Remove build-binary

Author: lukekim

Cargo.lock

@@ -35,7 +35,6 @@ exclude = ["*.proto"]
 rustc-args = ["--cfg", "docsrs"]
 
 [features]
-build-binary = ["aws-config", "aws-credential-types", "clap", "object_store"]
 docsrs = []
 # Used for testing ONLY: causes all values to hash to the same value (test for collisions)
 force_hash_collisions = ["datafusion/force_hash_collisions"]
@@ -45,11 +44,10 @@ vortex = ["vortex-array", "vortex-buffer", "vortex-dtype", "vortex-error", "vort
 [dependencies]
 arrow-flight = { workspace = true }
 async-trait = { workspace = true }
-aws-config = { version = "1.6.0", optional = true }
-aws-credential-types = { version = "1.2.0", optional = true }
+aws-config = { version = "1.6.0" }
+aws-credential-types = { version = "1.2.0" }
 bytes = { workspace = true }
 chrono = { version = "0.4", default-features = false }
-clap = { workspace = true, optional = true }
 dashmap = "6"
 datafusion = { workspace = true }
 datafusion-proto = { workspace = true }
@@ -58,7 +56,7 @@ futures = { workspace = true }
 itertools = "0.14"
 log = { workspace = true }
 md-5 = { version = "^0.10.0" }
-object_store = { workspace = true, features = ["aws", "azure", "http"], optional = true }
+object_store = { workspace = true, features = ["aws", "azure", "http"] }
 parking_lot = { workspace = true }
 prost = { workspace = true }
 prost-types = { workspace = true }

ballista/core/Cargo.toml

@@ -468,7 +468,6 @@ impl datafusion::config::ConfigExtension for BallistaConfig {
 /// Ballista supports both push-based and pull-based task scheduling.
 /// It is recommended that you try both to determine which is the best for your use case.
 #[derive(Clone, Copy, Debug, serde::Deserialize, Default)]
-#[cfg_attr(feature = "build-binary", derive(clap::ValueEnum))]
 pub enum TaskSchedulingPolicy {
     /// Pull-based scheduling works in a similar way to Apache Spark
     #[default]
@@ -485,18 +484,23 @@ impl Display for TaskSchedulingPolicy {
     }
 }
 
-#[cfg(feature = "build-binary")]
 impl std::str::FromStr for TaskSchedulingPolicy {
     type Err = String;
 
     fn from_str(s: &str) -> std::result::Result<Self, Self::Err> {
-        clap::ValueEnum::from_str(s, true)
+        match s.to_lowercase().as_str() {
+            "pull-staged" | "pullstaged" => Ok(TaskSchedulingPolicy::PullStaged),
+            "push-staged" | "pushstaged" => Ok(TaskSchedulingPolicy::PushStaged),
+            _ => Err(format!(
+                "Invalid scheduling policy '{}'. Valid options: 'pull-staged', 'push-staged'",
+                s
+            )),
+        }
     }
 }
 
 /// Configures the log file rotation policy.
 #[derive(Clone, Copy, Debug, serde::Deserialize, Default)]
-#[cfg_attr(feature = "build-binary", derive(clap::ValueEnum))]
 pub enum LogRotationPolicy {
     /// Rotate log files every minute.
     Minutely,
@@ -520,12 +524,20 @@ impl Display for LogRotationPolicy {
     }
 }
 
-#[cfg(feature = "build-binary")]
 impl std::str::FromStr for LogRotationPolicy {
     type Err = String;
 
     fn from_str(s: &str) -> std::result::Result<Self, Self::Err> {
-        clap::ValueEnum::from_str(s, true)
+        match s.to_lowercase().as_str() {
+            "minutely" => Ok(LogRotationPolicy::Minutely),
+            "hourly" => Ok(LogRotationPolicy::Hourly),
+            "daily" => Ok(LogRotationPolicy::Daily),
+            "never" => Ok(LogRotationPolicy::Never),
+            _ => Err(format!(
+                "Invalid rotation policy '{}'. Valid options: 'minutely', 'hourly', 'daily', 'never'",
+                s
+            )),
+        }
     }
 }
 

ballista/core/src/config.rs

@@ -28,13 +28,9 @@ use std::result;
 use std::sync::Arc;
 use std::task::{Context, Poll};
 
-#[cfg(feature = "build-binary")]
 use object_store::ObjectStore;
-#[cfg(feature = "build-binary")]
 use object_store::aws::AmazonS3Builder;
-#[cfg(feature = "build-binary")]
 use object_store::azure::MicrosoftAzureBuilder;
-#[cfg(feature = "build-binary")]
 use url::Url;
 
 use crate::client::BallistaClient;
@@ -841,7 +837,6 @@ fn check_is_object_store_location(location: &PartitionLocation) -> bool {
     path.starts_with("s3://") || path.starts_with("abfs://") || path.starts_with("az://")
 }
 
-#[cfg(feature = "build-binary")]
 async fn fetch_partition_object_store(
     location: &PartitionLocation,
 ) -> result::Result<SendableRecordBatchStream, BallistaError> {
@@ -877,16 +872,6 @@ async fn fetch_partition_object_store(
     Ok(Box::pin(RecordBatchStreamAdapter::new(schema, stream)))
 }
 
-#[cfg(not(feature = "build-binary"))]
-async fn fetch_partition_object_store(
-    _location: &PartitionLocation,
-) -> result::Result<SendableRecordBatchStream, BallistaError> {
-    Err(BallistaError::NotImplemented(
-        "Object store support requires 'build-binary' feature".to_string(),
-    ))
-}
-
-#[cfg(feature = "build-binary")]
 async fn fetch_partition_object_store_inner(
     path: &str,
 ) -> result::Result<Vec<RecordBatch>, BallistaError> {

ballista/core/src/execution_plans/shuffle_reader.rs

@@ -45,7 +45,6 @@ pub mod event_loop;
 pub mod execution_plans;
 /// Extension traits and utilities for DataFusion integration.
 pub mod extension;
-#[cfg(feature = "build-binary")]
 /// Object store configuration and utilities for distributed file access.
 pub mod object_store;
 /// Query planning utilities for distributed execution.
@@ -57,7 +56,6 @@ pub mod remote_catalog;
 /// Serialization and deserialization for Ballista messages and plans.
 pub mod serde;
 /// Shuffle storage abstraction for local and object store backends.
-#[cfg(feature = "build-binary")]
 pub mod shuffle_storage;
 /// General utility functions for Ballista operations.
 pub mod utils;

ballista/core/src/lib.rs

@@ -33,7 +33,7 @@ path = "src/bin/main.rs"
 required-features = ["build-binary"]
 
 [features]
-build-binary = ["clap", "tracing-subscriber", "tracing-appender", "tracing", "ballista-core/build-binary"]
+build-binary = ["clap", "tracing-subscriber", "tracing-appender", "tracing"]
 default = ["build-binary", "mimalloc"]
 vortex = ["ballista-core/vortex", "vortex-array", "vortex-ipc"]
 

ballista/executor/Cargo.toml

@@ -33,7 +33,7 @@ path = "src/bin/main.rs"
 required-features = ["build-binary"]
 
 [features]
-build-binary = ["clap", "tracing-subscriber", "tracing-appender", "tracing", "ballista-core/build-binary"]
+build-binary = ["clap", "tracing-subscriber", "tracing-appender", "tracing"]
 default = ["build-binary", "substrait"]
 # job info can cache stage plans, in some cases where 
 # task plans can be re-computed, cache behavior may need to be disabled.