doc/graph_io.dox
author alpar
Wed, 12 Jul 2006 11:40:52 +0000
changeset 2135 15355b98fb84
parent 1959 264811b995f3
child 2391 14a343be7a5a
permissions -rw-r--r--
Must have been changed because of the private copy constructor.
alpar@1118
     1
namespace lemon {
deba@1114
     2
/*!
deba@1114
     3
deba@1114
     4
deba@1114
     5
\page graph-io-page Graph Input-Output
deba@1114
     6
athos@1540
     7
The standard graph IO enables one to store graphs and additional maps
athos@1540
     8
(i.e. functions on the nodes or edges) in a flexible and efficient way. 
athos@1540
     9
Before you read this page you should be familiar with LEMON 
athos@1540
    10
\ref graphs "graphs" and \ref maps-page "maps".
deba@1114
    11
deba@1114
    12
\section format The general file format
deba@1114
    13
deba@1532
    14
The file contains sections in the following order:
deba@1114
    15
deba@1114
    16
\li nodeset
deba@1114
    17
\li edgeset
deba@1114
    18
\li nodes
deba@1114
    19
\li edges
deba@1532
    20
\li attributes
deba@1114
    21
athos@1540
    22
Some of these sections can be omitted, but you will basicly need the nodeset
athos@1540
    23
section (unless your graph has no nodes at all) and the edgeset section
athos@1540
    24
(unless your graph has no edges at all). 
athos@1540
    25
athos@1540
    26
The nodeset section describes the nodes of your graph: it identifies the nodes
athos@1540
    27
and gives the maps defined on them, if any. It starts with the
athos@1540
    28
following line:
athos@1522
    29
athos@1522
    30
<tt>\@nodeset</tt>
athos@1522
    31
athos@1522
    32
The next line contains the names of the nodemaps, separated by whitespaces.  Each
athos@1522
    33
following line describes a node in the graph: it contains the values of the
deba@1901
    34
maps in the right order. The map named "label" should contain unique values
deba@1901
    35
because it is regarded as a label map. These labels need not be numbers but they
athos@1540
    36
must identify the nodes uniquely for later reference. For example:
deba@1114
    37
deba@1114
    38
\code
deba@1114
    39
@nodeset
deba@1901
    40
label  x-coord  y-coord  color
deba@1114
    41
3   1.0      4.0      blue
deba@1114
    42
5   2.3      5.7      red
deba@1114
    43
12  7.8      2.3      green
deba@1114
    44
\endcode
deba@1114
    45
deba@1114
    46
The edgeset section is very similar to the nodeset section, it has
athos@1522
    47
the same coloumn oriented structure. It starts with the line 
athos@1522
    48
athos@1522
    49
<tt>\@edgeset</tt>
athos@1522
    50
athos@1540
    51
The next line contains the whitespace separated list of names of the edge
athos@1540
    52
maps.  Each of the next lines describes one edge. The first two elements in
deba@1901
    53
the line are the labels of the source and target (or tail and head) nodes of the
deba@1901
    54
edge as they occur in the label node map of the nodeset section. You can also
deba@1901
    55
have an optional label map on the edges for later reference (which has to be
athos@1540
    56
unique in this case).
deba@1114
    57
deba@1114
    58
\code
deba@1114
    59
@edgeset
deba@1901
    60
             label      weight   note
deba@1901
    61
3   5        a          4.3      a-edge
deba@1901
    62
5   12       c          2.6      c-edge
deba@1901
    63
3   12       g          3.4      g-edge
deba@1114
    64
\endcode
deba@1114
    65
athos@1540
    66
The \e nodes section contains <em>labeled (distinguished) nodes</em> 
athos@1540
    67
(i.e. nodes having a special
alpar@1118
    68
label on them). The section starts with
athos@1522
    69
athos@1522
    70
<tt> \@nodes </tt>
athos@1522
    71
athos@1522
    72
Each of the next lines contains a label for a node in the graph 
deba@1901
    73
and then the label as described in the \e nodeset section.
deba@1114
    74
deba@1114
    75
\code
deba@1114
    76
@nodes 
deba@1114
    77
source 3
deba@1114
    78
target 12
deba@1114
    79
\endcode
deba@1114
    80
athos@1540
    81
The last section describes the <em>labeled (distinguished) edges</em>
deba@1333
    82
(i.e. edges having a special label on them). It starts with \c \@edges
deba@1901
    83
and then each line contains the name of the edge and the label.
