 | |
|---|
| 156 | <mime-type type="application/rtf"> |
|---|
| 157 | <glob pattern="*.rtf"/> |
|---|
| 158 | <alias type="text/rtf" /> |
|---|
| 159 | </mime-type> |
|---|
| 160 | |
|---|
| 161 | <mime-type type="application/rss+xml"> |
|---|
| 162 | <alias type="text/rss" /> |
|---|
| 163 | <root-XML localName="rss" /> |
|---|
| 164 | <root-XML namespaceURI="http://purl.org/rss/1.0/" /> |
|---|
| 165 | <glob pattern="*.rss" /> |
|---|
| 166 | </mime-type> |
|---|
| 167 | |
|---|
| 168 | <!-- added in by mattmann --> |
|---|
| 169 | <mime-type type="application/xml"> |
|---|
| 170 | <alias type="text/xml" /> |
|---|
| 171 | <glob pattern="*.xml" /> |
|---|
| 172 | </mime-type> |
|---|
| 173 | |
|---|
| 174 | <mime-type type="application/x-mif"> |
|---|
| 175 | <alias type="application/vnd.mif" /> |
|---|
| 176 | </mime-type> |
|---|
| 177 | |
|---|
| 178 | <mime-type type="application/vnd.wap.wbxml"> |
|---|
| 179 | <glob pattern="*.wbxml" /> |
|---|
| 180 | </mime-type> |
|---|
| 181 | |
|---|
| 182 | <mime-type type="application/vnd.wap.wmlc"> |
|---|
| 183 | <_comment>Compiled WML Document</_comment> |
|---|
| 184 | <glob pattern="*.wmlc" /> |
|---|
| 185 | </mime-type> |
|---|
| 186 | |
|---|
| 187 | <mime-type type="application/vnd.wap.wmlscriptc"> |
|---|
| 188 | <_comment>Compiled WML Script</_comment> |
|---|
| 189 | <glob pattern="*.wmlsc" /> |
|---|
| 190 | </mime-type> |
|---|
| 191 | |
|---|
| 192 | <mime-type type="text/vnd.wap.wmlscript"> |
|---|
| 193 | <_comment>WML Script</_comment> |
|---|
| 194 | <glob pattern="*.wmls" /> |
|---|
| 195 | </mime-type> |
|---|
| 196 | |
|---|
| 197 | <mime-type type="application/x-bzip"> |
|---|
| 198 | <alias type="application/x-bzip2" /> |
|---|
| 199 | </mime-type> |
|---|
| 200 | |
|---|
| 201 | <mime-type type="application/x-bzip-compressed-tar"> |
|---|
| 202 | <glob pattern="*.tbz" /> |
|---|
| 203 | <glob pattern="*.tbz2" /> |
|---|
| 204 | </mime-type> |
|---|
| 205 | |
|---|
| 206 | <mime-type type="application/x-cdlink"> |
|---|
| 207 | <_comment>Virtual CD-ROM CD Image File</_comment> |
|---|
| 208 | <glob pattern="*.vcd" /> |
|---|
| 209 | </mime-type> |
|---|
| 210 | |
|---|
| 211 | <mime-type type="application/x-director"> |
|---|
| 212 | <_comment>Shockwave Movie</_comment> |
|---|
| 213 | <glob pattern="*.dcr" /> |
|---|
| 214 | <glob pattern="*.dir" /> |
|---|
| 215 | <glob pattern="*.dxr" /> |
|---|
| 216 | </mime-type> |
|---|
| 217 | |
|---|
| 218 | <mime-type type="application/x-futuresplash"> |
|---|
| 219 | <_comment>Macromedia FutureSplash File</_comment> |
|---|
| 220 | <glob pattern="*.spl" /> |
|---|
| 221 | </mime-type> |
|---|
| 222 | |
|---|
| 223 | <mime-type type="application/x-java"> |
|---|
| 224 | <alias type="application/java" /> |
|---|
| 225 | </mime-type> |
|---|
