Qiskit / qiskit
89%
main: 88%

LAST BUILD BRANCH: gh-readonly-queue/main/pr-15094-96f5c08b3f54acfcc9552a34bf1bcc5402e07e7f
DEFAULT BRANCH: main
Repo Added
Files 868
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst

LAST BUILD ON BRANCH gh-readonly-queue/main/pr-14286-f367f9a2dcc6b6c451a1c9e9e9e1e40868ea0527
branch: gh-readonly-queue/main/pr-14286-f367f9a2dcc6b6c451a1c9e9e9e1e40868ea0527
CHANGE BRANCH
x

coverage: 88.718% (+0.02%) from 88.7%
14930451342

push

github

web-flow 
Split out separate qiskit-transpiler crate (#14286)

* Split out separate qiskit-transpiler crate

This commit creates a new qiskit-transpiler crate. As we move towards
exposing more rust functionality directly via the C API having more
explicit interface boundaries in the rust code is important for keeping
things organized and managing dependencies (both external and internal).
We also are building towards a case where the core functionality exists
without a dependency on python, so more explicitly splitting
functionality makes it clearer where the expected boundaries are. The
ideal end state is where we can separate the API of qiskit-transpiler
(and the existing qiskit-circuit) to be pure rust and then move the
python interface components to only live in the accelerate or pyext
crate. But that is a longer term vision and this is only one step
towards that goal.

The transpiler functionality in Rust has grown to the point where it is
both sufficiently self contained and also large enough to warrant its
own compilation unit. Especially because the transpiler is the next big
frontier in the C API, so having it be standalone will make it simpler
to expose via cext.

This PR isn't a 100% clean separation yet though. Ideally
qiskit-transpiler would not depend on qiskit-accelerate. However this PR
still has qiskit-transpiler dependency on qiskit-accelerate. This is for
2 reasons primarily. The first is there are synthesis and quantum info
pieces that are still in qiskit-accelerate which several passes depend.
The next step after this PR is to migrate those into a standalone crate
for synthesis and/or quantum info.  The second component is the nlayout
module. This should live in the transpiler crate, but was not migrated
as part of this PR. This will be done in a follow-up as ideally we
should move PhysicalQubit and VirtualQubit to qiskit-circuit, and the
NLayout struct should move to transpiler. This was deferred as part of
this commit to minimize the chan... (continued)

249 of 270 new or added lines in 26 files covered. (92.22%)

2 existing lines in 1 file now uncovered.

76052 of 85723 relevant lines covered (88.72%)

429514.14 hits per line

Relevant lines Covered
85723 RELEVANT LINES 76052 COVERED LINES
429514.14 HITS PER LINE
Source Files on gh-readonly-queue/main/pr-14286-f367f9a2dcc6b6c451a1c9e9e9e1e40868ea0527

Recent builds

Builds Branch Commit Type Ran Committer Via Coverage
14930451342 gh-readonly-queue/main/pr-14286-f367f9a2dcc6b6c451a1c9e9e9e1e40868ea0527 Split out separate qiskit-transpiler crate (#14286) * Split out separate qiskit-transpiler crate This commit creates a new qiskit-transpiler crate. As we move towards exposing more rust functionality directly via the C API having more explicit i... push web-flow github
88.72
See All Builds (13861)