deba@1114
    84
deba@1114
    85
\code
athos@1540
    86
@edges 
deba@1114
    87
observed c
deba@1114
    88
\endcode
deba@1114
    89
deba@1114
    90
deba@1114
    91
The file may contain empty lines and comment lines. The comment lines
deba@1114
    92
start with an \c # character.
deba@1114
    93
deba@1532
    94
The attributes section can handle some information about the graph. It
athos@1540
    95
contains key-value pairs in each line (a key and the mapped value to key). The
athos@1540
    96
key should be a string without whitespaces, the value can be of various types.
deba@1532
    97
deba@1532
    98
\code
deba@1532
    99
@attributes
alpar@1959
   100
title "Four colored planar graph"
deba@1532
   101
author "Balazs DEZSO"
deba@1532
   102
copyright "Lemon Library"
deba@1532
   103
version 12
deba@1532
   104
\endcode
deba@1532
   105
deba@1901
   106
Finally, the file should be closed with \c \@end line.
athos@1522
   107
deba@1114
   108
deba@1114
   109
\section use Using graph input-output
athos@1540
   110
athos@1540
   111
athos@1540
   112
The graph input and output is based on <em> reading and writing
athos@1540
   113
commands</em>. The user gives reading and writing commands to the reader or
athos@1540
   114
writer class, then he calls the \c run() method that executes all the given
athos@1540
   115
commands.
deba@1114
   116
deba@1114
   117
\subsection write Writing a graph
deba@1114
   118
alpar@1631
   119
The \ref lemon::GraphWriter "GraphWriter" template class
alpar@1631
   120
provides the graph output. To write a graph
athos@1526
   121
you should first give writing commands to the writer. You can declare
athos@1540
   122
writing command as \c NodeMap or \c EdgeMap writing and labeled Node and
deba@1114
   123
Edge writing.
deba@1114
   124
deba@1114
   125
\code
deba@1333
   126
GraphWriter<ListGraph> writer(std::cout, graph);
deba@1114
   127
\endcode
deba@1114
   128
alpar@1631
   129
The \ref lemon::GraphWriter::writeNodeMap() "writeNodeMap()"
alpar@1631
   130
function declares a \c NodeMap writing command in the
alpar@1631
   131
\ref lemon::GraphWriter "GraphWriter".
alpar@1631
   132
You should give a name to the map and the map
deba@1901
   133
object as parameters. The NodeMap writing command with name "label" should write a 
deba@1901
   134
unique map because it will be regarded as a label map.
deba@1114
   135
deba@1114
   136
\see IdMap, DescriptorMap  
deba@1114
   137
deba@1114
   138
\code
deba@1901
   139
IdMap<ListGraph, Node> nodeLabelMap;
deba@1901
   140
writer.writeNodeMap("label", nodeLabelMap);
deba@1114
   141
deba@1394
   142
writer.writeNodeMap("x-coord", xCoordMap);
deba@1394
   143
writer.writeNodeMap("y-coord", yCoordMap);
deba@1394
   144
writer.writeNodeMap("color", colorMap);
deba@1114
   145
\endcode
deba@1114
   146
alpar@1631
   147
With the \ref lemon::GraphWriter::writeEdgeMap() "writeEdgeMap()"
alpar@1631
   148
member function you can give an edge map
deba@1333
   149
writing command similar to the NodeMaps.
deba@1114
   150
deba@1114
   151
\see IdMap, DescriptorMap  
athos@1522
   152
deba@1114
   153
\code
deba@1114
   154
DescriptorMap<ListGraph, Edge, ListGraph::EdgeMap<int> > edgeDescMap(graph);
deba@1394
   155
writer.writeEdgeMap("descriptor", edgeDescMap);
deba@1114
   156
deba@1394
   157
writer.writeEdgeMap("weight", weightMap);
deba@1901
   158
writer.writeEdgeMap("note", noteMap);
deba@1114
   159
\endcode
deba@1114
   160
alpar@1631
   161
With \ref lemon::GraphWriter::writeNode() "writeNode()"
alpar@1631
   162
and \ref lemon::GraphWriter::writeEdge() "writeEdge()"
alpar@1631
   163
functions you can designate Nodes and
athos@1522
   164
Edges in the graph. For example, you can write out the source and target node
athos@1522
   165
of a maximum flow instance.
