Drive coordinator upload and reduction from the workload catalog

This commit is contained in:
Emil
2026-08-02 17:47:47 +03:00
parent 700a96a259
commit 749396da05
26 changed files with 1473 additions and 184 deletions
+57
View File
@@ -0,0 +1,57 @@
// Package reducer contains deterministic, coordinator-side result reductions.
package reducer
import (
"bytes"
"encoding/csv"
"errors"
"fmt"
"io"
)
// ReduceOrderedConcat concatenates worker partial tables in shard order into a
// single table with one header. Every partial must carry the same header as the
// first partial and rows of the same width; anything else fails the job closed.
func ReduceOrderedConcat(partials []io.Reader) ([]byte, error) {
var out bytes.Buffer
writer := csv.NewWriter(&out)
var firstHeader []string
for _, partial := range partials {
reader := csv.NewReader(partial)
header, err := reader.Read()
if err != nil {
return nil, fmt.Errorf("read partial header: %w", err)
}
if len(header) == 0 {
return nil, fmt.Errorf("partial result has an empty header")
}
if firstHeader == nil {
firstHeader = header
if err := writer.Write(header); err != nil {
return nil, err
}
} else if !equalStrings(header, firstHeader) {
return nil, fmt.Errorf("partial result has an inconsistent header")
}
for {
row, err := reader.Read()
if errors.Is(err, io.EOF) {
break
}
if err != nil {
return nil, fmt.Errorf("read partial row: %w", err)
}
if len(row) != len(header) {
return nil, fmt.Errorf("partial result has a row with an inconsistent width")
}
if err := writer.Write(row); err != nil {
return nil, err
}
}
}
writer.Flush()
if err := writer.Error(); err != nil {
return nil, err
}
return out.Bytes(), nil
}
@@ -0,0 +1,64 @@
package reducer
import (
"io"
"strings"
"testing"
)
func TestReduceOrderedConcatJoinsPartialsInOrderWithOneHeader(t *testing.T) {
first := strings.NewReader("chembl_id,canonical_smiles\nA,CC\nB,CCC\n")
second := strings.NewReader("chembl_id,canonical_smiles\nC,CCCC\n")
output, err := ReduceOrderedConcat([]io.Reader{first, second})
if err != nil {
t.Fatal(err)
}
want := "chembl_id,canonical_smiles\nA,CC\nB,CCC\nC,CCCC\n"
if string(output) != want {
t.Fatalf("output = %q, want %q", output, want)
}
}
func TestReduceOrderedConcatIsDeterministicAcrossInputOrder(t *testing.T) {
left := strings.NewReader("id,rows\nA,1\nB,2\n")
right := strings.NewReader("id,rows\nC,3\n")
first, err := ReduceOrderedConcat([]io.Reader{left, right})
if err != nil {
t.Fatal(err)
}
left, right = strings.NewReader("id,rows\nA,1\nB,2\n"), strings.NewReader("id,rows\nC,3\n")
second, err := ReduceOrderedConcat([]io.Reader{left, right})
if err != nil {
t.Fatal(err)
}
if string(first) != string(second) {
t.Fatalf("concat is not deterministic: %q != %q", first, second)
}
}
func TestReduceOrderedConcatRejectsInconsistentHeaders(t *testing.T) {
first := strings.NewReader("a,b\n1,2\n")
second := strings.NewReader("a,c\n1,2\n")
if _, err := ReduceOrderedConcat([]io.Reader{first, second}); err == nil {
t.Fatal("inconsistent headers must fail")
}
}
func TestReduceOrderedConcatRejectsRaggedRows(t *testing.T) {
partial := strings.NewReader("a,b\n1,2,3\n")
if _, err := ReduceOrderedConcat([]io.Reader{partial}); err == nil {
t.Fatal("ragged rows must fail")
}
}
func TestReduceOrderedConcatEmptyPartials(t *testing.T) {
output, err := ReduceOrderedConcat(nil)
if err != nil {
t.Fatal(err)
}
if len(output) != 0 {
t.Fatalf("empty input must produce empty output, got %q", output)
}
}