A high-performance inference engine for AI models
fix(ci): avoid runner temp parse error (#314)
Artur Mullakhmetovf6d2e6071b55
diff --git a/.github/workflows/run-benchmark.yml b/.github/workflows/run-benchmark.ymlindex fed4f884d..5ff03ae24 100644--- a/.github/workflows/run-benchmark.yml+++ b/.github/workflows/run-benchmark.yml@@ -34,7 +34,6 @@ jobs:environment: prefect-benchmarksenv:- ARTIFACT_DIR: ${{ runner.temp }}/prefect-benchmarkBENCHMARK_ARGS_JSON: ${{ inputs.benchmark_args }}GITHUB_ACTOR_NAME: ${{ github.actor }}GITHUB_TRIGGERING_ACTOR_NAME: ${{ github.triggering_actor }}@@ -52,6 +51,7 @@ jobs:env:PREFECT_API_KEY: ${{ secrets.PREFECT_API_KEY }}run: |+ ARTIFACT_DIR="$RUNNER_TEMP/prefect-benchmark"python3 scripts/prefect_benchmark_run.py \--api-url "$PREFECT_API_URL" \--pool "$POOL_NAME" \@@ -65,12 +65,13 @@ jobs:uses: actions/upload-artifact@v4with:name: prefect-benchmark-${{ github.run_id }}- path: ${{ env.ARTIFACT_DIR }}+ path: ${{ runner.temp }}/prefect-benchmarkif-no-files-found: warn- name: Write workflow summaryif: ${{ always() }}run: |+ ARTIFACT_DIR="$RUNNER_TEMP/prefect-benchmark"SUMMARY_PATH="$ARTIFACT_DIR/summary.md"if [[ -f "$SUMMARY_PATH" ]]; thencat "$SUMMARY_PATH" >> "$GITHUB_STEP_SUMMARY"diff --git a/scripts/prefect_benchmark_run.py b/scripts/prefect_benchmark_run.pyindex 08aa3833e..1c359fcb0 100644--- a/scripts/prefect_benchmark_run.py+++ b/scripts/prefect_benchmark_run.py@@ -14,6 +14,9 @@ from pathlib import Pathfrom typing import Any, Callable, Protocol+type LogSink = Callable[[str], None]++class PoolName(str, Enum):MACOS_A18_PRO = "macos-a18-pro-pool"MACOS_M1 = "macos-m1-pool"@@ -63,6 +66,8 @@ POOL_DEPLOYMENTS: dict[PoolName, str] = {PoolName.MACOS_M4_PRO: "run-benchmark-worker-flow/run-benchmark-m4-pro",}+PREFECT_LOG_PAGE_SIZE = 200+@dataclass(frozen=True)class FlowRun:@@ -273,6 +278,7 @@ def execute_benchmark_run(sleep: Callable[[float], None] = time.sleep,actor: str | None = None,triggering_actor: str | None = None,+ log_sink: LogSink | None = None,) -> BenchmarkRunResult:deployment_name = deployment_name_for_pool(pool)deployment_id = client.read_deployment_id(deployment_name)@@ -283,6 +289,7 @@ def execute_benchmark_run(poll_interval_seconds=poll_interval_seconds,timeout_seconds=timeout_seconds,sleep=sleep,+ log_sink=log_sink,)logs = read_all_logs(client, created_run.id)failure_reason = None@@ -317,26 +324,78 @@ def wait_for_terminal_flow_run(poll_interval_seconds: int,timeout_seconds: int,sleep: Callable[[float], None],+ log_sink: LogSink | None = None,) -> FlowRun:current_run = initial_run+ next_log_offset = stream_new_logs(+ client=client,+ flow_run=current_run,+ next_log_offset=0,+ log_sink=log_sink,+ )deadline = time.monotonic() + timeout_secondswhile not current_run.state_type.is_terminal and time.monotonic() < deadline:if poll_interval_seconds > 0:sleep(poll_interval_seconds)current_run = client.read_flow_run(current_run.id)+ next_log_offset = stream_new_logs(+ client=client,+ flow_run=current_run,+ next_log_offset=next_log_offset,+ log_sink=log_sink,+ )+ stream_new_logs(+ client=client,+ flow_run=current_run,+ next_log_offset=next_log_offset,+ log_sink=log_sink,+ )return current_run+def stream_new_logs(+ client: PrefectClientProtocol,+ flow_run: FlowRun,+ next_log_offset: int,+ log_sink: LogSink | None,+) -> int:+ if log_sink is None:+ return next_log_offset++ new_logs, updated_log_offset = read_logs_since_offset(+ client=client,+ flow_run_id=flow_run.id,+ offset=next_log_offset,+ )+ for log in new_logs:+ log_sink(format_log_line(flow_run, log))+ return updated_log_offset+++def read_logs_since_offset(+ client: PrefectClientProtocol,+ flow_run_id: str,+ offset: int,+) -> tuple[tuple[PrefectLog, ...], int]:+ next_log_offset = offset+ new_logs: list[PrefectLog] = []+ while True:+ page = client.read_logs_page(flow_run_id, limit=PREFECT_LOG_PAGE_SIZE, offset=next_log_offset)+ new_logs.extend(page)+ next_log_offset += len(page)+ if len(page) < PREFECT_LOG_PAGE_SIZE:+ return tuple(new_logs), next_log_offset++def read_all_logs(client: PrefectClientProtocol, flow_run_id: str) -> tuple[PrefectLog, ...]:offset = 0- page_size = 200all_logs: list[PrefectLog] = []while True:- page = client.read_logs_page(flow_run_id, limit=page_size, offset=offset)+ page = client.read_logs_page(flow_run_id, limit=PREFECT_LOG_PAGE_SIZE, offset=offset)all_logs.extend(page)- if len(page) < page_size:+ if len(page) < PREFECT_LOG_PAGE_SIZE:return tuple(all_logs)- offset += page_size+ offset += PREFECT_LOG_PAGE_SIZEdef write_run_artifacts(output_dir: Path | str, result: BenchmarkRunResult) -> None:@@ -383,14 +442,19 @@ def build_summary_markdown(result: BenchmarkRunResult) -> str:return "\n".join(lines) + "\n"+def format_log_line(flow_run: FlowRun, log: PrefectLog) -> str:+ return f"{log.timestamp} | {logging.getLevelName(log.level):7s} | Flow run {flow_run.name!r} - {log.message}"++def format_logs(flow_run: FlowRun, logs: tuple[PrefectLog, ...]) -> str:- lines = [- f"{log.timestamp} | {logging.getLevelName(log.level):7s} | Flow run {flow_run.name!r} - {log.message}"- for log in logs- ]+ lines = [format_log_line(flow_run, log) for log in logs]return "\n".join(lines) + ("\n" if lines else "")+def print_live_log_line(line: str) -> None:+ print(line, flush=True)++def parse_cli_args(argv: list[str] | None = None) -> CliConfig:parser = argparse.ArgumentParser(description="Trigger one Prefect benchmark worker deployment and fetch its logs.")parser.add_argument("--api-url", required=True, help="Prefect API URL")@@ -453,6 +517,7 @@ def main(argv: list[str] | None = None) -> int:timeout_seconds=config.timeout_seconds,actor=config.actor,triggering_actor=config.triggering_actor,+ log_sink=print_live_log_line,)write_run_artifacts(config.output_dir, result)print(build_summary_markdown(result))