deba@1114
   166
deba@1114
   167
\code
deba@1394
   168
writer.writeNode("source", sourceNode);
deba@1394
   169
writer.writeNode("target", targetNode);
deba@1114
   170
deba@1394
   171
writer.writeEdge("observed", edge);
deba@1114
   172
\endcode
deba@1114
   173
alpar@1631
   174
With \ref lemon::GraphWriter::writeAttribute() "writeAttribute()"
alpar@1631
   175
function you can write an attribute to the file.
deba@1532
   176
deba@1532
   177
\code
deba@1532
   178
writer.writeAttribute("author", "Balazs DEZSO");
deba@1532
   179
writer.writeAttribute("version", 12);
deba@1532
   180
\endcode
deba@1532
   181
alpar@1631
   182
After you give all write commands you must call the
alpar@1631
   183
\ref lemon::GraphWriter::run() "run()" member
athos@1522
   184
function, which executes all the writing commands.
deba@1114
   185
deba@1114
   186
\code
deba@1114
   187
writer.run();
deba@1114
   188
\endcode
deba@1114
   189
deba@1114
   190
\subsection reading Reading a graph
deba@1114
   191
athos@1540
   192
The file to be read may contain several maps and labeled nodes or edges.
deba@1114
   193
If you read a graph you need not read all the maps and items just those
alpar@1631
   194
that you need. The interface of the \ref lemon::GraphReader "GraphReader"
alpar@1631
   195
is very similar to
alpar@1631
   196
the \ref lemon::GraphWriter "GraphWriter"
alpar@1631
   197
but the reading method does not depend on the order of the
deba@1114
   198
given commands.
deba@1114
   199
deba@2100
   200
The reader object assumes that each not read value does not contain 
alpar@1118
   201
whitespaces, therefore it has some extra possibilities to control how
alpar@1118
   202
it should skip the values when the string representation contains spaces.
deba@1114
   203
deba@1114
   204
\code
deba@1333
   205
GraphReader<ListGraph> reader(std::cin, graph);
deba@1114
   206
\endcode
deba@1114
   207
alpar@1631
   208
The \ref lemon::GraphReader::readNodeMap() "readNodeMap()"
alpar@1631
   209
function reads a map from the \c nodeset section.
athos@1522
   210
If there is a map that you do not want to read from the file and there are
athos@1522
   211
whitespaces in the string represenation of the values then you should
alpar@1631
   212
call the \ref lemon::GraphReader::skipNodeMap() "skipNodeMap()"
alpar@1631
   213
template member function with proper parameters.
deba@1114
   214
deba@1114
   215
\see QuotedStringReader
athos@1522
   216
deba@1114
   217
\code
deba@1394
   218
reader.readNodeMap("x-coord", xCoordMap);
deba@1394
   219
reader.readNodeMap("y-coord", yCoordMap);
deba@1114
   220
deba@1394
   221
reader.readNodeMap<QuotedStringReader>("label", labelMap);
deba@1114
   222
reader.skipNodeMap<QuotedStringReader>("description");
deba@1114
   223
deba@1394
   224
reader.readNodeMap("color", colorMap);
deba@1114
   225
\endcode
deba@1114
   226
alpar@1631
   227
With the \ref lemon::GraphReader::readEdgeMap() "readEdgeMap()"
alpar@1631
   228
member function you can give an edge map
deba@1114
   229
reading command similar to the NodeMaps. 
deba@1114
   230
deba@1114
   231
\code
deba@1394
   232
reader.readEdgeMap("weight", weightMap);
deba@1394
   233
reader.readEdgeMap("label", labelMap);
deba@1114
   234
\endcode
deba@1114
   235
alpar@1631
   236
With \ref lemon::GraphReader::readNode() "readNode()"
alpar@1631
   237
and \ref lemon::GraphReader::readEdge() "readEdge()"
alpar@1631
   238
functions you can read labeled Nodes and
deba@1114
   239
Edges.
