conceptnet_lite

Module contents

class conceptnet_lite.PartOfEdge

Bases: enum.Enum

An enumeration.

ANY = 'any'
END = 'end'
START = 'start'
conceptnet_lite.connect(db_path='conceptnet.db', db_download_url='https://conceptnet-lite.fra1.cdn.digitaloceanspaces.com/conceptnet.db.zip', delete_compressed_db=True, dump_download_url='https://s3.amazonaws.com/conceptnet/downloads/2019/edges/conceptnet-assertions-5.7.0.csv.gz', load_dump_edge_count=34074917, delete_compressed_dump=True, delete_dump=True)

Connect to ConceptNet database.

This function connects to ConceptNet database. If it does not exists, there are two options: to download ready database or to download the compressed ConceptNet dump, extract it, and load it into database (pass db_download_url=None for this option).

Parameters

  • db_path (Union[Path, str]) – Path to the database.

  • db_download_url (Optional[str]) – Link to compressed ConceptNet database. Pass None to build the db from dump.

  • delete_compressed_db (bool) – Delete compressed database after extraction.

  • dump_download_url (str) – Link to compressed ConceptNet dump.

  • load_dump_edge_count (int) – Number of edges to load from the beginning of the dump file. Can be useful for testing.

  • delete_compressed_dump (bool) – Delete compressed dump after unpacking.

  • delete_dump (bool) – Delete dump after loading into database.

Return type

None

conceptnet_lite.edges(part_of_edge, concepts, relation=None, same_language=False)
Return type

ModelSelect

conceptnet_lite.edges_between(start_concepts, end_concepts, relation=None, two_way=False)
Return type

ModelSelect

conceptnet_lite.edges_for(concepts, relation=None, same_language=False)
Return type

ModelSelect

conceptnet_lite.edges_from(start_concepts, relation=None, same_language=False)
Return type

ModelSelect

conceptnet_lite.edges_to(end_concepts, relation=None, same_language=False)
Return type

ModelSelect

conceptnet_lite.db module

class conceptnet_lite.db.Concept(*args, **kwargs)

Bases: conceptnet_lite.db._BaseModel

Concept ORM class.

Concept represents node in ConceptNet knowledge graph. It provides properties language and text that are aliases for corresponding Label.language and Label.text fields.

This abstraction is not present in the original ConceptNet. Class is introduced for the purposes of normalization.

DoesNotExist

alias of ConceptDoesNotExist

edges_in
edges_out
classmethod get(*query, **filters)
id = <AutoField: Concept.id>
label = <ForeignKeyField: Concept.label>
label_id = <ForeignKeyField: Concept.label>
property language
Return type

Language

sense_label = <TextField: Concept.sense_label>
property text
Return type

str

property uri
Return type

str

class conceptnet_lite.db.Edge(*args, **kwargs)

Bases: conceptnet_lite.db._BaseModel

Edge ORM class.

See: https://github.com/commonsense/conceptnet5/wiki/Edges.

Everything except relation, start, and end nodes is stored in etc field that is plain dict.

DoesNotExist

alias of EdgeDoesNotExist

end = <ForeignKeyField: Edge.end>
end_id = <ForeignKeyField: Edge.end>
etc = <JSONField: Edge.etc>
classmethod get(*query, **filters)
id = <AutoField: Edge.id>
relation = <ForeignKeyField: Edge.relation>
relation_id = <ForeignKeyField: Edge.relation>
start = <ForeignKeyField: Edge.start>
start_id = <ForeignKeyField: Edge.start>
property uri
Return type

str

class conceptnet_lite.db.Label(*args, **kwargs)

Bases: conceptnet_lite.db._BaseModel

Label ORM class.

Label can be seen as a part of Concept. Label is basically a text on a certain language (most often, a word).

This abstraction is not present in the original ConceptNet. Class is introduced for the purposes of normalization.

DoesNotExist

alias of LabelDoesNotExist

concepts
classmethod get(*query, **filters)
id = <AutoField: Label.id>
language = <ForeignKeyField: Label.language>
language_id = <ForeignKeyField: Label.language>
text = <TextField: Label.text>
class conceptnet_lite.db.Language(*args, **kwargs)

Bases: conceptnet_lite.db._BaseModel

Language ORM class.

See: https://github.com/commonsense/conceptnet5/wiki/Languages.

DoesNotExist

alias of LanguageDoesNotExist

id = <AutoField: Language.id>
labels
name = <TextField: Language.name>
class conceptnet_lite.db.Relation(*args, **kwargs)

Bases: conceptnet_lite.db._BaseModel

Relation ORM class.

