Refactor of flow tools - XZ-corrections from pattern

[matulni]

This commit introduces tooling to extract an XZCorrections object from a Pattern. It drops all the remaining dependencies on graphix.gflow. Additionally it fixes two unreported bugs (see discussion below).

Description of changes

• Introduced the following methods:

  • graphix.optimization.StandardizedPattern.extract_xzcorrections

    • Constructs the X- and Z-corrections mappings by reading out the pattern's measurements and corrections.
    • Calls the constructor XZCorrections.from_measured_nodes_mapping.

  • graphix.pattern.Pattern.extract_xzcorrections

    • Wrapper which standardizes the pattern and calls the XZ-corrections extraction routine.

• Modified the constructor XZCorrections.from_measured_nodes_mapping so that it doesn't need to create a nx.DiGraph instance. Instead it exploits graphix.flow._partial_order.compute_topological_generations introduced in Refactor of flow tools - Flow from pattern #393. This simplies the code in the public API and fixes an unreported bug in the method (see Code snippets section).

• Fixed an unreported bug in OpenGraph.is_equal_structurally which failed to compare open graphs differing on the output nodes only.

• The modules graphix.gflow and graphix.find_pflow are removed. Note that graphix.visualization does not fully exploit the new flow API but this is out of the scope of this PR.

• Method graphix.Pattern.get_layers is removed (see bug Pattern.get_layers assumes that inputs cannot be outputs #367) in favor of existing graphix.Pattern.extract_partial_order_layers.

Code snippets

The following correction maps and open graph are incorrectly parsed by the current implementation of XZCorrections.from_measured_nodes_mapping in master. This is fixed in this PR.

from graphix.flow.core import XZCorrections
from graphix.opengraph import OpenGraph
from graphix.measurements import Measurement
from graphix.fundamentals import Plane
import networkx as nx

x_corr = {1: {2}, 2: {3}, 0: {4}, 4: {5}, 5: {6}, 6: {7}, 3: {8}, 8: {9}, 7: {10}, 10: {11}, 9: {12}, 12: {13}, 11: {14}, 14: {15}, 15: {16}, 16: {17}, 13: {18}, 18: {19}}
z_corr = {1: {0, 3}, 0: {3, 5}, 5: {7}, 3: {9}, 7: {9, 11}, 9: {11, 13}, 11: {15}, 14: {16}, 15: {17}, 13: {19}}

og = OpenGraph(graph= nx.Graph([(3, 4), (3, 8), (3, 2), (4, 5), (4, 0), (12, 13), (12, 9), (12, 11), (13, 18), (0, 2), (2, 1), (8, 9), (9, 10), (11, 14), (11, 10), (14, 15), (7, 10), (7, 6), (18, 19), (5, 6), (15, 16), (16, 17)]),input_nodes=[0, 1], output_nodes=[17, 19], measurements={1: Measurement(angle=0.0, plane=Plane.XY), 2: Measurement(angle=0.0, plane=Plane.XY), 0: Measurement(angle=0.0, plane=Plane.XY), 4: Measurement(angle=0.0, plane=Plane.XY), 5: Measurement(angle=0.0, plane=Plane.XY), 6: Measurement(angle=-1, plane=Plane.XY), 3: Measurement(angle=-0.5, plane=Plane.XY), 8: Measurement(angle=0.0, plane=Plane.XY), 7: Measurement(angle=0.0, plane=Plane.XY), 10: Measurement(angle=0.0, plane=Plane.XY), 9: Measurement(angle=0.0, plane=Plane.XY), 12: Measurement(angle=0.0, plane=Plane.XY), 11: Measurement(angle=0.5, plane=Plane.XY), 14: Measurement(angle=1.0, plane=Plane.XY), 15: Measurement(angle=-0.5, plane=Plane.XY), 16: Measurement(angle=0.0, plane=Plane.XY), 13: Measurement(angle=0.0, plane=Plane.XY), 18: Measurement(angle=0.25, plane=Plane.XY)})

xz = XZCorrections.from_measured_nodes_mapping(og=og, x_corrections=x_corr, z_corrections=z_corr)

print(xz.partial_order_layers[:2]) # -> (frozenset({17, 19}), frozenset({16, 19})). Node 19 appears twice
xz.to_pattern() # Assertion error due to incorrect partial order

[codecov]

Codecov Report

❌ Patch coverage is 98.03922% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 86.56%. Comparing base (611e4bf) to head (a34eeb4).

Files with missing lines	Patch %	Lines
graphix/flow/core.py	96.77%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #407      +/-   ##
==========================================
+ Coverage   85.60%   86.56%   +0.95%     
==========================================
  Files          46       44       -2     
  Lines        6677     6126     -551     
==========================================
- Hits         5716     5303     -413     
+ Misses        961      823     -138     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

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

[thierry-martinez]

Looks good! Two minor comments.

[thierry-martinez]

I think casting to dict is not necessary.

[matulni]

Opss, you are right. I had used a defaultdict somewhere before but I changed everything to dict. Fixed in 06dad95

[thierry-martinez]

The casting to tuple is no longer necessary since _corrections_to_partial_order_layers returns a tuple.

[matulni]

Fixed in 06dad95

[thierry-martinez]

LGTM! Thanks.