Skip to content

[SPN-2931] Tier generation status + decouple quality checks + silence RUF001/003 #26

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
drago1216 merged 1 commit into from
May 29, 2026
Merged

[SPN-2931] Tier generation status + decouple quality checks + silence RUF001/003 #26

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
62 changes: 46 additions & 16 deletions .github/workflows/sdk-update.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -143,7 +143,14 @@ jobs:
path.write_text(json.dumps(data, indent=2) + "\n")
'

# ----- Generate (each step continue-on-error so we can still PR) ------
# ----- Generate + quality checks --------------------------------------
# `make generate` is the only step the others truly depend on (they all
# need the regenerated tree). Format / lint / smoke / test all run
# independently after generate so reviewers see every signal on a single
# run instead of just "the first thing that broke." Each is
# continue-on-error so the workflow proceeds to open the bot PR
# regardless, and the "Determine generation status" step tiers the
# outcomes into one of three states.
- name: make generate
if: steps.fetch.outputs.changed == 'true'
id: generate
Expand All @@ -157,35 +164,48 @@ jobs:
run: make format

- name: make lint
if: steps.fetch.outputs.changed == 'true' && steps.format.outcome == 'success'
if: steps.fetch.outputs.changed == 'true' && steps.generate.outcome == 'success'
id: lint
continue-on-error: true
run: make lint

- name: make smoke-test
if: steps.fetch.outputs.changed == 'true' && steps.lint.outcome == 'success'
if: steps.fetch.outputs.changed == 'true' && steps.generate.outcome == 'success'
id: smoke
continue-on-error: true
run: make smoke-test

- name: make test
if: steps.fetch.outputs.changed == 'true' && steps.smoke.outcome == 'success'
if: steps.fetch.outputs.changed == 'true' && steps.generate.outcome == 'success'
id: test
continue-on-error: true
run: make test

# Three-tier status so reviewers can tell apart the structural failure
# ("the SDK isn't buildable, don't merge") from the quality failure
# ("the SDK is buildable but has lint/test issues to look at"):
#
# generation-failed — `make generate` itself didn't succeed; the
# regenerated tree is missing or broken. PR opens
# with the spec update committed but no usable
# SDK; reviewer needs to investigate before merge.
# quality-issues — generate succeeded but one or more of
# format/lint/smoke/test failed. SDK is structurally
# valid; the issues are usually small and fixable.
# success — everything green; safe to merge once reviewed.
- name: Determine generation status
if: steps.fetch.outputs.changed == 'true'
id: status
run: |
if [ "${{ steps.generate.outcome }}" = "success" ] \
&& [ "${{ steps.format.outcome }}" = "success" ] \
&& [ "${{ steps.lint.outcome }}" = "success" ] \
&& [ "${{ steps.smoke.outcome }}" = "success" ] \
&& [ "${{ steps.test.outcome }}" = "success" ]; then
if [ "${{ steps.generate.outcome }}" != "success" ]; then
echo "result=generation-failed" >> "$GITHUB_OUTPUT"
elif [ "${{ steps.format.outcome }}" = "success" ] \
&& [ "${{ steps.lint.outcome }}" = "success" ] \
&& [ "${{ steps.smoke.outcome }}" = "success" ] \
&& [ "${{ steps.test.outcome }}" = "success" ]; then
echo "result=success" >> "$GITHUB_OUTPUT"
else
echo "result=failed" >> "$GITHUB_OUTPUT"
echo "result=quality-issues" >> "$GITHUB_OUTPUT"
fi

# ----- Skip-if-no-meaningful-changes ----------------------------------
Expand Down Expand Up @@ -283,19 +303,29 @@ jobs:
TITLE="[auto] SDK update (${{ steps.classify.outputs.level }}) — spec ${{ steps.fetch.outputs.source_commit }}"

# Build the exact label set for this run. `auto-generated` is
# always present; `breaking` and `generation-failed` are toggled.
# always present; `breaking`, `generation-failed`, and
# `quality-issues` are toggled based on this run's outcomes.
# See the "Determine generation status" step for tier definitions.
ADD_LABELS=("auto-generated")
REMOVE_LABELS=()
if [ "${{ steps.analyze.outputs.has_breaking }}" = "true" ]; then
ADD_LABELS+=("breaking")
else
REMOVE_LABELS+=("breaking")
fi
if [ "${{ steps.status.outputs.result }}" = "failed" ]; then
ADD_LABELS+=("generation-failed")
else
REMOVE_LABELS+=("generation-failed")
fi
case "${{ steps.status.outputs.result }}" in
generation-failed)
ADD_LABELS+=("generation-failed")
REMOVE_LABELS+=("quality-issues")
;;
quality-issues)
ADD_LABELS+=("quality-issues")
REMOVE_LABELS+=("generation-failed")
;;
success)
REMOVE_LABELS+=("generation-failed" "quality-issues")
;;
esac

ADD_LABELS_CSV="$(IFS=,; echo "${ADD_LABELS[*]}")"