deba@1114
   240
deba@1114
   241
\code
deba@1394
   242
reader.readNode("source", sourceNode);
deba@1394
   243
reader.readNode("target", targetNode);
deba@1114
   244
deba@1394
   245
reader.readEdge("observed", edge);
deba@1114
   246
\endcode
deba@1114
   247
alpar@1631
   248
With \ref lemon::GraphReader::readAttribute() "readAttribute()"
alpar@1631
   249
function you can read an attribute from the file.
deba@1532
   250
deba@1532
   251
\code
deba@1532
   252
std::string author;
deba@1532
   253
writer.readAttribute("author", author);
deba@1532
   254
int version;
deba@1532
   255
writer.writeAttribute("version", version);
deba@1532
   256
\endcode
deba@1532
   257
alpar@1631
   258
After you give all read commands you must call the
alpar@1631
   259
\ref lemon::GraphReader::run() "run()" member
athos@1522
   260
function, which executes all the commands.
deba@1114
   261
deba@1114
   262
\code
deba@1114
   263
reader.run();
deba@1114
   264
\endcode
deba@1114
   265
athos@1540
   266
\anchor rwbackground
athos@1527
   267
\section types Background of Reading and Writing
athos@1540
   268
athos@1540
   269
athos@1527
   270
To read a map (on the nodes or edges)
alpar@1631
   271
the \ref lemon::GraphReader "GraphReader"
alpar@1631
   272
should know how to read a Value from the given map.
deba@1114
   273
By the default implementation the input operator reads a value from
deba@2100
   274
the stream and the type of the read value is the value type of the given map.
deba@1114
   275
When the reader should skip a value in the stream, because you do not
athos@1527
   276
want to store it in a map, the reader skips a character sequence without 
athos@1540
   277
whitespaces. 
deba@1114
   278
deba@1114
   279
If you want to change the functionality of the reader, you can use
deba@1114
   280
template parameters to specialize it. When you give a reading
deba@1114
   281
command for a map you can give a Reader type as template parameter.
deba@1333
   282
With this template parameter you can control how the Reader reads
deba@1114
   283
a value from the stream.
deba@1114
   284
deba@1114
   285
The reader has the next structure: 
deba@1114
   286
\code
deba@1114
   287
struct TypeReader {
deba@1114
   288
  typedef TypeName Value;
deba@1114
   289
deba@1114
   290
  void read(std::istream& is, Value& value);
deba@1114
   291
};
deba@1114
   292
\endcode
deba@1114
   293
athos@1527
   294
For example, the \c "strings" nodemap contains strings and you do not need
athos@1540
   295
the value of the string just the length. Then you can implement an own Reader
deba@1114
   296
struct.
deba@1114
   297
deba@1114
   298
\code
deba@1114
   299
struct LengthReader {
deba@1114
   300
  typedef int Value;
deba@1114
   301
deba@1114
   302
  void read(std::istream& is, Value& value) {
deba@1114
   303
    std::string tmp;
deba@1114
   304
    is >> tmp;
deba@1114
   305
    value = tmp.length();
deba@1114
   306
  }
deba@1114
   307
};
deba@1114
   308
...
deba@1394
   309
reader.readNodeMap<LengthReader>("strings", lengthMap);
deba@1114
   310
\endcode  
deba@1114
   311
deba@1114
   312
The global functionality of the reader class can be changed by giving a
athos@1526
   313
special template parameter to the GraphReader class. By default, the
alpar@1118
   314
template parameter is \c DefaultReaderTraits. A reader traits class 
deba@1901
   315
should provide a nested template class Reader for each type, and a 
deba@1114
   316
DefaultReader for skipping a value.
deba@1114
   317
deba@1901
   318
The specialization of writing is very similar to that of reading.
deba@1114
   319
klao@1909
   320
\section u Undirected graphs
deba@1532
   321
klao@1909
   322
In a file describing an undirected graph (ugraph, for short) you find an
klao@1909
   323
\c uedgeset section instead of the \c edgeset section. The first line of
athos@1540
   324
the section describes the names of the maps on the undirected egdes and all
athos@1540
   325
next lines describe one undirected edge with the the incident nodes and the
athos@1540
   326
values of the map.
deba@1532
   327
athos@1540
   328
The format handles directed edge maps as a syntactical sugar???, if there
athos@1540
   329
are two maps with names being the same with a \c '+' and a \c '-' prefix
athos@1540
   330
then this will be read as a directed map.
