Issue #1373: feat(sandbox): add multi-platform support through --platform CLI argument (#1629)

* fixup, add multi-platform support with GOARCH detection and `TARGETPLATFORM` env in gha_pr_changed_files.py

Author: jiridanek

Makefile

@@ -71,7 +71,7 @@ define build_image
 
 	$(info # Building $(IMAGE_NAME) image...)
 
-	$(ROOT_DIR)/scripts/sandbox.py --dockerfile '$(2)' -- \
+	$(ROOT_DIR)/scripts/sandbox.py --dockerfile '$(2)' --platform '$(BUILD_ARCH)' -- \
 		$(CONTAINER_ENGINE) build $(CONTAINER_BUILD_CACHE_ARGS) --platform=$(BUILD_ARCH) --label release=$(RELEASE) --tag $(IMAGE_NAME) --file '$(2)' $(BUILD_ARGS) {}\;
 endef
 

ci/cached-builds/gha_pr_changed_files.py

@@ -3,10 +3,12 @@
 import logging
 import os
 import pathlib
+import platform
 import re
 import shutil
 import subprocess
 import unittest
+from typing import Literal
 
 PROJECT_ROOT = pathlib.Path(__file__).parent.parent.parent.resolve()
 MAKE = shutil.which("gmake") or shutil.which("make")
@@ -87,7 +89,13 @@ def should_build_target(changed_files: list[str], target_directory: str) -> str:
     dockerfiles = find_dockerfiles(target_directory)
     for dockerfile in dockerfiles:
         stdout = subprocess.check_output(
-            [PROJECT_ROOT / "bin/buildinputs", target_directory + "/" + dockerfile], text=True, cwd=PROJECT_ROOT
+            [PROJECT_ROOT / "bin/buildinputs", target_directory + "/" + dockerfile],
+            env={
+                "TARGETPLATFORM": f"linux/{get_go_arch()}",
+                **os.environ,
+            },  # TODO(jdanek): still not ideal for qemu-user
+            text=True,
+            cwd=PROJECT_ROOT,
         )
         logging.debug(f"{target_directory=} {dockerfile=} {stdout=}")
         if stdout == "\n":
@@ -116,6 +124,24 @@ def filter_out_unchanged(targets: list[str], changed_files: list[str]) -> list[s
     return changed
 
 
+def get_go_arch() -> Literal["amd64", "arm64", "ppc64le", "s390x"]:
+    arch = os.environ.get("GOARCH")
+    if arch is not None:
+        return arch
+    match platform.machine().lower():
+        case "x86_64" | "amd64":
+            arch = "amd64"
+        case "aarch64" | "arm64":
+            arch = "arm64"
+        case "ppc64le":
+            arch = "ppc64le"
+        case "s390x":
+            arch = "s390x"
+        case _:
+            raise Exception(f"Unknown machine architecture: {platform.machine()}")
+    return arch
+
+
 class SelfTests(unittest.TestCase):
     def test_list_changed_files(self):
         """This is PR #556 in opendatahub-io/notebooks"""

scripts/buildinputs/buildinputs.go

@@ -3,12 +3,26 @@ package main
 import (
 	"flag"
 	"fmt"
+	"os"
+	"strings"
 )
 
 func main() {
+	targetPlatform := os.Getenv("TARGETPLATFORM")
+	platformFields := strings.Split(targetPlatform, "/")
+	if len(platformFields) != 2 {
+		panic(fmt.Sprint(targetPlatform, "format is invalid, should be os/arch"))
+	}
+	targetOs := platformFields[0]
+	targetArch := platformFields[1]
+
+	if targetOs != "linux" {
+		panic(fmt.Sprint(targetOs, "not supported"))
+	}
+
 	flag.Parse()
 	for _, dockerfile := range flag.Args() {
-		deps := getDockerfileDeps(dockerfile)
+		deps := getDockerfileDeps(dockerfile, targetArch)
 		fmt.Println(deps)
 	}
 }

scripts/buildinputs/buildinputs_test.go

@@ -30,7 +30,7 @@ func TestParseAllDockerfiles(t *testing.T) {
 
 	for _, dockerfile := range dockerfiles {
 		t.Run(dockerfile, func(t *testing.T) {
-			result := getDockerfileDeps(dockerfile)
+			result := getDockerfileDeps(dockerfile, "amd64")
 			if result == "" {
 				// no deps in the dockerfile
 				return

scripts/buildinputs/dockerfile.go

@@ -6,7 +6,6 @@ import (
 	"fmt"
 	"log"
 	"os"
-	"runtime"
 	"strings"
 
