Skip to content

chore: Cleanup assembly and shading #617

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
SemyonSinchenko merged 7 commits into from
Jul 2, 2025
Merged

chore: Cleanup assembly and shading #617

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
cb6d7ec
Try new method for shading
Kimahriman Jul 1, 2025
dc9b47c
Fix tests
Kimahriman Jul 1, 2025
702d0d6
Add exportJars
Kimahriman Jul 2, 2025
43d372d
Just use target dir
Kimahriman Jul 2, 2025
4fec8ec
Add comment
Kimahriman Jul 2, 2025
b7c51cd
Don't shade anything, just rename
Kimahriman Jul 2, 2025
687ea18
Don't actually need the extra project
Kimahriman Jul 2, 2025
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
45 changes: 23 additions & 22 deletions build.sbt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -116,29 +116,20 @@ lazy val root = (project in file("."))

Compile / unmanagedSourceDirectories += (Compile / baseDirectory).value / "src" / "main" / s"scala-spark-$sparkMajorVer",

// Assembly settings
assembly / test := {}, // No tests in assembly
assemblyPackageScala / assembleArtifact := false,
assembly / assemblyMergeStrategy := {
case PathList("META-INF", xs @ _*) => MergeStrategy.discard
case x if x.endsWith("module-info.class") => MergeStrategy.discard
case x =>
val oldStrategy = (assembly / assemblyMergeStrategy).value
oldStrategy(x)
},
Comment on lines -119 to -128
Copy link
Contributor Author

@Kimahriman Kimahriman Jul 2, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I removed this because I don't think there's any need to run assembly on the root project? Unless you want to keep the ability to manual build a fat JAR

Test / packageBin / publishArtifact := false,
Test / packageDoc / publishArtifact := false,
Test / packageSrc / publishArtifact := false,
Compile / packageBin / publishArtifact := true,
Compile / packageDoc / publishArtifact := true,
Compile / packageSrc / publishArtifact := true)

lazy val connect = (project in file("graphframes-connect"))
// Dedicated project for creating the shaded JAR that doesn't get published
lazy val connectAssembly = (project in file("graphframes-connect"))
.dependsOn(root)
.settings(
name := s"graphframes-connect-assembly",
moduleName := s"graphframes-connect-spark${sparkMajorVer}",
commonSetting,
name := s"graphframes-connect",
moduleName := s"${name.value}-spark${sparkMajorVer}",
Compile / unmanagedSourceDirectories += (Compile / baseDirectory).value / "src" / "main" / s"scala-spark-$sparkMajorVer",
Compile / PB.targets := Seq(PB.gens.java -> (Compile / sourceManaged).value),
Compile / PB.includePaths ++= Seq(file("src/main/protobuf")),
Expand All @@ -147,8 +138,8 @@ lazy val connect = (project in file("graphframes-connect"))
"org.apache.spark" %% "spark-connect" % sparkVer % "provided" cross CrossVersion.for3Use2_13),

// Assembly and shading
assembly / assemblyJarName := s"${moduleName.value}_${(scalaBinaryVersion).value}-${version.value}.jar",
assembly / test := {},
assemblyPackageScala / assembleArtifact := false,
assembly / assemblyShadeRules := Seq(
ShadeRule.rename("com.google.protobuf.**" -> protobufShadingPattern).inAll),
assembly / assemblyMergeStrategy := {
Expand All @@ -157,18 +148,28 @@ lazy val connect = (project in file("graphframes-connect"))
case x if x.endsWith("module-info.class") => MergeStrategy.discard
case x => MergeStrategy.first
},
assembly / assemblyExcludedJars := (Compile / fullClasspath).value.filter { className =>
className.data
.getName()
.contains("scala-library-") || className.data
.getName()
.contains("slf4j-api-")
assembly / assemblyExcludedJars := {
val cp = (assembly / fullClasspath).value
val allowedPrefixes = Set("protobuf-java")
cp.filter { f =>
!allowedPrefixes.exists(prefix => f.data.getName.startsWith(prefix))
}
},
publish / skip := false,
publish / skip := true,
Compile / packageBin := assembly.value,
Test / packageBin / publishArtifact := false,
Test / packageDoc / publishArtifact := false,
Test / packageSrc / publishArtifact := false,
Compile / packageBin / publishArtifact := true,
Compile / packageBin / publishArtifact := false,
Compile / packageDoc / publishArtifact := false,
Compile / packageSrc / publishArtifact := false)

// Publish the shaded JAR with the correct dependencies in the POM
lazy val connect = project
.dependsOn(root)
.settings(
commonSetting,
name := s"graphframes-connect",
moduleName := s"${name.value}-spark${sparkMajorVer}",
Compile / packageBin := (connectAssembly / Compile / assembly).value
)
4 changes: 2 additions & 2 deletions python/dev/build_jar.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,9 +16,9 @@ def build(spark_versions: Sequence[str] = ["3.5.5"]):
sbt_executable,
f"-Dspark.version={spark_version}",
"clean",
"assembly",
"package",
"connect/clean",
"connect/assembly"
"connect/package"
]
sbt_build = subprocess.Popen(
sbt_build_command,
Expand Down
4 changes: 2 additions & 2 deletions python/tests/conftest.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -36,7 +36,7 @@ def get_gf_jar_locations() -> Tuple[str, str]:
core_jar: Optional[str] = None
connect_jar: Optional[str] = None

for pp in core_dir.glob("graphframes-assembly-*.jar"):
for pp in core_dir.glob(f"graphframes-spark{spark_major_version}*.jar"):
assert isinstance(pp, pathlib.PosixPath) # type checking
core_jar = str(pp.absolute())

Expand All @@ -45,7 +45,7 @@ def get_gf_jar_locations() -> Tuple[str, str]:
f"Failed to find graphframes jar for Spark {spark_major_version} in {core_dir}"
)

for pp in connect_dir.glob("graphframes-connect-assembly-*.jar"):
for pp in connect_dir.glob(f"graphframes-connect-spark{spark_major_version}*.jar"):
assert isinstance(pp, pathlib.PosixPath) # type checking
connect_jar = str(pp.absolute())

Expand Down