I have a strong preference for analytics / diagnostics happening outside of the system. Export the copy of the database and the journals, start up the analytics notebook and sort it out like any other data science problem.

It really depends on the whys. The high end, there is a 'PROBLEM' with what the system is doing is to pull a copy of the graph and journals and work backwards to find root cause. As the problems get weirder, we will be building more tooling to automate the analysis.

The edge facets indicate the type of connection and maintain meta-data about the relationship. The nodes contain the data about the type of node, where the data came from, and other meta-data.

When the tasks are decomposed, those tasks receive IDs, they are not exposed outside of the system, but they are used to collect the requirements for composing the PR request.

Requirements get decomposed by the Knowledge Engine and relayed to the Actor. There is a journal for this. The Actor takes the decomposed tasks and generates development tasks which are provided to the Motivator. There is a journal for this. The Motivator provides the tasks to the code generation component (Spyder). Which works the code. When Spyder is finished with the code a new task is provided to the Motivator for the code to be tested. The code is recorded, the pass of is logged. The Motivator sends the code to the Actor who tests the code. If the code passes it is reviewed by the Regulator and sent to the Code repo as a commit. Logged in system and a code history from Git. When the criteria for a pull request is met, the Actor composes the PR text, and the Regulator creates the PR.

Initial development / testing has been using spoj.com/problems/classical/ as they are objective. Before it is released, the Python instance will be pointed at a number of idle projects. It won't be released until there is a public face project.

The Python Developer is the first persona that we are building. The team has enough domain expertise to ensure Tapestry capable at that task, and my approach is different than what everybody else is doing.