archivist: initial Go implementation of the written user stories (US-01..US-23)

archivist/internal/storage/s3.go

@@ -0,0 +1,123 @@
+// Package storage wraps the MinIO S3 client (US-01, US-10..US-17).
+package storage
+
+import (
+	"context"
+	"crypto/sha256"
+	"encoding/hex"
+	"fmt"
+	"io"
+	"regexp"
+	"strconv"
+	"strings"
+
+	"github.com/minio/minio-go/v7"
+	"github.com/minio/minio-go/v7/pkg/credentials"
+)
+
+type Client struct {
+	mc     *minio.Client
+	Bucket string
+}
+
+func New(endpoint, accessKey, secretKey, bucket string, useSSL bool) (*Client, error) {
+	mc, err := minio.New(endpoint, &minio.Options{
+		Creds:  credentials.NewStaticV4(accessKey, secretKey, ""),
+		Secure: useSSL,
+	})
+	if err != nil {
+		return nil, err
+	}
+	c := &Client{mc: mc, Bucket: bucket}
+
+	ctx := context.Background()
+	exists, err := mc.BucketExists(ctx, bucket)
+	if err != nil {
+		return nil, fmt.Errorf("bucket check: %w", err)
+	}
+	if !exists {
+		return nil, fmt.Errorf("bucket %q does not exist", bucket)
+	}
+	// US-01: the bucket must be versioned (delete markers keep data recoverable).
+	vc, err := mc.GetBucketVersioning(ctx, bucket)
+	if err != nil {
+		return nil, fmt.Errorf("versioning check: %w", err)
+	}
+	if vc.Status != "Enabled" {
+		return nil, fmt.Errorf("bucket %q is not versioned (status %q)", bucket, vc.Status)
+	}
+	return c, nil
+}
+
+// InboxObject is one object listed under inbox/.
+type InboxObject struct {
+	Key  string
+	Size int64
+	ETag string
+}
+
+// ListInbox lists the objects under the inbox/ prefix (US-10).
+func (c *Client) ListInbox(ctx context.Context) ([]InboxObject, error) {
+	var out []InboxObject
+	for obj := range c.mc.ListObjects(ctx, c.Bucket, minio.ListObjectsOptions{Prefix: "inbox/", Recursive: true}) {
+		if obj.Err != nil {
+			return nil, obj.Err
+		}
+		out = append(out, InboxObject{Key: obj.Key, Size: obj.Size, ETag: obj.ETag})
+	}
+	return out, nil
+}
+
+// Get downloads the object into memory and computes its sha256 on the fly
+// (US-12: streamed in memory, no disk writes).
+func (c *Client) Get(ctx context.Context, key string) ([]byte, string, error) {
+	obj, err := c.mc.GetObject(ctx, c.Bucket, key, minio.GetObjectOptions{})
+	if err != nil {
+		return nil, "", err
+	}
+	defer obj.Close()
+	data, err := io.ReadAll(obj)
+	if err != nil {
+		return nil, "", err
+	}
+	sum := sha256.Sum256(data)
+	return data, hex.EncodeToString(sum[:]), nil
+}
+
+func (c *Client) Exists(ctx context.Context, key string) bool {
+	_, err := c.mc.StatObject(ctx, c.Bucket, key, minio.StatObjectOptions{})
+	return err == nil
+}
+
+func (c *Client) Copy(ctx context.Context, srcKey, dstKey string) error {
+	_, err := c.mc.CopyObject(ctx,
+		minio.CopyDestOptions{Bucket: c.Bucket, Object: dstKey},
+		minio.CopySourceOptions{Bucket: c.Bucket, Object: srcKey})
+	return err
+}
+
+// Remove deletes the object; with a versioned bucket this creates a delete
+// marker and the data remains recoverable.
+func (c *Client) Remove(ctx context.Context, key string) error {
+	return c.mc.RemoveObject(ctx, c.Bucket, key, minio.RemoveObjectOptions{})
+}
+
+var extRe = regexp.MustCompile(`(\.[^./]+)$`)
+
+// ResolveConflict returns desired when no object exists under that key,
+// otherwise the name with an incremental numeric suffix before the extension
+// (US-15, US-17): "name.pdf" -> "name (1).pdf", "name (2).pdf", ...
+func (c *Client) ResolveConflict(ctx context.Context, desired string) (string, error) {
+	if !c.Exists(ctx, desired) {
+		return desired, nil
+	}
+	ext := extRe.FindString(desired)
+	base := strings.TrimSuffix(desired, ext)
+	for n := 1; n < 1000; n++ {
+		candidate := base + " (" + strconv.Itoa(n) + ")" + ext
+		if !c.Exists(ctx, candidate) {
+			return candidate, nil
+		}
+	}
+	return "", fmt.Errorf("cannot resolve conflict on %q", desired)
+}