See: https://github.com/commonsense/conceptnet5/wiki/Relations.

DoesNotExist

alias of RelationDoesNotExist

edges
id = <AutoField: Relation.id>
name = <TextField: Relation.name>
property uri
Return type

str

class conceptnet_lite.db.RelationName

Bases: object

Names of non-deprecated relations.

See: https://github.com/commonsense/conceptnet5/wiki/Relations.

ANTONYM = 'antonym'
AT_LOCATION = 'at_location'
CAPABLE_OF = 'capable_of'
CAUSES = 'causes'
CAUSES_DESIRE = 'causes_desire'
CREATED_BY = 'created_by'
DEFINED_AS = 'defined_as'
DERIVED_FROM = 'derived_from'
DESIRES = 'desires'
DISTINCT_FROM = 'distinct_from'
ETYMOLOGICALLY_DERIVED_FROM = 'etymologically_derived_from'
EXTERNAL_URL = 'external_url'
FORM_OF = 'form_of'
HAS_A = 'has_a'
HAS_CONTEXT = 'has_context'
HAS_FIRST_SUBEVENT = 'has_first_subevent'
HAS_LAST_SUBEVENT = 'has_last_subevent'
HAS_PREREQUISITE = 'has_prerequisite'
HAS_PROPERTY = 'has_property'
HAS_SUBEVENT = 'has_subevent'
IS_A = 'is_a'
LOCATED_NEAR = 'located_near'
MADE_OF = 'made_of'
MANNER_OF = 'manner_of'
MOTIVATED_BY_GOAL = 'motivated_by_goal'
OBSTRUCTED_BY = 'obstructed_by'
PART_OF = 'part_of'
RECEIVES_ACTION = 'receives_action'
RELATED_TO = 'related_to'
SIMILAR_TO = 'similar_to'
SYMBOL_OF = 'symbol_of'
SYNONYM = 'synonym'
USED_FOR = 'used_for'
conceptnet_lite.db.download_db(url='https://conceptnet-lite.fra1.cdn.digitaloceanspaces.com/conceptnet.db.zip', db_path='conceptnet.db', delete_compressed_db=True)

Download compressed ConceptNet dump and extract it.

Parameters

  • url (str) – Link to compressed ConceptNet database.

  • db_path (Union[Path, str]) – Path to resulting database.

  • delete_compressed_db (bool) – Delete compressed database after extraction.

Return type

None

conceptnet_lite.db.download_dump(url='https://s3.amazonaws.com/conceptnet/downloads/2019/edges/conceptnet-assertions-5.7.0.csv.gz', out_dir_path=PosixPath('/home/docs/checkouts/readthedocs.org/user_builds/conceptnet-lite/checkouts/latest/docs/source'))

Download compressed ConceptNet dump.

Parameters

  • url (str) – Link to the dump.

  • out_dir_path (Union[Path, str]) – Dir where to store dump.

conceptnet_lite.db.extract_compressed_dump(compressed_dump_path, delete_compressed_dump=True)

Extract compressed ConceptNet dump.

Parameters

  • compressed_dump_path (Union[Path, str]) – Path to compressed dump to extract.

  • delete_compressed_dump (bool) – Delete compressed dump after extraction.

conceptnet_lite.db.load_dump_to_db(dump_path, db_path, edge_count=34074917, delete_dump=True)

Load dump to database.

Parameters

  • dump_path (Union[Path, str]) – Path to dump to load.

  • db_path (Union[Path, str]) – Path to resulting database.

  • edge_count (int) – Number of edges to load from the beginning of the dump file. Can be useful for testing.

  • delete_dump (bool) – Delete dump after loading into database.

conceptnet_lite.db.prepare_db(db_path, dump_download_url='https://s3.amazonaws.com/conceptnet/downloads/2019/edges/conceptnet-assertions-5.7.0.csv.gz', load_dump_edge_count=34074917, delete_compressed_dump=True, delete_dump=True)

Prepare ConceptNet database.

This function downloads the compressed ConceptNet dump, extracts it, and loads it into database. First two steps are optional, and are executed only if needed.

Parameters

  • db_path (Union[Path, str]) – Path to the resulting database.

  • dump_download_url (str) – Link to compressed ConceptNet dump.

  • load_dump_edge_count (int) – Number of edges to load from the beginning of the dump file. Can be useful for testing.

  • delete_compressed_dump (bool) – Delete compressed dump after extraction.

  • delete_dump (bool) – Delete dump after loading into database.

database schema

_images/erd.svg