Reference guide¶

GParser¶

class gparser.GParser(validation: bool = True, lib: Optional[str] = None, patternmatcher: Optional[gpatternmatcher.GPatternMatcher] = None, graphbuilder: Optional[ggraphbuilder.GGraphBuilder] = None, graphtype: Literal['GDGraph', 'GAGraph'] = 'GDGraph', databuilder: Optional[gdatabuilder.GDataBuilder] = None, grammar: Optional[Union[lark.load_grammar.Grammar, str, IO[str]]] = None, parser: Literal['earley', 'lalr', 'cyk', 'auto'] = 'lalr', lexer: Union[Type[lark.lexer.Lexer], Literal['auto', 'standard', 'contextual', 'dynamic', 'dynamic_complete']] = 'contextual', start: Optional[Union[str, List[str]]] = 'sequence', cache: Union[bool, str] = '.lark-cache', debug: bool = False)¶

The GParser class, inherited from lark.Lark

Parameters

validation – If True, run indices validation after parsing
lib – The patternlibrary for the Pattern Matching. (see GPatternMatcher.lib). Default: None
patternmatcher (GPatternMatcher) – The Pattern Matching class. If None, using GPatternMatcher. Default: None
graphbuilder (GGraphBuilder) – The graph builder class. If None, using GGraphBuilder. Default: None
databuilder (GDataBuilder) – The data builder class. If None, using GDataBuilder. Default: None
grammar – The grammar file. Default: grammar.lark
parser – The parser algorithm. Default: lalr
lexer – The lexer. Default: contextual
start – The start point to the parsing. Default: sequence
cache – With cache parsing is x2~x3 faster! Default: True
debug – The debug option. Default: False

build_data(data: Union[List, Tuple, gtree.GTree, str], **kwargs) → Union[List, gdata.GData]¶

Builds GData from the data and returns the result

Parameters: data – The data to the graph building. May be GTree, list[GTree] or raw data, from that GTree will be parsed.
Returns: The builded List[GData]

build_graph(data: Union[List, Tuple, gtree.GTree], pattern_matching: bool = False, **kwargs) → Optional[Union[List, Tuple, gagraph.GAGraph, gdgraph.GDGraph, gdgraph.GDGraphHelper]]¶

Builds GAGraph or GDGraph from the data and returns the result

Parameters

data – The data to the graph building. May be GTree, list[GTree] or raw data, from that GTree will be parsed.
pattern_matching (bool) – If True makes Pattern Matching from lib or self.lib (is setted and lib arg is None). Default: False

Returns

The builded GAGraph or GDGraph

convert(data: lark.tree.Tree) → gtree.GTree¶
convert(data: lark.lexer.Token) → gtree.GToken
convert(data: Any) → Any: Convers data to the GTree or GToken

property databuilder: gdatabuilder.GDataBuilder¶: The self._databuilder getter

property graphbuilder: ggraphbuilder.GGraphBuilder¶: The self._graphbuilder getter

match(data: Union[List, Tuple, gdgraph.GDGraph, gdgraph.GDGraphHelper], lib: Optional[str] = None) → Union[List, Tuple, gdgraph.GDGraph, gdgraph.GDGraphHelper]¶

Makes Pattern Matching for the data and returns the result

Parameters

data – The data to Pattern Matching. May be GTree, list[GTree] or raw data, from that GTree will be parsed.
lib – The pattern patternlibrary. If None using from GParser.__init__. Default: None

Returns

The Pattern Matched GTree

parse(text: str, start: Optional[str] = None, on_error: Optional[Callable] = None, make_identifiers: bool = True, reset_counts: bool = True) → Union[gtree.GTree, List, Tuple]¶

Reimplemented lark.tree.Tree.parse method

Parameters

text (str) – Text to be parsed
start (str) – Required if GParser was given multiple possible start symbols (using the start option). Default: None
on_error (function) – If provided, will be called on UnexpectedToken error. Return True to resume parsing. Default: None
make_identifiers (bool) – If True makes unique identifiers for all the parts of this GTree (see GTree.make_identifiers). Default: True
reset_counts (bool) – If True resets the counts for the making identifiers (see GTree.reset_counts). Default: True