Expand Down
18 changes: 12 additions & 6 deletions pyproject.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -105,27 +105,33 @@ known-first-party = ["bamboohr_sdk"]
# the upstream templates and aren't worth fixing in our fork.
# Hand-written code lives under bamboohr_sdk/client/ and stays under the
# default rule set.
#
# RUF001/RUF002/RUF003 (ambiguous unicode in strings/docstrings/comments) are
# all ignored on generated paths — they fire on smart-quote characters (’, “”)
# that come straight from spec description text written by API designers. The
# characters aren't ambiguous in context; ruff is being pedantic about source
# we don't own.
[tool.ruff.lint.per-file-ignores]
"bamboohr_sdk/api/*.py" = [
"B028", "B904", "F841", "N805", "RUF002", "RUF012", "RUF059",
"B028", "B904", "F841", "N805", "RUF001", "RUF002", "RUF003", "RUF012", "RUF059",
"SIM101", "SIM105", "SIM108", "SIM118", "TC001", "TC003", "UP030", "UP031",
]
"bamboohr_sdk/models/*.py" = [
"F841", "N805", "N818", "RUF002", "RUF012", "SIM118", "TC001", "TC003",
"F841", "N805", "N818", "RUF001", "RUF002", "RUF003", "RUF012", "SIM118", "TC001", "TC003",
]
"bamboohr_sdk/api_client.py" = [
"B028", "B904", "F841", "RUF002", "RUF012", "RUF059",
"B028", "B904", "F841", "RUF001", "RUF002", "RUF003", "RUF012", "RUF059",
"SIM101", "SIM105", "SIM108", "TC001", "TC003", "UP030", "UP031",
]
"bamboohr_sdk/api_helper.py" = ["TC001", "TC003"]
"bamboohr_sdk/api_response.py" = ["TC001", "TC003"]
"bamboohr_sdk/configuration.py" = ["B028", "B904", "RUF002", "UP030", "UP031"]
"bamboohr_sdk/configuration.py" = ["B028", "B904", "RUF001", "RUF002", "RUF003", "UP030", "UP031"]
"bamboohr_sdk/exceptions.py" = ["N818", "SIM105"]
"bamboohr_sdk/rest.py" = ["B028", "B904", "RUF002", "SIM101", "SIM108"]
"bamboohr_sdk/rest.py" = ["B028", "B904", "RUF001", "RUF002", "RUF003", "SIM101", "SIM108"]
# Hand-written client code: relax rules for type-only imports and
# docstring-only stylistic noise. Real bugs would still surface via
# mypy and the test suite.
"bamboohr_sdk/client/**/*.py" = ["B904", "RUF002", "RUF059", "TC001", "TC003"]
"bamboohr_sdk/client/**/*.py" = ["B904", "RUF001", "RUF002", "RUF003", "RUF059", "TC001", "TC003"]

[tool.ruff.format]
quote-style = "double"
Expand Down
28 changes: 20 additions & 8 deletions scripts/build_pr_body.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,9 @@
--since-last-review Path to a pre-computed delta-vs-existing-branch text
file. Pass --since-last-review="" (or omit) to
render the "Initial PR" placeholder.
--generation-status One of: success, failed. Adds a banner if failed.
--generation-status One of: success, quality-issues, generation-failed.
Each tier shows a different banner (or none) at the
top of the PR body.
--old-version Previous version (e.g. 1.0.0).
--new-version Bumped version (e.g. 1.1.0); pass equal to old to
skip the version-bump section.
Expand Down Expand Up @@ -59,14 +61,24 @@ def parse_classification(text: str) -> tuple[str, str]:
def build(args: argparse.Namespace) -> str:
sections: list[str] = []

# --- Banner for generation failures ------------------------------------
if args.generation_status == "failed":
# --- Banner keyed off the three-tier generation status -----------------
if args.generation_status == "generation-failed":
sections.append(
"> [!CAUTION]\n"
"> **`make generate` failed on this run.**\n"
"> The committed tree contains the spec update but no regenerated\n"
"> SDK — the package is not buildable as-is. Reproduce locally\n"
"> with `make generate` and resolve the underlying error before\n"
"> merging."
)
elif args.generation_status == "quality-issues":
sections.append(
"> [!WARNING]\n"
"> **Generation, lint, or tests failed on this run.**\n"
"> The committed tree contains the spec update but may not contain\n"
"> a clean regenerated SDK. Reproduce locally with `make generate`\n"
"> then `make format && make lint && make test` before merging."
"> **Generation succeeded, but one or more quality checks failed.**\n"
"> The regenerated SDK is structurally valid; one or more of\n"
"> `make format` / `make lint` / `make smoke-test` / `make test`\n"
"> reported issues. See the workflow run for per-step details and\n"
"> reproduce locally with the failing target before merging."
)

# --- Spec source metadata ----------------------------------------------
Expand Down Expand Up @@ -167,7 +179,7 @@ def parse_args() -> argparse.Namespace:
parser.add_argument("--since-last-review", default="")
parser.add_argument(
"--generation-status",
choices=["success", "failed"],
choices=["success", "quality-issues", "generation-failed"],
default="success",
)
parser.add_argument("--old-version", default="")
Expand Down