| 226 | |
|---|
| 227 | <mime-type type="application/x-koan"> |
|---|
| 228 | <_comment>SSEYO Koan File</_comment> |
|---|
| 229 | <glob pattern="*.skp" /> |
|---|
| 230 | <glob pattern="*.skd" /> |
|---|
| 231 | <glob pattern="*.skt" /> |
|---|
| 232 | <glob pattern="*.skm" /> |
|---|
| 233 | </mime-type> |
|---|
| 234 | |
|---|
| 235 | <mime-type type="application/x-latex"> |
|---|
| 236 | <_comment>LaTeX Source Document</_comment> |
|---|
| 237 | <glob pattern="*.latex" /> |
|---|
| 238 | </mime-type> |
|---|
| 239 | |
|---|
| 240 | <!-- JC CHANGED |
|---|
| 241 | <mime-type type="application/x-mif"> |
|---|
| 242 | <_comment>FrameMaker MIF document</_comment> |
|---|
| 243 | <glob pattern="*.mif"/> |
|---|
| 244 | </mime-type> --> |
|---|
| 245 | |
|---|
| 246 | <mime-type type="application/x-ms-dos-executable"> |
|---|
| 247 | <alias type="application/x-dosexec" /> |
|---|
| 248 | </mime-type> |
|---|
| 249 | |
|---|
| 250 | <mime-type type="application/ogg"> |
|---|
| 251 | <alias type="application/x-ogg" /> |
|---|
| 252 | </mime-type> |
|---|
| 253 | |
|---|
| 254 | <mime-type type="application/x-rar"> |
|---|
| 255 | <alias type="application/x-rar-compressed" /> |
|---|
| 256 | </mime-type> |
|---|
| 257 | |
|---|
| 258 | <mime-type type="application/x-shellscript"> |
|---|
| 259 | <alias type="application/x-sh" /> |
|---|
| 260 | </mime-type> |
|---|
| 261 | |
|---|
| 262 | <mime-type type="application/xhtml+xml"> |
|---|
| 263 | <glob pattern="*.xht" /> |
|---|
| 264 | </mime-type> |
|---|
| 265 | |
|---|
| 266 | <mime-type type="audio/midi"> |
|---|
| 267 | <glob pattern="*.kar" /> |
|---|
| 268 | </mime-type> |
|---|
| 269 | |
|---|
| 270 | <mime-type type="audio/x-pn-realaudio"> |
|---|
| 271 | <alias type="audio/x-realaudio" /> |
|---|
| 272 | </mime-type> |
|---|
| 273 | |
|---|
| 274 | <mime-type type="image/tiff"> |
|---|
| 275 | <magic priority="50"> |
|---|
| 276 | <match value="0x4d4d2a00" type="string" offset="0" /> |
|---|
| 277 | <match value="0x49492a00" type="string" offset="0" /> |
|---|
| 278 | </magic> |
|---|
| 279 | </mime-type> |
|---|
| 280 | |
|---|
| 281 | <mime-type type="message/rfc822"> |
|---|
| 282 | <magic priority="50"> |
|---|
| 283 | <match type="string" value="Relay-Version:" offset="0" /> |
|---|
| 284 | <match type="string" value="#! rnews" offset="0" /> |
|---|
| 285 | <match type="string" value="N#! rnews" offset="0" /> |
|---|
| 286 | <match type="string" value="Forward to" offset="0" /> |
|---|
| 287 | <match type="string" value="Pipe to" offset="0" /> |
|---|
| 288 | <match type="string" value="Return-Path:" offset="0" /> |
|---|
| 289 | <match type="string" value="From:" offset="0" /> |
|---|
| 290 | <match type="string" value="Message-ID:" offset="0" /> |
|---|
| 291 | <match type="string" value="Date:" offset="0" /> |
