doc/graph_io.dox
author deba
Mon, 04 Jul 2005 17:22:03 +0000
changeset 1538 777834118f73
parent 1527 7ceab500e1f6
child 1540 7d028a73d7f2
permissions -rw-r--r--
NewUndirEdgeSetAdaptor class
some doc
some bug fix
     1 namespace lemon {
     2 /*!
     3 
     4 
     5 \page graph-io-page Graph Input-Output
     6 
     7 The standard graph IO enables to store graphs and additional maps
     8 in a flexible and efficient way. 
     9 
    10 \section format The general file format
    11 
    12 The file contains sections in the following order:
    13 
    14 \li nodeset
    15 \li edgeset
    16 \li nodes
    17 \li edges
    18 \li attributes
    19 
    20 The nodeset section starts with the following line:
    21 
    22 <tt>\@nodeset</tt>
    23 
    24 The next line contains the names of the nodemaps, separated by whitespaces.  Each
    25 following line describes a node in the graph: it contains the values of the
    26 maps in the right order. The map named "id" should contain unique values
    27 because it is regarded as an ID-map. For example:
    28 
    29 \code
    30 @nodeset
    31 id  x-coord  y-coord  color
    32 3   1.0      4.0      blue
    33 5   2.3      5.7      red
    34 12  7.8      2.3      green
    35 \endcode
    36 
    37 The edgeset section is very similar to the nodeset section, it has
    38 the same coloumn oriented structure. It starts with the line 
    39 
    40 <tt>\@edgeset</tt>
    41 
    42 The next line contains the whitespace separated list of names of the maps.
    43 Each of the next lines describes one edge. The first two elements in the line
    44 are the IDs of the source and target (or tail and head) node of the edge as they occur in the ID node
    45 map. You can also have an optional ID map on the edges for later reference.
    46 
    47 \code
    48 @edgeset
    49              id    weight   label
    50 3   5        a     4.3      a-edge
    51 5   12       c     2.6      c-edge
    52 3   12       g     3.4      g-edge
    53 \endcode
    54 
    55 The next section contains <em>labeled nodes</em> (i.e. nodes having a special
    56 label on them). The section starts with
    57 
    58 <tt> \@nodes </tt>
    59 
    60 Each of the next lines contains a label for a node in the graph 
    61 and then the ID described in the nodeset section.
    62 
    63 \code
    64 @nodes 
    65 source 3
    66 target 12
    67 \endcode
    68 
    69 The last section describes the <em>labeled edges</em>
    70 (i.e. edges having a special label on them). It starts with \c \@edges
    71 and then each line contains the name of the edge and the ID.
    72 
    73 \code
    74 @nodes 
    75 observed c
    76 \endcode
    77 
    78 
    79 The file may contain empty lines and comment lines. The comment lines
    80 start with an \c # character.
    81 
    82 The attributes section can handle some information about the graph. It
    83 contains in each line an key and the mapped value to key. The key should
    84 be a string without whitespace, the value can be from various type.
    85 
    86 \code
    87 @attributes
    88 title "Four colored plan graph"
    89 author "Balazs DEZSO"
    90 copyright "Lemon Library"
    91 version 12
    92 \endcode
    93 
    94 \code
    95 @end
    96 \endcode
    97 =======
    98 The file ends with the 
    99 
   100 <tt> \@end </tt>
   101 
   102 line.
   103 
   104 
   105 \section use Using graph input-output
   106 The graph input and output is based on reading and writing  commands. The user
   107 adds reading and writing commands to the reader or writer class, then he
   108 calls the \c run() method that executes all the given commands.
   109 
   110 \subsection write Writing a graph
   111 
   112 The \c GraphWriter class provides the graph output. To write a graph
   113 you should first give writing commands to the writer. You can declare
   114 write command as \c NodeMap or \c EdgeMap writing and labeled Node and
   115 Edge writing.
   116 
   117 \code
   118 GraphWriter<ListGraph> writer(std::cout, graph);
   119 \endcode
   120 
   121 The \c writeNodeMap() function declares a \c NodeMap writing command in the
   122 \c GraphWriter. You should give a name of the map and the map
   123 object as parameters. The NodeMap writing command with name "id" should write a 
   124 unique map because it is regarded as ID map.
