Topological level and shortest dependency path of dependency tree (directed acyclic graph)

[dachengx]

What is the problem / what does the code in this PR do

The dependency structure forms a directed acyclic graph (DAG), allowing us to define a topological level for each node, as implemented in #896. This new PR extends that by also computing the shortest dependency path within the DAG:

1. Topological level: level(node) = 1 + max(level(dep)) Represents the longest path from any source node to the given node — useful for understanding how late a node occurs in the dependency chain.
2. Shortest dependency path: level(node) = 1 + min(level(dep)) Captures the minimum number of steps from a source node to the current node, identifying the shallowest point of dependency.

By combining both metrics, we can better analyze the structure of the dependency tree. In particular, we can identify key nodes that act as "connectors" or "cut points": if we divide the tree at such a node, one of the resulting subtrees will contain all nodes that depend on it, effectively capturing a full dependency closure.

That node is the node with a unique topological level and shortest dependency path among all nodes.

Previously, in reprocessing, only data_types that depend on one depends_on are allowed in per-chunk storage. Now we allow multiple depends_on, if all depends_on eventually depend on a single data_type, and all intermediate data_types are not stored.

Can you briefly describe how it works?

Can you give a minimal working example (or illustrate with a figure)?

Please include the following if applicable:

• Update the docstring(s)
• Update the documentation
• Tests to check the (new) code is working as desired.
• Does it solve one of the open issues on github?

Please make sure that all automated tests have passed before asking for a review (you can save the PR as a draft otherwise).

[coveralls]

coverage: 88.956% (-0.1%) from 89.063%
when pulling 7dad875 on dag_shortest_path
into 328dc07 on master.