py4j/.github/workflows/perf-framework.yml at master · byteoak/py4j · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
name: perf-framework

# Run the perf-framework self-tests once on Linux + latest Python +
# latest LTS Java. The framework itself is local-only (it spawns its
# own JVMs), so running its self-tests on the full Python x Java x OS
# matrix (56 cells) gives no additional coverage — the tests are pure
# Python and the JVM-spawning path is the same on every cell. One
# representative cell suffices.
#
# Triggers:
#   - any change under py4j-python/src/py4j/tests/perf/
#   - any change to test_perf_framework.py
#   - any change to requirements-test.txt (deps the framework needs)
#   - any change to this workflow file
on:
  push:
    branches:
    - master
    paths:
    # Include perf-framework code, excluding doc-only changes within it
    # (GitHub Actions rejects `paths` + `paths-ignore` together in the
    # same trigger; the `!` negation prefix inside a single `paths:` list
    # is the supported way to combine include + exclude).
    - 'py4j-python/src/py4j/tests/perf/**'
    - '!py4j-python/src/py4j/tests/perf/**/*.md'
    - 'py4j-python/src/py4j/tests/test_perf_framework.py'
    - 'py4j-python/requirements-test.txt'
    - '.github/workflows/perf-framework.yml'
  pull_request:
    branches:
    - master
    paths:
    - 'py4j-python/src/py4j/tests/perf/**'
    - '!py4j-python/src/py4j/tests/perf/**/*.md'
    - 'py4j-python/src/py4j/tests/test_perf_framework.py'
    - 'py4j-python/requirements-test.txt'
    - '.github/workflows/perf-framework.yml'
  # Manual trigger from the Actions UI — useful when you want to
  # re-run perf coverage after merging upstream master or to test the
  # workflow itself.
  workflow_dispatch:

jobs:
  perf-framework-tests:
    name: Perf framework self-tests
    runs-on: ubuntu-latest
    timeout-minutes: 10
    steps:
      - uses: actions/checkout@v6

      - name: Setup Java 8 JDK for build
        # py4j-java targets source/target 1.6 so it has to compile with
        # Java 8 specifically — newer JDKs reject Source option 6.
        uses: actions/setup-java@v5
        with:
          java-version: '8'
          distribution: 'zulu'
          cache: 'gradle'

      - name: Setup Python 3.13
        uses: actions/setup-python@v6
        with:
          python-version: '3.13'

      - name: Install Python deps
        run: |
          cd py4j-python
          # -e . is required so `python -m py4j.tests.perf` (the
          # smoke-run step below) can find the module. Just installing
          # requirements-test.txt isn't enough — py4j itself has to be
          # on the path. pytest finds the source automatically via
          # pyproject/conftest discovery, but python -m does not.
          pip install -e .
          pip install -r requirements-test.txt

      - name: Build py4j-java classes (needed for JVM-spawn tests)
        run: |
          cd py4j-java
          ./gradlew classes testClasses

      - name: Setup Java 17 JDK for the perf-framework runtime
        # The framework spawns JVMs at runtime; latest LTS is fine
        # (py4j supports 8+). Java 8 stays available for any back-
        # compat smoke later.
        uses: actions/setup-java@v5
        with:
          java-version: '17'
          distribution: 'zulu'

      - name: Run perf-framework self-tests
        run: |
          cd py4j-python
          pytest src/py4j/tests/test_perf_framework.py -vvv

      - name: Smoke-run the framework end-to-end (single scenario, --quick)
        run: |
          cd py4j-python
          # Sanity: the framework can spawn a JVM, run a micro
          # scenario, and emit a report without crashing. Doesn't
          # gate on measurement quality (CI noise makes that
          # meaningless); just verifies the orchestration works.
          python -m py4j.tests.perf --quick --only M1 --no-renice \
            --output-dir /tmp/perf-smoke || exit 1
          test -f /tmp/perf-smoke/perf_report.json
          test -f /tmp/perf-smoke/perf_report.md
          echo "Smoke run OK"