Remove unused merger

This commit is contained in:
Dominik Werder
2021-05-08 14:28:51 +02:00
parent 9f5c7b48b4
commit 2df0be8ed2
2 changed files with 0 additions and 212 deletions

View File

@@ -1,5 +1,4 @@
use super::agg::IntoDim1F32Stream;
use super::merge::MergeDim1F32Stream;
use crate::agg::binnedt::IntoBinnedT;
use crate::agg::binnedx::IntoBinnedXBins1;
use futures_util::StreamExt;
@@ -145,74 +144,3 @@ async fn agg_x_dim_1_inner() {
.for_each(|_k| ready(()));
fut1.await;
}
#[test]
fn merge_0() {
taskrun::run(async {
merge_0_inner().await;
Ok(())
})
.unwrap();
}
async fn merge_0_inner() {
let query = netpod::AggQuerySingleChannel {
channel_config: ChannelConfig {
channel: Channel {
backend: "ks".into(),
name: "wave1".into(),
},
keyspace: 3,
time_bin_size: Nanos { ns: DAY },
array: true,
shape: Shape::Wave(17),
scalar_type: ScalarType::F64,
big_endian: true,
compression: true,
},
timebin: 0,
tb_file_count: 1,
buffer_size: 1024 * 8,
};
let range: NanoRange = err::todoval();
let streams = (0..13)
.into_iter()
.map(|k| make_test_node(k))
.map(|node| {
super::eventblobs::EventBlobsComplete::new(
range.clone(),
query.channel_config.clone(),
node.clone(),
query.buffer_size as usize,
)
.into_dim_1_f32_stream()
})
.collect();
MergeDim1F32Stream::new(streams)
.map(|_k| {
//info!("NEXT MERGED ITEM ts {:?}", k.as_ref().unwrap().tss);
})
.fold(0, |_k, _q| ready(0))
.await;
}
pub fn tmp_some_older_things() {
// I want to distinguish already in the outer part between dim-0 and dim-1 and generate
// separate code for these cases...
// That means that also the reading chain itself needs to be typed on that.
// Need to supply some event-payload converter type which has that type as Output type.
// Now the T-binning:
/*
T-aggregator must be able to produce empty-values of correct type even if we never get
a single value of input data.
Therefore, it needs the bin range definition.
How do I want to drive the system?
If I write the T-binner as a Stream, then I also need to pass it the input!
Meaning, I need to pass the Stream which produces the actual numbers from disk.
readchannel() -> Stream of timestamped byte blobs
.to_f32() -> Stream ? indirection to branch on the underlying shape
.agg_x_bins_1() -> Stream ? can I keep it at the single indirection on the top level?
*/
}

View File

@@ -1,6 +1,4 @@
use crate::agg::eventbatch::{MinMaxAvgScalarEventBatch, MinMaxAvgScalarEventBatchStreamItem};
use crate::agg::{Dim1F32Stream, Dim1F32StreamItem, ValuesDim1};
use crate::eventchunker::EventChunkerItem;
use crate::streamlog::LogItem;
use err::Error;
use futures_core::Stream;
@@ -12,144 +10,6 @@ use std::task::{Context, Poll};
#[allow(unused_imports)]
use tracing::{debug, error, info, trace, warn};
pub struct MergeDim1F32Stream<S> {
// yields Dim1F32StreamItem
inps: Vec<Dim1F32Stream<S>>,
current: Vec<CurVal>,
ixs: Vec<usize>,
errored: bool,
completed: bool,
batch: ValuesDim1,
}
impl<S> MergeDim1F32Stream<S> {
pub fn new(inps: Vec<Dim1F32Stream<S>>) -> Self {
let n = inps.len();
let mut current = vec![];
for _ in 0..n {
current.push(CurVal::None);
}
Self {
inps,
current: current,
ixs: vec![0; n],
batch: ValuesDim1::empty(),
errored: false,
completed: false,
}
}
}
impl<S> Stream for MergeDim1F32Stream<S>
where
S: Stream<Item = Result<EventChunkerItem, Error>> + Unpin,
{
type Item = Result<Dim1F32StreamItem, Error>;
fn poll_next(mut self: Pin<&mut Self>, cx: &mut Context) -> Poll<Option<Self::Item>> {
use Poll::*;
// TODO rewrite making the break the default and explicit continue.
'outer: loop {
if self.completed {
panic!("poll on complete stream");
}
if self.errored {
self.completed = true;
return Ready(None);
}
// can only run logic if all streams are either finished, errored or have some current value.
for i1 in 0..self.inps.len() {
match self.current[i1] {
CurVal::None => {
match self.inps[i1].poll_next_unpin(cx) {
Ready(Some(Ok(k))) => {
// TODO do I keep only the values as "current" or also the other kinds of items?
// Can I process the other kinds instantly?
match k {
Dim1F32StreamItem::Values(vals) => {
self.current[i1] = CurVal::Val(vals);
}
Dim1F32StreamItem::RangeComplete => {
todo!();
}
Dim1F32StreamItem::EventDataReadStats(_stats) => {
todo!();
}
}
}
Ready(Some(Err(e))) => {
self.current[i1] = CurVal::Err(Error::with_msg(format!(
"MergeDim1F32Stream error from upstream {:?}",
e
)));
return Ready(Some(Err(e)));
}
Ready(None) => {
self.current[i1] = CurVal::Finish;
}
Pending => {
// TODO is this behaviour correct?
return Pending;
}
}
}
_ => (),
}
}
let mut lowest_ix = usize::MAX;
let mut lowest_ts = u64::MAX;
for i1 in 0..self.inps.len() {
match &self.current[i1] {
CurVal::Finish => {}
CurVal::Val(val) => {
let u = self.ixs[i1];
if u >= val.tss.len() {
self.ixs[i1] = 0;
self.current[i1] = CurVal::None;
continue 'outer;
} else {
let ts = val.tss[u];
if ts < lowest_ts {
lowest_ix = i1;
lowest_ts = ts;
}
}
}
_ => panic!(),
}
}
if lowest_ix == usize::MAX {
// TODO all inputs in finished state
break Ready(None);
} else {
//trace!("decided on next lowest ts {} ix {}", lowest_ts, lowest_ix);
self.batch.tss.push(lowest_ts);
let rix = self.ixs[lowest_ix];
match &mut self.current[lowest_ix] {
CurVal::Val(ref mut k) => {
let k = std::mem::replace(&mut k.values[rix], vec![]);
self.batch.values.push(k);
}
_ => panic!(),
}
self.ixs[lowest_ix] += 1;
}
if self.batch.tss.len() >= 64 {
let k = std::mem::replace(&mut self.batch, ValuesDim1::empty());
let ret = Dim1F32StreamItem::Values(k);
break Ready(Some(Ok(ret)));
}
}
}
}
enum CurVal {
None,
Finish,
Err(Error),
Val(ValuesDim1),
}
pub struct MergedMinMaxAvgScalarStream<S>
where
S: Stream<Item = Result<MinMaxAvgScalarEventBatchStreamItem, Error>>,