In addition, track permissions (i.e., can-view, can-edit and can-comment) and ownership on the connection between a user and a topic map.

Source:

• topic-db/topicdb/core/store/topicstore.py

  Line 2137 in 444856a

  def initialise_topic_map(self, map_identifier: int, user_identifier: int) -> None:

• topic-db/topicdb/core/store/topicstore.py

  Line 70 in 444856a

  self.base_topics = {

See URL: http://doc.pytest.org/en/latest/contents.html

Currently, the GetTopicIdentifiers command class returns topic identifiers regardless of the type (i.e., instance of) of the topics. It should be possible to filter what type of topic identifiers should be returned. This is already relevant for StoryTechnologies' scene browser.

This change will have consequences downstream (i.e., Contextualise).

See URL: https://www.python.org/downloads/release/python-3110b5/

This will require refactoring of the StoryTechnologies Story Engine (https://github.com/brettkromkamp/story_engine) code base.

Current set of tests are whole insufficient: https://github.com/brettkromkamp/topic-db/tree/master/tests

The tests need to be reviewed (probably updated) and appropriately extended to ensure sufficient coverage.

Expected output:

elon-musk
         lyndon-rive
                 solar-city
                         solar-energy
         space-x
                 falcon-9
                 falcon-heavy
                 dragon
         tesla
                 model-s
                 model-x
                 model-y
                 roadster

Actual (erroneous) output:

elon-musk
         tesla
                 roadster
                         model-y
                                 model-x
                 model-s
                 space-x
                         dragon
                                 falcon-heavy
                         falcon-9
         lyndon-rive
                 solar-city
                         solar-energy

Order of the siblings is not necessarily relevant, but overall "shape" of the tree/graph is.

Method:

Currently, the topic map engine uses SQLite as its persistence store. It should be possible to use different persistence stores (e.g., PostgreSQL).

Implementation detail: I will probably implement this using the strategy pattern.

Examples:

• from topicdb.core import * (imports both commands and models)
• from topicdb.core.commands import * (imports all of the command submodules)
• from topicdb.core.models import * (imports all of the models plus the tree submodule)

See URL: https://www.sqlite.org/fasterthanfs.html

See URL: https://www.psycopg.org/articles/2021/06/16/psycopg-29-released/

Currently, the different command classes (e.g., GetAssociations, GetAttributes, GetTopicsHierarchy) allow for retrieving topic map-related data. Nonetheless, it should be possible to return data for user-defined queries.

See URL: https://docs.python.org/3/library/contextlib.html#contextlib.closing

Refactor TopicDB to be asynchronous.

Relevant links:

• aiosqlite: https://github.com/omnilib/aiosqlite (and https://aiosqlite.omnilib.dev/en/stable/)

The GetAssociations command will be renamed to GetTopicAssociations (retrieve all associations for a given topic). And, the GetAssociations will become a command class to retrieve associations, optionally filtered by 'instance of' and 'scope'.