deba@1532
   331
deba@1532
   332
\code
klao@1909
   333
@uedgeset
deba@1901
   334
             label      capacity        +flow   -flow
deba@1901
   335
32   2       1          4.3             2.0     0.0
deba@1901
   336
21   21      5          2.6             0.0     2.6
deba@1901
   337
21   12      8          3.4             0.0     0.0
deba@1532
   338
\endcode
deba@1532
   339
klao@1909
   340
The \c edges section is changed to \c uedges section. This section
deba@1532
   341
describes labeled edges and undirected edges. The directed edge label
athos@1540
   342
should start with a \c '+' or a \c '-' prefix to decide the direction
deba@1532
   343
of the edge. 
deba@1532
   344
deba@1532
   345
\code
klao@1909
   346
@uedges
klao@1909
   347
uedge 1
deba@1532
   348
+edge 5
deba@1532
   349
-back 5
deba@1532
   350
\endcode
deba@1532
   351
alpar@1631
   352
There are similar classes to the \ref lemon::GraphReader "GraphReader" and
alpar@1631
   353
\ref lemon::GraphWriter "GraphWriter" which
alpar@1631
   354
handle the undirected graphs. These classes are
klao@1909
   355
the \ref lemon::UGraphReader "UGraphReader"
klao@1909
   356
and \ref lemon::UGraphWriter "UGraphWriter".
deba@1532
   357
klao@1909
   358
The \ref lemon::UGraphReader::readUEdgeMap() "readUEdgeMap()"
alpar@1631
   359
function reads an undirected map and the
klao@1909
   360
\ref lemon::UGraphReader::readUEdge() "readUEdge()"
alpar@1631
   361
reads an undirected edge from the file, 
deba@1532
   362
deba@1532
   363
\code
klao@1909
   364
reader.readUEdgeMap("capacity", capacityMap);
deba@1532
   365
reader.readEdgeMap("flow", flowMap);
deba@1532
   366
...
klao@1909
   367
reader.readUEdge("u_edge", u_edge);
deba@1532
   368
reader.readEdge("edge", edge);
deba@1532
   369
\endcode
deba@1532
   370
deba@1532
   371
\section advanced Advanced features
deba@1532
   372
athos@1540
   373
The graph reader and writer classes give an easy way to read and write
athos@1540
   374
graphs. But sometimes we want more advanced features. In this case we can
athos@1540
   375
use the more general <tt>lemon reader and writer</tt> interface.
deba@1532
   376
athos@1540
   377
The LEMON file format is a section oriented file format. It contains one or
athos@1540
   378
more sections, each starting with a line identifying its type 
athos@1540
   379
(the word starting with the \c \@  character).
deba@1532
   380
The content of the section this way cannot contain line with \c \@ first
deba@1532
   381
character. The file may contains comment lines with \c # first character.
deba@1532
   382
alpar@1631
   383
The \ref lemon::LemonReader "LemonReader"
alpar@1631
   384
and \ref lemon::LemonWriter "LemonWriter"
alpar@1631
   385
gives a framework to read and
deba@1532
   386
write sections. There are various section reader and section writer
alpar@1631
   387
classes which can be attached to a \ref lemon::LemonReader "LemonReader"
alpar@1631
   388
or a \ref lemon::LemonWriter "LemonWriter".
deba@1532
   389
deba@1532
   390
There are default section readers and writers for reading and writing
athos@1540
   391
item sets, and labeled items in the graph. These read and write
deba@1532
   392
the format described above. Other type of data can be handled with own
deba@1532
   393
section reader and writer classes which are inherited from the
alpar@1631
   394
\c LemonReader::SectionReader or the
alpar@1631
   395
\ref lemon::LemonWriter::SectionWriter "LemonWriter::SectionWriter"
alpar@1631
   396
classes.
