The Session Object

Session

 Session (register_stdlib=True)

A Session object is the main interface to the spannerlog engine. It is used to parse, check semantics, plan and execute queries. It allows importing data and callbacks to the Spannerlog engine and exporting data from the engine back to python.

	Type	Default	Details
register_stdlib	bool	True	if True, registers the standard library of IEs and AGGs

source

Session.clear

 Session.clear (register_stdlib=True)

Resets the engine and clears all relations, functions and rules.

	Type	Default	Details
register_stdlib	bool	True	if True, registers the standard library of IEs and AGGs

Importing information to spannerlog

source

Session.register

 Session.register (name, func, in_schema, out_schema)

Registers an IE function with the spannerlog engine.

	Details
name	name of the IE function in spannerlog
func	the python function that implements the IE
in_schema	the schema of the input relation
out_schema	the schema of the output relation

source

Session.register_agg

 Session.register_agg (name, func, in_schema, out_schema)

Registers an AGG function with the spannerlog engine.

	Details
name	name of the AGG function in spannerlog
func	the python function that implements the AGG
in_schema	the schema of the input relation, can be of arity 1 only
out_schema	the schema of the output relation, can be of arity 1 only

source

Session.import_rel

 Session.import_rel (name:str,
                     data:Union[str,pathlib.Path,pandas.core.frame.DataFra
                     me], delim:str=None, header=None)

Imports a relation into the current session, either from a dataframe or from a csv file.

	Type	Default	Details
name	str		name of the relation in spannerlog
data	Union		either a pandas dataframe or a path to a csv file
delim	str	None	the delimiter of the csv file
header	NoneType	None	the header of the csv file

source

Session.import_var

 Session.import_var (name, value)

Imports a variable into the current session.

	Details
name	name of the variable in spannerlog
value	the value of the variable

Exporting data from spannerlog to python

source

Session.export

 Session.export (code:str, display_results=False, draw_query=False,
                 plan_query=False, return_statements_meta=False)

Takes a string of spannerlog code, and executes it, returning the value of the last statement in the code string. All statements that are not queries, return None.

	Type	Default	Details
code	str		the spannerlog code to execute
display_results	bool	False	if True, displays the results of the query to screen
draw_query	bool	False	if True, draws the query graph of queries to screen
plan_query	bool	False	if True, if last statement is a query, plans the query and returns the query graph and root node.
return_statements_meta	bool	False	if True, returns both the return value and the statements meta data, used internally.

source

Session.get_all_functions

 Session.get_all_functions ()

Returns all the IEs and AGGs in the engine, as a nested dictionary of the form: { ‘ie’:{name:IEFunction}, ‘agg’:{name:AGGFunction} }

source

Session.print_rules

 Session.print_rules ()

Prints all the rules in the engine. and returns them as a list

Removing information

These functions are mostly used when debugging spannerlog code, to remove rules and relations we want to redefine.

source

Session.remove_rule

 Session.remove_rule (rule:str)

removes a rule from the engine, rule string must be identical to the rule defined previously

	Type	Details
rule	str	the rule string to remove

source

Session.remove_head

 Session.remove_head (head:str)

removes all rules of a given head relation

source

Session.remove_all_rules

 Session.remove_all_rules ()

removes all rules from the engine

source

Session.remove_relation

 Session.remove_relation (relation:str)

removes a relation from the engine, either a extrinsic or intrinsic relation

Examples

sess = Session()
df1 = pd.DataFrame([['John Doe', 35],['Jane Smith', 28]],columns=['X','Y'])
df2 = pd.DataFrame([['John Doe', 30]],columns=['X','Y'])

sess.import_rel("AgeOfKids",df1)
display.display(sess.export("?AgeOfKids(X,Y)"))
sess.import_rel("AgeOfKids",df2)
display.display(sess.export("?AgeOfKids(X,Y)"))

	X	Y
0	Jane Smith	28
1	John Doe	35

	X	Y
0	Jane Smith	28
1	John Doe	30
2	John Doe	35

# basic export with metadata
sess = Session()

res = sess.export("""
new A(int)
""")
display.display(res)

None

sess.engine.Relation_defs

{'A': RelationDefinition(name='A', scheme=[<class 'int'>])}