Diffstat (limited to 'internal/client/input')
| -rw-r--r-- | internal/client/input/b2.go | 52 | ||||
| -rw-r--r-- | internal/client/input/input_client_interface.go | 25 | ||||
| -rw-r--r-- | internal/client/input/local_unix.go | 108 | ||||
| -rw-r--r-- | internal/client/input/s3.go | 157 |
4 files changed, 317 insertions, 25 deletions
diff --git a/internal/client/input/b2.go b/internal/client/input/b2.go index 67fb14e..fe94333 100644 --- a/internal/client/input/b2.go +++ b/internal/client/input/b2.go @@ -16,11 +16,12 @@ var _ InputClient = (*B2InputClient)(nil) type B2InputClient struct { prefix string bucket *b2.Bucket + bucketName string b2cl *b2.Client knownExtensions []string } -func NewB2InputClient(cfg *config.InputConfig) (*B2InputClient, error) { +func NewB2InputClient(cfg *config.InputConfig) (InputClient, error) { if cfg.Storage.Type != "b2" { return nil, fmt.Errorf("invalid storage type for B2InputClient") } @@ -36,7 +37,7 @@ func NewB2InputClient(cfg *config.InputConfig) (*B2InputClient, error) { return nil, err } - return &B2InputClient{b2cl: b2cl, bucket: bucket, prefix: b2cfg.Prefix, knownExtensions: cfg.KnownExtensions}, nil + return &B2InputClient{b2cl: b2cl, bucket: bucket, bucketName: b2cfg.BucketName, prefix: b2cfg.Prefix, knownExtensions: cfg.KnownExtensions}, nil } func (c *B2InputClient) Scan() ([]string, error) { @@ -72,7 +73,7 @@ func (c *B2InputClient) Scan() ([]string, error) { } } - filePaths = append(filePaths, name) + filePaths = append(filePaths, strings.TrimPrefix(name, c.prefix)) } if err := iter.Err(); err != nil { @@ -82,8 +83,8 @@ func (c *B2InputClient) Scan() ([]string, error) { return filePaths, nil } -func (c *B2InputClient) ReadMetadata(path string) (map[string]string, error) { - obj := c.bucket.Object(path) +func (c *B2InputClient) ReadMetadata(path string) (*MetadataStruct, error) { + obj := c.bucket.Object(c.prefix + path) if obj == nil { return nil, fmt.Errorf("object not found in B2 bucket") } @@ -93,37 +94,42 @@ func (c *B2InputClient) ReadMetadata(path string) (map[string]string, error) { return nil, fmt.Errorf("get attributes for object: %w", err) } - metadata := attrs.Info - metadata["Name"] = attrs.Name - metadata["Size"] = fmt.Sprintf("%d", attrs.Size) - metadata["ContentType"] = attrs.ContentType - metadata["LastModified"] = attrs.LastModified.Format("2006-01-02T15:04:05Z") - metadata["SHA1"] = attrs.SHA1 - metadata["UploadTimestamp"] = attrs.UploadTimestamp.Format("2006-01-02T15:04:05Z") + metadata := MetadataStruct{ + Name: attrs.Name, + StorageType: "b2", + Hash: attrs.SHA1, + ContentType: attrs.ContentType, + FirstCreated: attrs.UploadTimestamp, + LastModified: attrs.LastModified, + Misc: attrs.Info, + Size: attrs.Size, + } switch attrs.Status { case b2.Uploaded: - metadata["Status"] = "Uploaded" + metadata.Misc["b2-status"] = "Uploaded" case b2.Folder: - metadata["Status"] = "Folder" + metadata.Misc["b2-status"] = "Folder" case b2.Hider: - metadata["Status"] = "Hider" + metadata.Misc["b2-status"] = "Hider" case b2.Started: - metadata["Status"] = "Started" + metadata.Misc["b2-status"] = "Started" default: - metadata["Status"] = "Unknown" + metadata.Misc["b2-status"] = "Unknown" } - return metadata, nil + return &metadata, nil +} + +func (c *B2InputClient) ID(path string) string { + return fmt.Sprintf("b2://%s/%s%s", c.bucketName, c.prefix, path) } -func (c *B2InputClient) GetReader(path string) (io.Reader, error) { - obj := c.bucket.Object(path) +func (c *B2InputClient) GetReader(path string) (io.ReadCloser, error) { + obj := c.bucket.Object(c.prefix + path) if obj == nil { return nil, fmt.Errorf("failed to reference object in B2 bucket") } - reader := obj.NewReader(context.Background()) - - return reader, nil + return obj.NewReader(context.Background()), nil } diff --git a/internal/client/input/input_client_interface.go b/internal/client/input/input_client_interface.go index 4bbbae5..dca64db 100644 --- a/internal/client/input/input_client_interface.go +++ b/internal/client/input/input_client_interface.go @@ -2,10 +2,31 @@ package input import ( "io" + "time" + + "github.com/SayaAndy/saya-today-thumbnail-generator/config" ) type