deba@1532
   397
deba@1532
   398
The next example defines a special section reader which reads the
deba@1532
   399
\c \@description sections into a string:
deba@1532
   400
deba@1532
   401
\code 
deba@1532
   402
class DescriptionReader : LemonReader::SectionReader {
deba@1532
   403
protected:
deba@1532
   404
  virtual bool header(const std::string& line) {
deba@1532
   405
    std::istringstream ls(line);
deba@1532
   406
    std::string head;
deba@1532
   407
    ls >> head;
deba@1532
   408
    return head == "@description";
deba@1532
   409
  }
deba@1532
   410
deba@1532
   411
  virtual void read(std::istream& is) {
deba@1532
   412
    std::string line;
deba@1532
   413
    while (getline(is, line)) {
deba@1532
   414
      desc += line;
deba@1532
   415
    }
deba@1532
   416
  }
deba@1532
   417
public:
deba@1532
   418
deba@1532
   419
  typedef LemonReader::SectionReader Parent;
deba@1532
   420
  
deba@1532
   421
  DescriptionReader(LemonReader& reader) : Parent(reader) {}
deba@1532
   422
deba@1532
   423
  const std::string& description() const {
deba@1532
   424
    return description;
deba@1532
   425
  }
deba@1532
   426
deba@1532
   427
private:
deba@1532
   428
  std::string desc;
deba@1532
   429
};
deba@1532
   430
\endcode
deba@1532
   431
deba@1532
   432
The other advanced stuff of the generalized file format is that 
deba@1532
   433
multiple edgesets can be stored to the same nodeset. It can be used 
athos@1540
   434
for example as a network traffic matrix.
deba@1532
   435
athos@1540
   436
In our example there is a network with symmetric links and there are assymetric
deba@1532
   437
traffic request on the network. This construction can be stored in an
deba@1842
   438
undirected graph and in a directed \c ListEdgeSet class. The example
alpar@1631
   439
shows the input with the \ref lemon::LemonReader "LemonReader" class:
deba@1532
   440
deba@1532
   441
\code
klao@1909
   442
ListUGraph network;
klao@1909
   443
ListUGraph::UEdgeMap<double> capacity;
klao@1909
   444
ListEdgeSet<ListUGraph> traffic(network);
klao@1909
   445
ListEdgeSet<ListUGraph>::EdgeMap<double> request(network);
deba@1532
   446
deba@1532
   447
LemonReader reader(std::cin);
klao@1909
   448
NodeSetReader<ListUGraph> nodesetReader(reader, network);
klao@1909
   449
UEdgeSetReader<ListUGraph> 
klao@1909
   450
  uEdgesetReader(reader, network, nodesetReader);
klao@1909
   451
uEdgesetReader.readEdgeMap("capacity", capacity);
klao@1909
   452
EdgeSetReader<ListEdgeSet<ListUGraph> > 
deba@1848
   453
  edgesetReader(reader, traffic, nodesetReader, "traffic");
deba@1532
   454
edgesetReader.readEdgeMap("request", request);
deba@1532
   455
deba@1532
   456
reader.run();
deba@1532
   457
\endcode
deba@1532
   458
alpar@1631
   459
Because both the \ref lemon::GraphReader "GraphReader"
klao@1909
   460
and the \ref lemon::UGraphReader "UGraphReader" can be converted
alpar@1631
   461
to \ref lemon::LemonReader "LemonReader"
deba@1901
   462
and it can resolve the label's of the items, the previous
klao@1909
   463
result can be achived with the \ref lemon::UGraphReader "UGraphReader"
alpar@1631
   464
class, too.
deba@1532
   465
deba@1532
   466
deba@1532
   467
\code
klao@1909
   468
ListUGraph network;
klao@1909
   469
ListUGraph::UEdgeSet<double> capacity;
klao@1909
   470
ListEdgeSet<ListUGraph> traffic(network);
klao@1909
   471
ListEdgeSet<ListUGraph>::EdgeMap<double> request(network);
deba@1532
   472
klao@1909
   473
UGraphReader<ListUGraph> reader(std::cin, network);
deba@1532
   474
reader.readEdgeMap("capacity", capacity);
klao@1909
   475
EdgeSetReader<ListEdgeSet<ListUGraph> > 
deba@1848
   476
  edgesetReader(reader, traffic, reader, "traffic");
deba@1532
   477
edgesetReader.readEdgeMap("request", request);
deba@1532
   478
deba@1532
   479
reader.run();
deba@1532
   480
\endcode
deba@1532
   481
deba@1333
   482
\author Balazs Dezso
deba@1114
   483
*/
alpar@1631
   484
}