parquet_pushdown(M1-C3): add Loader interface and LocalLoader

Introduce a small Loader interface so the service can stay oblivious
to the storage backend behind a Parquet path: tests inject a
LocalLoader backed by os.File, and the daemon will plug in a
filer-backed Loader in a later milestone.

FileHandle is the parser-friendly view: io.ReaderAt + io.Closer +
Size(). LocalLoader rejects missing files and directories up front so
service-level error paths see a clear "open failed" rather than a
later ReadAt EOF.

The Service does not yet consume Loader — that wiring lands in M1-C4
along with the predicate-less Pushdown handler that actually issues
loads.
This commit is contained in:
Chris Lu committed 2026-04-25 15:15:46 -07:00
1 parent ca42c5bbe4
commit 737cc0f330
2 files changed
+105

No files matched your search

+57
View File
@@ -0,0 +1,57 @@
package parquet_pushdown
import (
"context"
"fmt"
"io"
"os"
)
// FileHandle is what the service hands to the parser: a positional
// reader bound to a known size, with a Close so the caller can
// release any underlying resources.
type FileHandle interface {
io.ReaderAt
io.Closer
Size() int64
}
// Loader resolves a Parquet file's path to a FileHandle. The split
// between "local file" and "filer-backed" implementations lives here
// so the service stays oblivious to the storage backend, and so
// integration tests can plug in deterministic local files instead of
// spinning up a filer.
type Loader interface {
Open(ctx context.Context, path string) (FileHandle, error)
}
// LocalLoader treats `path` as a filesystem path. Used by integration
// tests and for local-development smoke runs against a Parquet file
// that is not under filer management.
type LocalLoader struct{}
func (LocalLoader) Open(_ context.Context, path string) (FileHandle, error) {
f, err := os.Open(path)
if err != nil {
return nil, fmt.Errorf("open %q: %w", path, err)
}
st, err := f.Stat()
if err != nil {
_ = f.Close()
return nil, fmt.Errorf("stat %q: %w", path, err)
}
if st.IsDir() {
_ = f.Close()
return nil, fmt.Errorf("%q is a directory", path)
}
return &localFile{f: f, size: st.Size()}, nil
}
type localFile struct {
f *os.File
size int64
}
func (l *localFile) ReadAt(p []byte, off int64) (int, error) { return l.f.ReadAt(p, off) }
func (l *localFile) Close() error { return l.f.Close() }
func (l *localFile) Size() int64 { return l.size }
+48
View File
@@ -0,0 +1,48 @@
package parquet_pushdown
import (
"context"
"os"
"path/filepath"
"testing"
)
func TestLocalLoader_Open(t *testing.T) {
dir := t.TempDir()
path := filepath.Join(dir, "a.bin")
want := []byte("hello pushdown")
if err := os.WriteFile(path, want, 0o600); err != nil {
t.Fatalf("write: %v", err)
}
h, err := LocalLoader{}.Open(context.Background(), path)
if err != nil {
t.Fatalf("open: %v", err)
}
defer h.Close()
if h.Size() != int64(len(want)) {
t.Errorf("Size = %d, want %d", h.Size(), len(want))
}
got := make([]byte, len(want))
n, err := h.ReadAt(got, 0)
if err != nil {
t.Fatalf("ReadAt: %v", err)
}
if n != len(want) || string(got) != string(want) {
t.Errorf("got %q (n=%d), want %q", got[:n], n, want)
}
}
func TestLocalLoader_RejectsMissing(t *testing.T) {
if _, err := (LocalLoader{}).Open(context.Background(), filepath.Join(t.TempDir(), "missing.bin")); err == nil {
t.Error("expected error opening missing file")
}
}
func TestLocalLoader_RejectsDirectory(t *testing.T) {
dir := t.TempDir()
if _, err := (LocalLoader{}).Open(context.Background(), dir); err == nil {
t.Error("expected error opening a directory")
}
}