Skip to content
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.

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

add toolbox for ar compress and uncompress #529

Merged
merged 5 commits into from
Jun 30, 2023
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
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
1 change: 1 addition & 0 deletions .gitignore
Original file line number Diff line number Diff line change
Expand Up @@ -40,3 +40,4 @@ tools/state_db
tools/index_doc_db
tools/index_meta_db
bridge
sdk/dist
Binary file added src/node/resources/test/37829_37968.gz.parquet
Binary file not shown.
325 changes: 325 additions & 0 deletions src/node/src/ar_toolbox.rs
Original file line number Diff line number Diff line change
@@ -0,0 +1,325 @@
use arrow::array::{
ArrayRef, BinaryArray, BinaryBuilder, StringArray, StringBuilder, UInt32Array, UInt32Builder,
UInt64Array, UInt64Builder,
};
use arrow::datatypes::{DataType, Field, Schema, SchemaRef};
use arrow::record_batch::RecordBatch;
use db3_error::{DB3Error, Result};
use db3_proto::db3_mutation_v2_proto::{MutationBody, MutationHeader};
use db3_storage::ar_fs::{ArFileSystem, ArFileSystemConfig};
use parquet::arrow::arrow_reader::ParquetRecordBatchReaderBuilder;
use parquet::arrow::ArrowWriter;
use parquet::basic::{Compression, GzipLevel};
use parquet::file::properties::WriterProperties;
use std::fs::File;
use std::path::Path;
use std::sync::atomic::{AtomicU64, Ordering};
use std::sync::Arc;
use tempdir::TempDir;
use tracing::info;
pub struct ArToolBox {
pub network_id: Arc<AtomicU64>,
pub schema: SchemaRef,
pub ar_filesystem: ArFileSystem,
pub temp_data_path: String,
}

impl ArToolBox {
pub fn new(
key_root_path: String,
arweave_url: String,
temp_data_path: String,
network_id: Arc<AtomicU64>,
) -> Result<Self> {
let ar_fs_config = ArFileSystemConfig {
key_root_path,
arweave_url,
};
let ar_filesystem = ArFileSystem::new(ar_fs_config)?;
let schema = Arc::new(Schema::new(vec![
Field::new("payload", DataType::Binary, true),
Field::new("signature", DataType::Utf8, true),
Field::new("block", DataType::UInt64, true),
Field::new("order", DataType::UInt32, true),
]));

Ok(Self {
network_id,
schema,
ar_filesystem,
temp_data_path,
})
}
pub async fn get_ar_account(&self) -> Result<(String, String)> {
let addr = self.ar_filesystem.get_address();
let balance = self.ar_filesystem.get_balance().await?;
Ok((addr, balance.to_string()))
}

pub async fn compress_and_upload_record_batch(
&self,
tx: String,
last_end_block: u64,
current_block: u64,
recordbatch: &RecordBatch,
) -> Result<(String, u64, u64, u64)> {
let tmp_dir = TempDir::new_in(&self.temp_data_path, "compression")
.map_err(|e| DB3Error::RollupError(format!("{e}")))?;
let file_path = tmp_dir.path().join("rollup.gz.parquet");
let (num_rows, size) = Self::dump_recordbatch(&file_path, recordbatch)?;
let filename = format!("{}_{}.gz.parquet", last_end_block, current_block);
//TODO add tx status confirmation
let (id, reward) = self
.ar_filesystem
.upload_file(
&file_path,
tx.as_str(),
last_end_block,
current_block,
self.network_id.load(Ordering::Relaxed),
filename.as_str(),
)
.await?;
Ok((id, reward, num_rows, size))
}

/// Compress recordbatch to parquet file
pub fn dump_recordbatch(path: &Path, recordbatch: &RecordBatch) -> Result<(u64, u64)> {
let properties = WriterProperties::builder()
.set_compression(Compression::GZIP(GzipLevel::default()))
.build();
let fd = File::create(path).map_err(|e| DB3Error::RollupError(format!("{e}")))?;

let mut writer = ArrowWriter::try_new(fd, recordbatch.schema(), Some(properties))
.map_err(|e| DB3Error::RollupError(format!("{e}")))?;
writer
.write(recordbatch)
.map_err(|e| DB3Error::RollupError(format!("{e}")))?;
let meta = writer
.close()
.map_err(|e| DB3Error::RollupError(format!("{e}")))?;
let metadata =
std::fs::metadata(path).map_err(|e| DB3Error::RollupError(format!("{e}")))?;
Ok((meta.num_rows as u64, metadata.len()))
}

