[ziofil]

Context:
The beamsplitter in the Fock representation suffers from numerical instabilities even for moderate cutoffs (~40)

Description of the Change:
This ports the stable BS implementation from MrMustard, which is just as fast as the unstable BS and offers numerical stability up to arbitrary cutoffs (or at least I ran out of memory before I could notice any).

Benefits:
BS can now be applied to high photon number inputs and outputs. It is as fast as the unstable implementation:

But stable:

Possible Drawbacks:
None

Related GitHub Issues:
None

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (12633ca) to head (94459d6).
⚠️ Report is 4 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff            @@
##            master      #405   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           28        28           
  Lines         1922      1946   +24     
=========================================
+ Hits          1922      1946   +24     

Files with missing lines	Coverage Δ
thewalrus/fock_gradients.py	100.00% <100.00%> (ø)

---

Continue to review full report in Codecov by Harness.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 54b1062...94459d6. Read the comment docs.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[nquesada]

You might aswell remove the old one, no?

[ziofil]

Do you want to keep the cutoff as an integer? the new implementation supports any shape.

[ziofil]

@nquesada regarding the cutoff argument, I preserved the old API and made it possible to pass the full shape.