 	"github.com/containerd/platforms"
@@ -21,7 +20,7 @@ import (
 	"github.com/pkg/errors"
 )
 
-func getDockerfileDeps(dockerfile string) string {
+func getDockerfileDeps(dockerfile string, targetArch string) string {
 	ctx := context.Background()
 	data := noErr2(os.ReadFile(dockerfile))
 
@@ -32,11 +31,11 @@ func getDockerfileDeps(dockerfile string) string {
 			// random digest value
 			digest:   "sha256:a1c7d58d98df3f9a67eda799200655b923ebc7a41cad1d9bb52723ae1c81ad17",
 			dir:      "/",
-			platform: "linux/" + runtime.GOARCH,
+			platform: "linux/" + targetArch,
 		},
 		Config: dockerui.Config{
 			BuildArgs:      map[string]string{"BASE_IMAGE": "fake-image"},
-			BuildPlatforms: []ocispecs.Platform{{OS: "linux", Architecture: runtime.GOARCH}},
+			BuildPlatforms: []ocispecs.Platform{{OS: "linux", Architecture: targetArch}},
 		},
 		Warn: func(rulename, description, url, fmtmsg string, location []parser.Range) {
 			log.Printf(rulename, description, url, fmtmsg, location)

scripts/sandbox.py

@@ -2,13 +2,14 @@
 
 import argparse
 import logging
+import os
 import pathlib
 import shutil
 import subprocess
 import sys
 import tempfile
 import json
-from typing import cast
+from typing import cast, Literal
 
 ROOT_DIR = pathlib.Path(__file__).parent.parent
 MAKE = shutil.which("gmake") or shutil.which("make")
@@ -19,12 +20,16 @@
 
 class Args(argparse.Namespace):
     dockerfile: pathlib.Path
+    platform: Literal["linux/amd64", "linux/arm64", "linux/s390x", "linux/ppc64le"]
     remaining: list[str]
 
 
 def main() -> int:
     p = argparse.ArgumentParser(allow_abbrev=False)
     p.add_argument("--dockerfile", type=pathlib.Path, required=True)
+    p.add_argument("--platform", type=str,
+                   choices=["linux/amd64", "linux/arm64", "linux/s390x", "linux/ppc64le"],
+                   required=True)
     p.add_argument('remaining', nargs=argparse.REMAINDER)
 
     args = cast(Args, p.parse_args())
@@ -38,7 +43,7 @@ def main() -> int:
         print("must give a `{};` parameter that will be replaced with new build context")
         return 1
 
-    prereqs = buildinputs(dockerfile=args.dockerfile)
+    prereqs = buildinputs(dockerfile=args.dockerfile, platform=args.platform)
 
     with tempfile.TemporaryDirectory(delete=True) as tmpdir:
         setup_sandbox(prereqs, pathlib.Path(tmpdir))
@@ -52,11 +57,15 @@ def main() -> int:
     return 0
 
 
-def buildinputs(dockerfile: pathlib.Path | str) -> list[pathlib.Path]:
+def buildinputs(
+        dockerfile: pathlib.Path | str,
+        platform: Literal["linux/amd64", "linux/arm64", "linux/s390x", "linux/ppc64le"] = "linux/amd64"
+) -> list[pathlib.Path]:
     if not (ROOT_DIR / "bin/buildinputs").exists():
         subprocess.check_call([MAKE, "bin/buildinputs"], cwd=ROOT_DIR)
     stdout = subprocess.check_output([ROOT_DIR / "bin/buildinputs", str(dockerfile)],
-                                     text=True, cwd=ROOT_DIR)
+                                     text=True, cwd=ROOT_DIR,
+                                     env={"TARGETPLATFORM": platform, **os.environ})
     prereqs = [pathlib.Path(file) for file in json.loads(stdout)] if stdout != "\n" else []
     print(f"{prereqs=}")
     return prereqs