/// Parse recordbatch from parquet file
pub fn parse_gzip_file(path: &Path) -> Result<Vec<RecordBatch>> {
let fd = File::open(path).map_err(|e| DB3Error::RollupError(format!("{e}")))?;
// Create a sync parquet reader with batch_size.
// batch_size is the number of rows to read up to buffer once from pages, defaults to 1024
let parquet_reader = ParquetRecordBatchReaderBuilder::try_new(fd)
.map_err(|e| DB3Error::RollupError(format!("{e}")))?
.with_batch_size(8192)
.build()
.map_err(|e| DB3Error::RollupError(format!("{e}")))?;

let mut batches = Vec::new();

for batch in parquet_reader {
let each = batch.map_err(|e| DB3Error::RollupError(format!("{e}")))?;
batches.push(each);
}
Ok(batches)
}

/// Parse mutation body, block and order from recordbatch
pub fn convert_recordbatch_to_mutation(
record_batch: &RecordBatch,
) -> Result<Vec<(MutationBody, u64, u32)>> {
let mut mutations = Vec::new();
let payloads = record_batch
.column_by_name("payload")
.unwrap()
Copy link
Contributor

@imotai imotai Jun 30, 2023

Choose a reason for hiding this comment

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

unwrap will panic if some error.

.as_any()
.downcast_ref::<BinaryArray>()
.unwrap();
let signatures = record_batch
.column_by_name("signature")
.unwrap()
.as_any()
.downcast_ref::<StringArray>()
.unwrap();
let blocks = record_batch
.column_by_name("block")
.unwrap()
.as_any()
.downcast_ref::<UInt64Array>()
.unwrap();
let orders = record_batch
.column_by_name("order")
.unwrap()
.as_any()
.downcast_ref::<UInt32Array>()
.unwrap();

for i in 0..record_batch.num_rows() {
let payload = payloads.value(i);
let signature = signatures.value(i);
let block = blocks.value(i);
let order = orders.value(i);
let mutation = MutationBody {
payload: payload.to_vec(),
signature: signature.to_string(),
};
mutations.push((mutation, block, order));
}
Ok(mutations)
}

/// convert mutation to recordbatch
/// encode mutation body, block and order to recordbatch
pub fn convert_mutations_to_recordbatch(
&self,
mutations: &[(MutationHeader, MutationBody)],
) -> Result<RecordBatch> {
//TODO limit the memory usage
let mut payload_builder = BinaryBuilder::new();
let mut signature_builder = StringBuilder::new();
let mut block_builder = UInt64Builder::new();
let mut order_builder = UInt32Builder::new();
for (header, body) in mutations {
let body_ref: &[u8] = &body.payload;
payload_builder.append_value(body_ref);
signature_builder.append_value(body.signature.as_str());
block_builder.append_value(header.block_id);
order_builder.append_value(header.order_id);
}
let array_refs: Vec<ArrayRef> = vec![
Arc::new(payload_builder.finish()),
Arc::new(signature_builder.finish()),
Arc::new(block_builder.finish()),
Arc::new(order_builder.finish()),
];
let record_batch = RecordBatch::try_new(self.schema.clone(), array_refs)
.map_err(|e| DB3Error::RollupError(format!("{e}")))?;
info!(
"convert {} into recordbatch with memory {}",
mutations.len(),
record_batch.get_array_memory_size()
);
Ok(record_batch)
}
}

#[cfg(test)]
mod tests {
use super::*;
use arrow::array::{Array, AsArray, BinaryArray, StringArray, UInt32Array, UInt64Array};
use arrow::datatypes::{BinaryType, DataType, Field, Schema};
use std::path::PathBuf;
#[test]
fn it_works() {}