Returns

Returns parsed GTree

pattern_matching(data: Union[List, Tuple, gdgraph.GDGraph, gdgraph.GDGraphHelper], lib: Optional[str] = None) → Union[List, Tuple, gdgraph.GDGraph, gdgraph.GDGraphHelper]¶

Makes Pattern Matching for the data and returns the result

Parameters

data – The data to Pattern Matching. May be GTree, list[GTree] or raw data, from that GTree will be parsed.
lib – The pattern patternlibrary. If None using from GParser.__init__. Default: None

Returns

The Pattern Matched GTree

property patternmatcher: gpatternmatcher.GPatternMatcher¶: The self._patternmatcher getter

tree_to_file(data: Union[gtree.GTree, lark.tree.Tree], filename: str, rankdir: Literal['TB', 'LR', 'BT', 'RL'] = 'LR', **kwargs) → None¶

Saves the data to filename

Parameters

data – GTree or lark.tree.Tree to saving
filename (str) – The filename to saving
rankdir (str) – Possible values are “TB”, “LR” (default), “BT”, “RL”, corresponding to directed graphs drawn from top to bottom, from left to right, from bottom to top, and from right to left, respectively.
kwargs – kwargs can be any graph attribute (e.g. dpi=200). For a list of possible attributes, see https://www.graphviz.org/doc/info/attrs.html.

GPatternMatcher¶

class gpatternmatcher.GPatternMatcher(parsemethod: Optional[Callable] = None, datamethod: Optional[Callable] = None, lib: Optional[str] = None, debug: bool = False, **kwargs)¶

The GPatternMatcher class

Designed for the Pattern Matching, in other words, to search the library for special rules for replacing any constructs in the tree, thereby simplifying the tree.

Use match-method for the Pattern Matching.

Parameters

parser – The parser class, that provides parse-method
lib – The library with patterns

Note

Now supports only YAML libraries! Example:

patternname:
    expr: "DSL @ Expression * to + Pattern - Matching"
    label: "label"

build_pgraphs() → list¶: Builds the list of the patterns GDGraph`s `GPatternMatcher.pgraphs

property lib: Optional[str]¶: The library with patterns

match(graph: Union[gdgraph.GDGraph, gdgraph.GDGraphHelper]) → Union[gdgraph.GDGraph, gdgraph.GDGraphHelper]¶

Find patterns in the lib and match corresponding object in the tree

Parameters: tree – The graph to pattern matching

read_patterns(lib: Optional[str] = None) → dict¶

Reads the patterns from lib and returns as dict

Parameters: lib – The library with patterns

Note

Now supports only YAML libraries!

GGraphBuilder¶

class ggraphbuilder.GGraphBuilder(databuilder: Optional[gdatabuilder.GDataBuilder] = None, graphtype: Literal['GDGraph', 'GAGraph'] = 'GDGraph', debug: bool = False)¶

The GGraphBuilder class

Designed for the building of GDGraphHelper or GAGraph from GTree, GData or Sequence[GData]

Parameters

databuilder – The GData builder class. If None, use a new instance of GDataBuilder. Default: None
graphtype – The type of the graph: GDGraph or GAGraph
debug – The debug option

abuild(data: Union[gtree.GTree, gdata.GData, Sequence], orig: Optional[gagraph.GAGraph] = None, **kwargs) → gagraph.GAGraph¶

Builds GAGraph from GTree or GData

Parameters

data – The data from which the GAGraph will be collected. May be GTree, GData or Sequence.
orig – If orig is GAGraph, starting from the graph and modifying it while the building graph from data. Else create a new instance of GAGraph. Default: None

property databuilder: gdatabuilder.GDataBuilder¶: The class to build GData objects

dbuild(_data: Union[gtree.GTree, gdata.GData, Sequence], orig: Optional[gdgraph.GDGraph] = None, **kwargs) → Optional[gdgraph.GDGraph]¶

Builds GDGraph from GTree or GData

Parameters

data – The data from which the GDGraph will be collected. May be GTree, GData or Sequence.
orig – If orig is GDGraph, starting from the graph and modifying it while the building graph from data. Else create a new instance of GDGraph. Default: None

property debug: bool¶: If True do not collapse extra nodes

property graphtype: str¶: If True do not collapse extra nodes

togagraph(gdgraph: Union[gdgraph.GDGraph, gdgraph.GDGraphHelper], **kwargs) → gagraph.GAGraph¶: Converts GDGraph or GDGraphHelper to GAGraph

GDataBuilder¶

class gdatabuilder.GDataBuilder(validation: bool = False, debug: bool = False)¶

The GDataBuilder class

Designed for the building of GData from GTree or list[GTree].

Use build for the building GData from GTree.

Parameters: debug – If True do not collapse extra nodes. Default: False

build(tree: Union[gtree.GTree, List, Tuple], assignments: Optional[Dict[str, gdata.GData]] = None, instances: Optional[Dict[str, int]] = None, **kwargs) → Union[gdata.GData, List[gdata.GData]]¶: Building GData or List[GData] from GTree or List[GTree]

property debug: bool¶: If True do not collapse extra nodes

property validation: bool¶: If True call the validation procedure after a parsing

GTree¶

class gagraph.GAGraph(thing=None, filename=None, data=None, string=None, handle=None, name='', strict=True, directed=False, **attr)¶

Inherited class from pygraphviz.AGraph

Reimplemented to work with GData class

add_edge(u: Union[gdata.GData, Tuple[gdata.GData, gdata.GData], Any], v: Optional[gdata.GData] = None, key: Optional[str] = None, **attr) → pygraphviz.agraph.Edge¶

Adding edge between nodes to the graph

Reimplemented for the work with GData-class

add_node(data: Any = 1, **attr) → pygraphviz.agraph.Node¶

Adding node to the graph

Reimplemented for the work with GData-class

GData¶

class gdata.GData(orig: Optional[Dict] = None)¶

The GData class

The class is like a python dict:

name (str);

sname (str);

type (str);

namespace (str);

label (dict):

default (str);

... (str);

instance (int);

arguments (list);

indices (dict):

explicit (list);

implicit (list);

reduction (list).

sname <-> storage name .uname() <-> unique name

Parameters: orig – Initialization data (GData, dict, str). Default: None

argnames() → list¶: Returns the arguments list without instances, for example, Transf_1 -> Transf, Object_1_2_3 -> Object_1_2

copy() → gdata.GData¶: Returns a copy of the object

classmethod from_str(data: str) → gdata.GData¶: Build the GData object from string

set_sname(val: Optional[str] = None) → None¶: Sets the storage name sname

uname() → str¶: If type is transformation returns name_instance-string, else returns name

GDGraph¶

class gdgraph.GDGraph(name: str, data: Any, children: List[GDGraph] = [])

The light graph class. The main idea is an absence of nodes and self-similarity, i.e. the children (like nodes, but it would be more correct to say the inner layers) are GDGraph`s as the head layer. Any `GDGraph instance contains the name (usual the same as data[“name”]), the data field (see GData class) and children. Supports some kind of caching (see GDMeta class), contains a pretty print method, a dict representation, shape and depth methods. It is possible to get a graph by the name or list of graphs at the some layer via the get method or using brackets []. The class does not contain any graphical representation, but it is possible to convert GDGraph class to GAGraph (via GGraphBuilder.togagraph method), which has the PyGraphviz features (see ggraphbuilder and gagraph modules).

Parameters

name – The name of the graph
data – The data of the graph
children – The list of the children

asdict(include_data: bool = False) → dict

Returns the dict representation of the graph

Parameters: include_data – If True, includes data field, else uses only name and children fields

depth() → int: Returns the depth (maximum layer number) of the graph

classmethod from_gdata(data: gdata.GData, children: Optional[Sequence[gdgraph.GDGraph]] = None) → gdgraph.GDGraph

Build the class from GData

Parameters

data – The GData object
children – The children (inputs, nodes) list

get(key: Union[int, str], default: Optional[Any] = None) → Optional[Union[gdgraph.GDGraph, List, Any]]

Returns the graph or list of graphs by layer number or name

Parameters

key – The layer number or name of the graph
default – If the graph does not contain the desired element(s), returns default value

getbylayer(layer: int) → Union[gdgraph.GDGraph, List]

Returns a graph or list of graphs corresponding to the layer number. It is possible to use negative indexes to get objects from the end

Parameters: layer – The layer number of the graph to find

getbyname(name: str) → gdgraph.GDGraph

Returns a graph corresponding to the name

Parameters: name – The name of the graph to find

isliketo(other: Any, strict_order: bool = True) → bool

Checks the name and arguments of the objects

Parameters

other – The object to compare
strict_order – If False do not check the children (arguments) order

iter_subgraphs(): Depth-first iteration. Iterates over all the subgraphs, never returning to the same subgraph twice.

iter_subgraphs_topdown(): Breadth-first iteration. Iterates over all the subgraphs.

objectscount() → int: Returns all the count of the inner graphs, including itself

print(include_data: bool = False) → None

Pretty print of the graph. The dict representation (see GDGraph.asdict) is used

Parameters: include_data – If True, includes data field, else uses only name and children fields

shape() → tuple: Returns the shape of the graph

todatalist() → list: Returns all the GData objects: the parent data and all the data from the children

tolist() → list: Returns all the GDGraph objects: the parent GDGraph and all the graphs from the children

class gdgraph.GDGraphHelper(graphs: Optional[Union[Sequence[Any], Any]] = None, strict: bool = True, debug: bool = False, **kwargs)

The helper to work with GDGraph objects. Contains the list of graphs and the dict with the same graphs as values and the names as keys.

It is possible to modify graphs by the list or by the dict, but it is important to update the dict after adding new graph to the graph: using the argument bind of the add_graph and add_graphs methods (by default) or by the direct calling the bind_graphs method.

There is no edges, but every graph has children. The method add_graphs, if bind=True, collecting children by the iteration of the data[“arguments”] of the every graph. If a single graph is added using the method add_graph, and it is necessary to iterate through all the graphs and update the children, then a direct call of the bind_children method is necessary.

add_graph(data: Any, name: Optional[str] = None, children: Optional[Sequence[gdgraph.GDGraph]] = None, bind: bool = True) → gdgraph.GDGraph

Adding the graph to the graph

Parameters

data – The data of the graph
name – The name of the graph
children – The list of the children
bind – If True, adding the graph to the dict GDGraphHelper.graphsdict

add_graphs(graphs: Sequence[Any], bind: bool = True) → None

Adding the list of the graphs

Parameters

graphs – The graphs sequence
bind – If True, filling also the dict of the graphs GDGraphHelper.graphsdict and binding children

asdict(include_data: bool = False) → dict

Returns the dict representation of the graph

Parameters: include_data – If True, includes data field, else uses only name and children fields of a graph

bind_children() → None: Collecting and adding children to all the graphs from the graph.data[“arguments”]

bind_graphs() → dict: Filling the dict of the graphs graphsdict

depths() → tuple: Returns the tuple of the depths of the GDGraphHelper.graphs (see GDGraph.depth)

main_graph(validate: bool = True) → Optional[gdgraph.GDGraph]: Returns the deepest graph (see GDGraph.depth)

print(include_data: bool = False) → None

Pretty print of the graph. The dict representation (see GDGraphHelper.asdict) is used

Parameters: include_data – If True, includes data field, else uses only name and children fields of a graph

shapes() → tuple: Returns the tuple of the shapes of the GDGraphHelper.graphs (see GDGraph.shape)

todatalist() → List[gdata.GData]: Returns the list of the GData objects

validate(graph: gdgraph.GDGraph, recursive: bool = True) → None: Check for equivalence of the children with the data[“arguments”]

validateall(recursive: bool = True) → None: Validation of all the graphs

GAGraph¶

class gagraph.GAGraph(thing=None, filename=None, data=None, string=None, handle=None, name='', strict=True, directed=False, **attr)

Inherited class from pygraphviz.AGraph

Reimplemented to work with GData class

add_edge(u: Union[gdata.GData, Tuple[gdata.GData, gdata.GData], Any], v: Optional[gdata.GData] = None, key: Optional[str] = None, **attr) → pygraphviz.agraph.Edge

Adding edge between nodes to the graph

Reimplemented for the work with GData-class

add_node(data: Any = 1, **attr) → pygraphviz.agraph.Node

Adding node to the graph

Reimplemented for the work with GData-class

Introduction