Steve Kieffer

The funny part is happening in `mpmath`, not in `sympy`. ```python import mpmath mpmath.identify(4678.0) Out[3]: '(3**(194/35)*7**(19/7))/(2**(108/35)*5**(17/35))' ``` For us the call happens here: https://github.com/sympy/sympy/blob/39de9a2698ad4bb90681c0fdb70b30a78233145f/sympy/simplify/simplify.py#L1480-L1481

Rebased on 6.8.0. The `--cut-deps` feature is now only in the `LegacyResolver`.

> Where did #1645 come from? It addresses problems like [this one](https://github.com/aio-libs/aioredis-py/issues/663) and [this one](https://github.com/apache/tinkerpop/pull/1733). As I see it, both #1645 and #1509 are important, because they are complementary. #1509...

Actually it would be very easy to add a third usage pattern to #1645, like this: ```shell pip-compile --cut-deps :C ``` (note the leading `:`), meaning, "Cut all _incoming_ edges...

It looks like both outer and inner self-loops can be used in the same model: [example](https://rtsys.informatik.uni-kiel.de/elklive/json.html?compressedContent=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-2yOMMvKJJMFMVwhAAwjhfMk9D52lAAGIASQAGgBRAAiAH1ZaXiyL7bQY3iCfGBSFKUDpmC5pCqTJhgA3PQEOhJFLwzKe5M5548S31-YCG62SBO2AupBBtBQb2+-2u4NL9G-aqzk7+YoAdWLATldb23hBAA9SPuEdldflOkUBO9PMvnfTAw9LcfT9AMkDdT4rkPCNj3GcQzwEYsNWvW8dB8R9XU6ZFOGwzRPX1Q1jVNBQKFEd8uBDedbTtT0VzXMA3Rfett1AwCJzDDFIxPBs435UlWyTfACDTDM8Czbwc0gAsSwrKsa1QqAeSbPjE3bEEZnBeY9F7aAByHEdUlMcioKEp4dk-KjrgUiY6IA9cgK9EDd3ApjlGgzExkuSBgWHARLxQyC53vOgnyQVyeFwmhjItCzrQsH8l1stjgJ3MCGPC3h3K4uDvIQqAkICnL0JCzCNjfUw7j1A0jRNM02FIzJMFIjIgA) by simply setting `insideSelfLoops.yo` to be `true` or `false`, according to what you want....

This seems strange because the whole point of eclipse/elk#775 was that edge containments are now automatically computed, meaning it shouldn't matter where you define the edge. [The docs state](https://www.eclipse.org/elk/documentation/tooldevelopers/graphdatastructure.html#working-with-the-graph-data-structure) that...

> If I add self loop from same to same port the edge is just twisted. What is the expected behavior?

One obvious fix for all this would be to supply a switch to recover legacy behavior, but I don't think it makes sense to do that if that behavior was...

Okay, so it sounds like containment really is functional. In other words, if the graph is given formally as $(V, E, H)$ (vertices, edges, hierarchy), then there is a well-defined...

It sounds like it will be solved by [this PR](https://github.com/eclipse/elk/pull/775), which is slated for inclusion in ELK 0.8.0 (hence also elkjs 0.8.0).