| graph | wolffd@0:: | wolffd@0:[ strict ] (graph | digraph) [ ID ] '{' stmt_list '}' | wolffd@0:
| stmt_list | wolffd@0:: | wolffd@0:[ stmt [ ';' ] [ stmt_list ] ] | wolffd@0:
| stmt | wolffd@0:: | wolffd@0:node_stmt | wolffd@0:
| wolffd@0: | | | wolffd@0:edge_stmt | wolffd@0:
| wolffd@0: | | | wolffd@0:attr_stmt | wolffd@0:
| wolffd@0: | | | wolffd@0:ID '=' ID | wolffd@0:
| wolffd@0: | | | wolffd@0:subgraph | wolffd@0:
| attr_stmt | wolffd@0:: | wolffd@0:(graph | node | edge) attr_list | wolffd@0:
| attr_list | wolffd@0:: | wolffd@0:'[' [ a_list ] ']' [ attr_list ] | wolffd@0:
| a_list | wolffd@0:: | wolffd@0:ID [ '=' ID ] [ ',' ] [ a_list ] | wolffd@0:
| edge_stmt | wolffd@0:: | wolffd@0:(node_id | subgraph) edgeRHS [ attr_list ] | wolffd@0:
| edgeRHS | wolffd@0:: | wolffd@0:edgeop (node_id | subgraph) [ edgeRHS ] | wolffd@0:
| node_stmt | wolffd@0:: | wolffd@0:node_id [ attr_list ] | wolffd@0:
| node_id | wolffd@0:: | wolffd@0:ID [ port ] | wolffd@0:
| port | wolffd@0:: | wolffd@0:':' ID [ ':' compass_pt ] | wolffd@0:
| wolffd@0: | | | wolffd@0:':' compass_pt | wolffd@0:
| subgraph | wolffd@0:: | wolffd@0:[ subgraph [ ID ] ] '{' stmt_list '}' | wolffd@0:
| compass_pt | wolffd@0:: | wolffd@0:(n | ne | e | se | s | sw | w | nw | c | _) | wolffd@0:
wolffd@0: The keywords node, edge, graph, digraph, wolffd@0: subgraph, and strict are case-independent. wolffd@0: Note also that the allowed compass point values are not keywords, so wolffd@0: these strings can be used elsewhere as ordinary identifiers and, conversely, wolffd@0: the parser will actually accept any identifier. wolffd@0:
wolffd@0: An ID is one of the following: wolffd@0:
wolffd@0: Note that in HTML strings, angle wolffd@0: brackets must occur in matched pairs, and unescaped newlines are allowed. wolffd@0: In addition, the content must be legal XML, so that the special XML wolffd@0: escape sequences for ", &, <, and > may be necessary wolffd@0: in order to embed these characters in attribute values or raw text. wolffd@0:wolffd@0: Both quoted strings and HTML strings are scanned as a unit, so wolffd@0: any embedded comments will be treated as part of the strings. wolffd@0:
wolffd@0: An edgeop is -> in directed graphs and -- in wolffd@0: undirected graphs. wolffd@0:
wolffd@0: An a_list clause of the form ID is equivalent to wolffd@0: ID=true. wolffd@0:
wolffd@0: The language supports C++-style comments: /* */ and //. wolffd@0: In addition, a line beginning with a '#' character is considered a line wolffd@0: output from a C preprocessor (e.g., # 34 to indicate line 34 ) and discarded. wolffd@0:
wolffd@0: Semicolons aid readability but are not required except in the rare case wolffd@0: that a named subgraph with no body immediately preceeds an anonymous wolffd@0: subgraph, since the precedence rules cause this sequence to be parsed as wolffd@0: a subgraph with a heading and a body. wolffd@0: Also, any amount of whitespace may be inserted between terminals. wolffd@0:
wolffd@0: As another aid for readability, dot allows single logical lines to wolffd@0: span multiple physical lines using the standard C convention of a wolffd@0: backslash immediately preceding a newline character. In addition, wolffd@0: double-quoted strings can be concatenated using a '+' operator. wolffd@0: As HTML strings can contain newline characters, they do not support the wolffd@0: concatenation operator. wolffd@0:
wolffd@0: Note, in particular, that a subgraph receives the attribute settings of wolffd@0: its parent graph at the time of its definition. This can be useful; for wolffd@0: example, one can assign a font to the root graph and all subgraphs will wolffd@0: also use the font. For some attributes, however, this property is wolffd@0: undesirable. If one attaches a label to the root graph, it is probably wolffd@0: not the desired effect to have the label used by all subgraphs. Rather wolffd@0: than listing the graph attribute at the top of the graph, and the wolffd@0: resetting the attribute as needed in the subgraphs, one can simple defer wolffd@0: the attribute definition if the graph until the appropriate subgraphs wolffd@0: have been defined. wolffd@0:
wolffd@0: If an edge belongs to a cluster, its endpoints belong to that cluster. wolffd@0: Thus, where you put an edge can effect a layout, as clusters are sometimes wolffd@0: laid out recursively. wolffd@0:
wolffd@0: By default, DOT assumes the UTF-8 character encoding. It also accepts wolffd@0: the Latin1 (ISO-8859-1) character set, assuming the input graph uses wolffd@0: the charset attribute to wolffd@0: specify this. For graphs using other wolffd@0: character sets, there are usually programs, such as iconv, which wolffd@0: will translate from one character set to another. wolffd@0:
wolffd@0: Another way to avoid non-ascii characters in labels is to use HTML entities wolffd@0: for special characters. During label evaluation, these entities are wolffd@0: translated into the underlying character. This wolffd@0: wolffd@0: table shows the supported entities, with their Unicode value, a typical wolffd@0: glyph, and the HTML entity name. Thus, to include a lower-case Greek beta wolffd@0: into a string, one can use the ascii sequence β. wolffd@0: In general, one should only use entities that are allowed in the output wolffd@0: character set, and for which there is a glyph in the font. wolffd@0: