pyaiml21.graphmaster.graphmaster.GraphMaster

class pyaiml21.graphmaster.graphmaster.GraphMaster(bot_config: pyaiml21.botconfig.config.BotConfig, policy: Union[str, Callable[[Any, Any], Any]] = 'keep_last')

GraphMaster is a pattern matching data structure.

It stores AIML categories in the tree-like structure (directed word graph/tree) and uses backtracking to find the first match.

Each AIML category is uniquely identified by its pattern path, which consists of the pattern, that and topic elements. The GraphMaster stores the corresponding template node in such a way, that it can be retrieved by the match method iff the input path matches the pattern path.

Typical order of operations over graphmaster consists of:

1. storing categories:

   1. validate - make sure the category conforms the AIML

      specification and/or the available tag

   2. save(category)

2. retrieving: - realized via match()

Methods

add_pattern_tag	Add pattern tag to the known tags of this graphmaster.
dump	Store the graphmaster into the specified file.
generate_html	Store graphmaster to html file.
load	Load the graphmaster from the file, after it was saved using dump.
load_obj	Load the object given its stored_representation.
match	Find the matching Pattern Path.
save	Save the category into the graphmaster.
save_obj	Store obj and return its stored representation.
size	Return the number of AIML categories.
validate	Validate the category, check its syntax.
vocabulary	Implement <vocabulary> tag.

Attributes

THAT
TOPIC

add_pattern_tag(tag: pyaiml21.graphmaster.pattern.Pattern)

Add pattern tag to the known tags of this graphmaster.

If the same priority tag is present, this tag will be matched first.

dump(filename: str) → None

Store the graphmaster into the specified file.

generate_html(filename: str, *args, **kwargs)

Store graphmaster to html file.

This method can be used for debugging purposes.

Opens the browser with the generated html. It is necessary that pyvis library is also installed.

By default, the graph is directed, 1000 x 1000 px.

Parameters

• filename – name of the file to store generated html into

• args – positional arguments passed to pyvis.network.Network

• kwargs – keyword arguments passed to pyvis.network.Network

Returns

the generated graph

static load(filename: str) → pyaiml21.graphmaster.graphmaster.GraphMaster

Load the graphmaster from the file, after it was saved using dump.

Parameters

filename – name of the file to load graphmaster from

Returns

loaded graphmaster

load_obj(stored_representation: Any) → Any

Load the object given its stored_representation.

See save_obj for more info.

The following should hold for all x:

self.save_obj(self.load_obj(x)) = x self.load_obj(self.save_obj(x)) = x

Parameters

stored_representation – representation of object stored in the nodemappers

Returns

the real object represented by stored_representation

match(input_: str, that_: str, topic_: str) → Optional[Tuple[Any, pyaiml21.graphmaster.stars.StarBindings, List[float]]]

Find the matching Pattern Path.

Given input, that (last bot’s response) and topic, construct input path and find the matching category, retrieving the stored object with the matched stars and a sequence of priorities of matching.

All the arguments should be already normalized.

Parameters

• input – user’s input

• that – last bot’s response

• topic – conversation topic

Returns

None if no matching category was found, else a tuple with - [0] matched object saved with the best-matching category - [1] star bindings for the matching - [2] a sequence of priorities used to find the matching, can be used to compare different matching-s (the bigger value and sooner, the better match)

save(category: pyaiml21.ast.category.Category)

Save the category into the graphmaster.

Traverse the graphmaster according to the pattern path specified by the category, creating new nodes if necessary, and store the template on the specified place, so it can be later retrieved by match.

Successfully calling self.validate is a precondition of correct behaviour of this method.

Parameters

category – a category used to build a pattern path

save_obj(obj: Any) → Any

Store obj and return its stored representation.

For example id which will be stored in the memory.

Parameters

obj – the object that should be stored in the graphmaster

Returns

representation of the object saved in-memory

size() → int

Return the number of AIML categories.

validate(category: pyaiml21.ast.category.Category, logger: pyaiml21.parser.logger.Logger) → None

Validate the category, check its syntax.

Make sure the pattern, that and topic tags contain valid elements. The validation results are stored in the logger instance.

Calling this function successfully (without errors in the logger) is a precondition of storing objects based on this category into this graphmaster.

Parameters

• category – category to validate

• logger – logger to log errors or warning

Returns

nothing, use logger to check for errors

vocabulary() → Set[str]

Implement <vocabulary> tag.

Return a set of words this graphmaster understands (including stars, excluding any word containing underscores).

Approximates <vocabulary/> tag in the sense that it does not count the words in <set>, as it is undefined how many words infinite sets (:number:) have.