|---|
| 292 | </magic> |
|---|
| 293 | </mime-type> |
|---|
| 294 | |
|---|
| 295 | <mime-type type="application/x-javascript"> |
|---|
| 296 | <glob pattern="*.js" /> |
|---|
| 297 | </mime-type> |
|---|
| 298 | |
|---|
| 299 | |
|---|
| 300 | <mime-type type="image/vnd.wap.wbmp"> |
|---|
| 301 | <_comment>Wireless Bitmap File Format</_comment> |
|---|
| 302 | <glob pattern="*.wbmp" /> |
|---|
| 303 | </mime-type> |
|---|
| 304 | |
|---|
| 305 | <mime-type type="image/x-psd"> |
|---|
| 306 | <alias type="image/photoshop" /> |
|---|
| 307 | </mime-type> |
|---|
| 308 | |
|---|
| 309 | <mime-type type="image/x-xcf"> |
|---|
| 310 | <alias type="image/xcf" /> |
|---|
| 311 | <magic priority="50"> |
|---|
| 312 | <match type="string" value="gimp xcf " offset="0" /> |
|---|
| 313 | </magic> |
|---|
| 314 | </mime-type> |
|---|
| 315 | |
|---|
| 316 | <mime-type type="application/x-shockwave-flash"> |
|---|
| 317 | <glob pattern="*.swf"/> |
|---|
| 318 | <magic priority="50"> |
|---|
| 319 | <match type="string" value="FWS" offset="0"/> |
|---|
| 320 | <match type="string" value="CWS" offset="0"/> |
|---|
| 321 | </magic> |
|---|
| 322 | </mime-type> |
|---|
| 323 | |
|---|
| 324 | <mime-type type="model/iges"> |
|---|
| 325 | <_comment> |
|---|
| 326 | Initial Graphics Exchange Specification Format |
|---|
| 327 | </_comment> |
|---|
| 328 | <glob pattern="*.igs" /> |
|---|
| 329 | <glob pattern="*.iges" /> |
|---|
| 330 | </mime-type> |
|---|
| 331 | |
|---|
| 332 | <mime-type type="model/mesh"> |
|---|
| 333 | <glob pattern="*.msh" /> |
|---|
| 334 | <glob pattern="*.mesh" /> |
|---|
| 335 | <glob pattern="*.silo" /> |
|---|
| 336 | </mime-type> |
|---|
| 337 | |
|---|
| 338 | <mime-type type="model/vrml"> |
|---|
| 339 | <glob pattern="*.vrml" /> |
|---|
| 340 | </mime-type> |
|---|
| 341 | |
|---|
| 342 | <mime-type type="text/x-tcl"> |
|---|
| 343 | <alias type="application/x-tcl" /> |
|---|
| 344 | </mime-type> |
|---|
| 345 | |
|---|
| 346 | <mime-type type="text/x-tex"> |
|---|
| 347 | <alias type="application/x-tex" /> |
|---|
| 348 | </mime-type> |
|---|
| 349 | |
|---|
| 350 | <mime-type type="text/x-texinfo"> |
|---|
| 351 | <alias type="application/x-texinfo" /> |
|---|
| 352 | </mime-type> |
|---|
| 353 | |
|---|
| 354 | <mime-type type="text/x-troff-me"> |
|---|
| 355 | <alias type="application/x-troff-me" /> |
|---|
| 356 | </mime-type> |
|---|
| 357 | |
|---|
| 358 | <mime-type type="video/vnd.mpegurl"> |
|---|
| 359 | <glob pattern="*.mxu" /> |
|---|
| 360 | </mime-type> |
|---|
| 361 | |
|---|
| 362 | <mime-type type="x-conference/x-cooltalk"> |
|---|
| 363 | <_comment>Cooltalk Audio</_comment> |
|---|
| 364 | <glob pattern="*.ice" /> |
|---|
| 365 | </mime-type> |
|---|
| 366 | |
|---|
| 367 | </mime-info> |
|---|