   125 
   126 \see IdMap, DescriptorMap  
   127 
   128 \code
   129 IdMap<ListGraph, Node> nodeIdMap;
   130 writer.writeNodeMap("id", nodeIdMap);
   131 
   132 writer.writeNodeMap("x-coord", xCoordMap);
   133 writer.writeNodeMap("y-coord", yCoordMap);
   134 writer.writeNodeMap("color", colorMap);
   135 \endcode
   136 
   137 With the \c writeEdgeMap() member function you can give an edge map
   138 writing command similar to the NodeMaps.
   139 
   140 \see IdMap, DescriptorMap  
   141 
   142 \code
   143 DescriptorMap<ListGraph, Edge, ListGraph::EdgeMap<int> > edgeDescMap(graph);
   144 writer.writeEdgeMap("descriptor", edgeDescMap);
   145 
   146 writer.writeEdgeMap("weight", weightMap);
   147 writer.writeEdgeMap("label", labelMap);
   148 \endcode
   149 
   150 With \c writeNode() and \c writeEdge() functions you can designate Nodes and
   151 Edges in the graph. For example, you can write out the source and target node
   152 of a maximum flow instance.
   153 
   154 \code
   155 writer.writeNode("source", sourceNode);
   156 writer.writeNode("target", targetNode);
   157 
   158 writer.writeEdge("observed", edge);
   159 \endcode
   160 
   161 With \c writeAttribute() function you can write an attribute to the file.
   162 
   163 \code
   164 writer.writeAttribute("author", "Balazs DEZSO");
   165 writer.writeAttribute("version", 12);
   166 \endcode
   167 
   168 After you give all write commands you must call the \c run() member
   169 function, which executes all the writing commands.
   170 
   171 \code
   172 writer.run();
   173 \endcode
   174 
   175 \subsection reading Reading a graph
   176 
   177 The given file format may contain several maps and labeled nodes or edges.
   178 If you read a graph you need not read all the maps and items just those
   179 that you need. The interface of the \c GraphReader is very similar to
   180 the GraphWriter but the reading method does not depend on the order of the
   181 given commands.
   182 
   183 The reader object assumes that each not readed value does not contain 
   184 whitespaces, therefore it has some extra possibilities to control how
   185 it should skip the values when the string representation contains spaces.
   186 
   187 \code
   188 GraphReader<ListGraph> reader(std::cin, graph);
   189 \endcode
   190 
   191 The \c readNodeMap() function reads a map from the \c \@nodeset section.
   192 If there is a map that you do not want to read from the file and there are
   193 whitespaces in the string represenation of the values then you should
   194 call the \c skipNodeMap() template member function with proper parameters.
   195 
   196 \see QuotedStringReader
   197 
   198 \code
   199 reader.readNodeMap("x-coord", xCoordMap);
   200 reader.readNodeMap("y-coord", yCoordMap);
   201 
   202 reader.readNodeMap<QuotedStringReader>("label", labelMap);
   203 reader.skipNodeMap<QuotedStringReader>("description");
   204 
   205 reader.readNodeMap("color", colorMap);
   206 \endcode
   207 
   208 With the \c readEdgeMap() member function you can give an edge map
   209 reading command similar to the NodeMaps. 
   210 
   211 \code
   212 reader.readEdgeMap("weight", weightMap);
   213 reader.readEdgeMap("label", labelMap);
   214 \endcode
   215 
   216 With \c readNode() and \c readEdge() functions you can read labeled Nodes and
   217 Edges.
   218 
   219 \code
   220 reader.readNode("source", sourceNode);
   221 reader.readNode("target", targetNode);
   222 
   223 reader.readEdge("observed", edge);
   224 \endcode
   225 
   226 With \c readAttribute() function you can read an attribute from the file.
   227 
   228 \code
   229 std::string author;
   230 writer.readAttribute("author", author);
   231 int version;
   232 writer.writeAttribute("version", version);
   233 \endcode
   234 
   235 After you give all read commands you must call the \c run() member
   236 function, which executes all the commands.