fn mock_batch_record() -> RecordBatch {
let schema = Arc::new(Schema::new(vec![
Field::new("payload", DataType::Binary, true),
Field::new("signature", DataType::Utf8, true),
Field::new("block", DataType::UInt64, true),
Field::new("order", DataType::UInt32, true),
]));
let mut payload_builder = BinaryBuilder::new();
let mut signature_builder = StringBuilder::new();
let mut block_builder = UInt64Builder::new();
let mut order_builder = UInt32Builder::new();
for block in 0..10 {
let body_ref: &[u8] = "this is a payload sample".as_bytes();
payload_builder.append_value(body_ref);
signature_builder.append_value("0x1234567890");
block_builder.append_value(block);
order_builder.append_value((block * 10) as u32);
}
let array_refs: Vec<ArrayRef> = vec![
Arc::new(payload_builder.finish()),
Arc::new(signature_builder.finish()),
Arc::new(block_builder.finish()),
Arc::new(order_builder.finish()),
];
RecordBatch::try_new(schema.clone(), array_refs).unwrap()
}
#[test]
fn dump_recordbatch_ut() {
let tmp_dir_path = TempDir::new("dump_recordbatch_ut").expect("create temp dir");

let record_batch = mock_batch_record();
let (num_rows, size) = ArToolBox::dump_recordbatch(
Path::new(tmp_dir_path.path().join("test.parquet").to_str().unwrap()),
&record_batch,
)
.unwrap();
assert_eq!(num_rows, 10);
assert_eq!(size, 1862);
}
#[test]
fn parse_gzip_file_ut() {
let tmp_dir_path = TempDir::new("dump_recordbatch_ut").expect("create temp dir");

let parquet_file = tmp_dir_path.path().join("test.parquet");
let record_batch = mock_batch_record();
let (num_rows, size) = ArToolBox::dump_recordbatch(&parquet_file, &record_batch).unwrap();
assert_eq!(num_rows, 10);
assert_eq!(size, 1862);
let res = ArToolBox::parse_gzip_file(parquet_file.as_path()).unwrap();
assert_eq!(res.len(), 1);
let rec = res[0].clone();
println!("schema: {}", rec.schema());
assert!(rec.num_columns() == 4);
assert_eq!(rec.num_rows(), 10);
let payloads = rec
.column_by_name("payload")
.unwrap()
.as_any()
.downcast_ref::<BinaryArray>()
.unwrap();
assert_eq!(payloads.len(), 10);
assert_eq!(payloads.value(5), "this is a payload sample".as_bytes());

let signatures = rec
.column_by_name("signature")
.unwrap()
.as_any()
.downcast_ref::<StringArray>()
.unwrap();
assert_eq!(signatures.len(), 10);
assert_eq!(signatures.value(5), "0x1234567890");

let blocks = rec
.column_by_name("block")
.unwrap()
.as_any()
.downcast_ref::<UInt64Array>()
.unwrap();
assert_eq!(blocks.len(), 10);
assert_eq!(blocks.value(5), 5);

let orders = rec
.column_by_name("order")
.unwrap()
.as_any()
.downcast_ref::<UInt32Array>()
.unwrap();
assert_eq!(orders.len(), 10);
assert_eq!(orders.value(5), 50);
}

#[test]
fn parse_sample_ar_parquet_ut() {
let mut path = PathBuf::from(env!("CARGO_MANIFEST_DIR"));
path.push("resources/test/37829_37968.gz.parquet");

let res = ArToolBox::parse_gzip_file(path.as_path()).unwrap();
assert_eq!(res.len(), 1);
let rec = res[0].clone();
println!("schema: {}", rec.schema());
println!("num_rows: {}", rec.num_rows());
assert_eq!(rec.num_columns(), 4);
assert_eq!(rec.num_rows(), 204);

let mutations = ArToolBox::convert_recordbatch_to_mutation(&rec).unwrap();
assert_eq!(mutations.len(), 204);
let (mutation, block, order) = mutations[0].clone();
assert_eq!(block, 37829);
assert_eq!(order, 1);
assert_eq!(mutation.signature, "0xf6afe1165ae87fa09375eabccdedc61f3e5af4ed1e5c6456f1b63d397862252667e1f13f0f076f30609754f787c80135c52f7c249e95c9b8fab1b9ed27846c1b1c");
}
}
1 change: 1 addition & 0 deletions src/node/src/lib.rs
Original file line number Diff line number Diff line change
Expand Up @@ -16,6 +16,7 @@
//
#![feature(iter_intersperse)]
pub mod abci_impl;
pub mod ar_toolbox;
pub mod auth_storage;
pub mod command;
pub mod context;
Expand Down
Loading