| graph | : | [ strict ] (graph | digraph) [ ID ] '{' stmt_list '}' |
| stmt_list | : | [ stmt [ ';' ] [ stmt_list ] ] |
| stmt | : | node_stmt |
| | | edge_stmt | |
| | | attr_stmt | |
| | | ID '=' ID | |
| | | subgraph | |
| attr_stmt | : | (graph | node | edge) attr_list |
| attr_list | : | '[' [ a_list ] ']' [ attr_list ] |
| a_list | : | ID '=' ID [ (';' | ',') ] [ a_list ] |
| edge_stmt | : | (node_id | subgraph) edgeRHS [ attr_list ] |
| edgeRHS | : | edgeop (node_id | subgraph) [ edgeRHS ] |
| node_stmt | : | node_id [ attr_list ] |
| node_id | : | ID [ port ] |
| port | : | ':' ID [ ':' compass_pt ] |
| | | ':' compass_pt | |
| subgraph | : | [ subgraph [ ID ] ] '{' stmt_list '}' |
| compass_pt | : | (n | ne | e | se | s | sw | w | nw | c | _) |
The keywords node, edge, graph, digraph, subgraph, and strict are case-independent. Note also that the allowed compass point values are not keywords, so these strings can be used elsewhere as ordinary identifiers and, conversely, the parser will actually accept any identifier.
An ID is one of the following:
An ID is just a string; the lack of quote characters in the first two forms is just for simplicity. There is no semantic difference between abc_2 and "abc_2", or between 2.34 and "2.34". Obviously, to use a keyword as an ID, it must be quoted. Note that, in HTML strings, angle brackets must occur in matched pairs, and newlines and other formatting whitespace characters are allowed. In addition, the content must be legal XML, so that the special XML escape sequences for ", &, <, and > may be necessary in order to embed these characters in attribute values or raw text.Both quoted strings and HTML strings are scanned as a unit, so any embedded comments will be treated as part of the strings.
An edgeop is -> in directed graphs and -- in undirected graphs.
The language supports C++-style comments: /* */ and //. In addition, a line beginning with a '#' character is considered a line output from a C preprocessor (e.g., # 34 to indicate line 34 ) and discarded.
Semicolons aid readability but are not required except in the rare case that a named subgraph with no body immediately preceeds an anonymous subgraph, since the precedence rules cause this sequence to be parsed as a subgraph with a heading and a body. Also, any amount of whitespace may be inserted between terminals.
As another aid for readability, dot allows double-quoted strings to span multiple physical lines using the standard C convention of a backslash immediately preceding a newline character2. In addition, double-quoted strings can be concatenated using a '+' operator. As HTML strings can contain newline characters, which are used solely for formatting, the language does not allow escaped newlines or concatenation operators to be used within them.
In the second role, a subgraph can provide a context for setting attributes. For example, a subgraph could specify that blue is the default color for all nodes defined in it. In the context of graph drawing, a more interesting example is:
subgraph {
rank = same; A; B; C;
}
This (anonymous) subgraph specifies that the nodes A, B and C
should all be placed on the same rank if drawn using dot.
The third role for subgraphs directly involves how the graph will be laid out by certain layout engines. If the name of the subgraph begins with cluster, Graphviz notes the subgraph as a special cluster subgraph. If supported, the layout engine will do the layout so that the nodes belonging to the cluster are drawn together, with the entire drawing of the cluster contained within a bounding rectangle. Note that, for good and bad, cluster subgraphs are not part of the DOT language, but solely a syntactic convention adhered to by certain of the layout engines.
Note, in particular, that a subgraph receives the attribute settings of its parent graph at the time of its definition. This can be useful; for example, one can assign a font to the root graph and all subgraphs will also use the font. For some attributes, however, this property is undesirable. If one attaches a label to the root graph, it is probably not the desired effect to have the label used by all subgraphs. Rather than listing the graph attribute at the top of the graph, and the resetting the attribute as needed in the subgraphs, one can simply defer the attribute definition in the graph until the appropriate subgraphs have been defined.
If an edge belongs to a cluster, its endpoints belong to that cluster. Thus, where you put an edge can effect a layout, as clusters are sometimes laid out recursively.
There are certain restrictions on subgraphs and clusters. First, at present, the names of a graph and it subgraphs share the same namespace. Thus, each subgraph must have a unique name. Second, although nodes can belong to any number of subgraphs, it is assumed clusters form a strict hierarchy when viewed as subsets of nodes and edges.
By default, DOT assumes the UTF-8 character encoding. It also accepts the Latin1 (ISO-8859-1) character set, assuming the input graph uses the charset attribute to specify this. For graphs using other character sets, there are usually programs, such as iconv, which will translate from one character set to another.
Another way to avoid non-ascii characters in labels is to use HTML entities for special characters. During label evaluation, these entities are translated into the underlying character. This table shows the supported entities, with their Unicode value, a typical glyph, and the HTML entity name. Thus, to include a lower-case Greek beta into a string, one can use the ascii sequence β. In general, one should only use entities that are allowed in the output character set, and for which there is a glyph in the font.