   237 
   238 \code
   239 reader.run();
   240 \endcode
   241 
   242 \section types Background of Reading and Writing
   243 To read a map (on the nodes or edges)
   244 the \c GraphReader should know how to read a Value from the given map.
   245 By the default implementation the input operator reads a value from
   246 the stream and the type of the readed value is the value type of the given map.
   247 When the reader should skip a value in the stream, because you do not
   248 want to store it in a map, the reader skips a character sequence without 
   249 whitespace. 
   250 
   251 If you want to change the functionality of the reader, you can use
   252 template parameters to specialize it. When you give a reading
   253 command for a map you can give a Reader type as template parameter.
   254 With this template parameter you can control how the Reader reads
   255 a value from the stream.
   256 
   257 The reader has the next structure: 
   258 \code
   259 struct TypeReader {
   260   typedef TypeName Value;
   261 
   262   void read(std::istream& is, Value& value);
   263 };
   264 \endcode
   265 
   266 For example, the \c "strings" nodemap contains strings and you do not need
   267 the value of the string just the length. Then you can implement own Reader
   268 struct.
   269 
   270 \code
   271 struct LengthReader {
   272   typedef int Value;
   273 
   274   void read(std::istream& is, Value& value) {
   275     std::string tmp;
   276     is >> tmp;
   277     value = tmp.length();
   278   }
   279 };
   280 ...
   281 reader.readNodeMap<LengthReader>("strings", lengthMap);
   282 \endcode  
   283 
   284 The global functionality of the reader class can be changed by giving a
   285 special template parameter to the GraphReader class. By default, the
   286 template parameter is \c DefaultReaderTraits. A reader traits class 
   287 should provide an inner template class Reader for each type, and an 
   288 DefaultReader for skipping a value.
   289 
   290 The specialization of  writing should be very similar to that of reading.
   291 
   292 \section undir Undir graphs
   293 
   294 In the undir graph format there is an \c undiredgeset section instead of
   295 the \c edgeset section. The first line of the section describes the 
   296 undirected egdes' names and all next lines describes one undirected edge
   297 with the the incident nodes and the values of the map.
   298 
   299 The format handles the directed edge maps as a syntactical sugar, if there
   300 is two map which names are the same with a \c '+' and a \c '-' prefix
   301 then it can be read as an directed map.
   302 
   303 \code
   304 @undiredgeset
   305              id    capacity +flow -flow
   306 32   2       1     4.3      2.0	  0.0
   307 21   21      5     2.6      0.0   2.6
   308 21   12      8     3.4      0.0   0.0
   309 \endcode
   310 
   311 The \c edges section changed to \c undiredges section. This section
   312 describes labeled edges and undirected edges. The directed edge label
   313 should start with a \c '+' and a \c '-' prefix what decide the direction
   314 of the edge. 
   315 
   316 \code
   317 @undiredges
   318 undiredge 1
   319 +edge 5
   320 -back 5
   321 \endcode
   322 
   323 There are similar classes to the \c GraphReader ans \c GraphWriter
   324 which handle the undirected graphs. These classes are the 
   325 \c UndirGraphReader and \UndirGraphWriter.
   326 
   327 The \c readUndirMap() function reads an undirected map and the
   328 \c readUndirEdge() reads an undirected edge from the file, 
   329 
   330 \code
   331 reader.readUndirEdgeMap("capacity", capacityMap);
   332 reader.readEdgeMap("flow", flowMap);
   333 ...
   334 reader.readUndirEdge("undir_edge", undir_edge);
   335 reader.readEdge("edge", edge);
   336 \endcode
   337 
   338 \section advanced Advanced features
   339 
   340 The graph reader and writer classes gives an easy way to read and write
   341 graphs. But sometimes we want more advanced features. This way we can
   342 use the more general lemon reader and writer interface.
   343 
   344 The lemon format is an section oriented file format. It contains one or
   345 more section, each starts with a line with \c \@ first character.
   346 The content of the section this way cannot contain line with \c \@ first
   347 character. The file may contains comment lines with \c # first character.
   348 
   349 The \c LemonReader and \c LemonWriter gives a framework to read and
   350 write sections. There are various section reader and section writer
   351 classes which can be attached to a \c LemonReader or a \c LemonWriter.
   352 
   353 There are default section readers and writers for reading and writing
   354 item sets, and labeled items in the graph. These reads and writes
   355 the format described above. Other type of data can be handled with own
   356 section reader and writer classes which are inherited from the
   357 \c LemonReader::SectionReader or the \c LemonWriter::SectionWriter classes.
   358 
   359 The next example defines a special section reader which reads the
   360 \c \@description sections into a string:
   361 
   362 \code 
   363 class DescriptionReader : LemonReader::SectionReader {
   364 protected:
   365   virtual bool header(const std::string& line) {
   366     std::istringstream ls(line);
   367     std::string head;
   368     ls >> head;
   369     return head == "@description";
   370   }
   371 
   372   virtual void read(std::istream& is) {
   373     std::string line;
   374     while (getline(is, line)) {
   375       desc += line;
   376     }
   377   }
   378 public:
   379 
   380   typedef LemonReader::SectionReader Parent;
   381   
   382   DescriptionReader(LemonReader& reader) : Parent(reader) {}
   383 
   384   const std::string& description() const {
   385     return description;
   386   }
   387 
   388 private:
   389   std::string desc;
   390 };
   391 \endcode
   392 
   393 The other advanced stuff of the generalized file format is that 
   394 multiple edgesets can be stored to the same nodeset. It can be used 
   395 by example as a network traffic matrix.
   396 
   397 In example there is a network with symmetric links and there are assymetric
   398 traffic request on the network. This construction can be stored in an
   399 undirected graph and in an directed NewEdgeSetAdaptor class. The example
   400 shows the input with the LemonReader class:
   401 
   402 \code
   403 UndirListGraph network;
   404 UndirListGraph::UndirEdgeSet<double> capacity;
   405 NewEdgeSetAdaptor<UndirListGraph> traffic(network);
   406 NewEdgeSetAdaptor<UndirListGraph>::EdgeSet<double> request(network);
   407 
   408 LemonReader reader(std::cin);
   409 NodeSetReader nodesetReader(reader, network);
   410 UndirEdgeSetReader undirEdgesetReader(reader, network, nodesetReader);
   411 undirEdgesetReader.readEdgeMap("capacity", capacity);
   412 EdgeSetReader edgesetReader(reader, traffic, nodesetReader);
   413 edgesetReader.readEdgeMap("request", request);
   414 
   415 reader.run();
   416 \endcode
   417 
   418 Because the GraphReader and the UndirGraphReader can be converted
   419 to LemonReader and it can resolve the ID's of the items, the previous
   420 result can be achived with the UndirGraphReader class also.
   421 
   422 
   423 \code
   424 UndirListGraph network;
   425 UndirListGraph::UndirEdgeSet<double> capacity;
   426 NewEdgeSetAdaptor<UndirListGraph> traffic(network);
   427 NewEdgeSetAdaptor<UndirListGraph>::EdgeSet<double> request(network);
   428 
   429 UndirGraphReader reader(std::cin, network);
   430 reader.readEdgeMap("capacity", capacity);
   431 EdgeSetReader edgesetReader(reader, traffic, reader);
   432 edgesetReader.readEdgeMap("request", request);
   433 
   434 reader.run();
   435 \endcode
   436 
   437 \author Balazs Dezso
   438 */
   439 }