InputClient interface { Scan() ([]string, error) - ReadMetadata(string) (map[string]string, error) - GetReader(string) (io.Reader, error) + ReadMetadata(string) (*MetadataStruct, error) + GetReader(string) (io.ReadCloser, error) + ID(path string) string +} + +type MetadataStruct struct { + Name string + StorageType string + Hash string + ContentType string + FirstCreated time.Time + LastModified time.Time + Size int64 + Misc map[string]string +} + +var NewInputClientMap = map[string]func(cfg *config.InputConfig) (InputClient, error){ + "b2": NewB2InputClient, + "s3": NewS3InputClient, + "local-unix": NewLocalUnixInputClient, } diff --git a/internal/client/input/local_unix.go b/internal/client/input/local_unix.go new file mode 100644 index 0000000..21b08a9 --- /dev/null +++ b/internal/client/input/local_unix.go @@ -0,0 +1,108 @@ +package input + +import ( + "fmt" + "io" + "log/slog" + "mime" + "os" + "slices" + "strconv" + "strings" + "syscall" + "time" + + "github.com/SayaAndy/saya-today-thumbnail-generator/config" +) + +var _ InputClient = (*LocalUnixInputClient)(nil) + +type LocalUnixInputClient struct { + path string + maxDepth int + knownExtensions []string +} + +func NewLocalUnixInputClient(cfg *config.InputConfig) (InputClient, error) { + if cfg.Storage.Type != "local-unix" { + return nil, fmt.Errorf("invalid storage type for LocalUnixInputClient") + } + localCfg := cfg.Storage.Config.(*config.InputLocalUnixConfig) + + return &LocalUnixInputClient{ + path: localCfg.Path, + maxDepth: localCfg.MaxDepth, + knownExtensions: cfg.KnownExtensions, + }, nil +} + +func (c *LocalUnixInputClient) Scan() ([]string, error) { + return c.recursiveScan(c.path, c.maxDepth) +} + +func (c *LocalUnixInputClient) recursiveScan(dir string, depth int) ([]string, error) { + filePaths := make([]string, 0) + entries, err := os.ReadDir(dir) + if err != nil { + return nil, fmt.Errorf("fail to read directory: %w", err) + } + + for _, entry := range entries { + if entry.IsDir() && depth > 0 { + subFilePaths, err := c.recursiveScan(dir+entry.Name()+"/", depth-1) + if err != nil { + return nil, fmt.Errorf("fail to scan subdirectory '%s': %w", entry.Name(), err) + } + filePaths = append(filePaths, subFilePaths...) + } else if !entry.IsDir() { + nameParts := strings.Split(entry.Name(), ".") + if len(nameParts) < 2 { + continue + } + if slices.Contains(c.knownExtensions, strings.ToLower(nameParts[len(nameParts)-1])) { + filePaths = append(filePaths, strings.TrimPrefix(dir+entry.Name(), c.path)) + } + fmt.Println(entry.Name()) + } + } + + return filePaths, nil +} + +func (c *LocalUnixInputClient) ReadMetadata(path string) (*MetadataStruct, error) { + nodePathParts := strings.Split(path, "/") + nodeName := nodePathParts[len(nodePathParts)-1] + nodeNameParts := strings.Split(nodeName, ".") + nodeExt := "" + if len(nodeNameParts) >= 2 { + nodeExt = nodeNameParts[len(nodeNameParts)-1] + } + slog.Debug("got a file extension", slog.String("extension", nodeExt), slog.String("path", path), slog.String("filename", nodeName)) + + fileInfo, err := os.Stat(c.path + path) + if err != nil { + return nil, fmt.Errorf("fail to read file info: %w", err) + } + + stat_t := fileInfo.Sys().(*syscall.Stat_t) + creationTime := time.Unix(stat_t.Ctimespec.Sec, stat_t.Ctimespec.Nsec) + + return &MetadataStruct{ + Name: fileInfo.Name(), + StorageType: "local-unix", + Hash: strconv.FormatInt(fileInfo.ModTime().Unix(), 16), + ContentType: mime.TypeByExtension("." + nodeExt), + FirstCreated: creationTime, + LastModified: fileInfo.ModTime(), + Size: fileInfo.Size(), + Misc: map[string]string{}, + }, nil +} + +func (c *LocalUnixInputClient) ID(path string) string { + return fmt.Sprintf("local-unix://%s%s", c.path, path) +} + +func (c *LocalUnixInputClient) GetReader(path string) (io.ReadCloser, error) { + return os.Open(c.path + path) +} diff --git a/internal/client/input/s3.go b/internal/client/input/s3.go new file mode 100644 index 0000000..00684c2 --- /dev/null +++ b/internal/client/input/s3.go @@ -0,0 +1,157 @@ +package input + +import ( + "context" + "fmt" + "io" + "slices" + "strings" + + "github.com/aws/aws-sdk-go-v2/aws" + awsconfig "github.com/aws/aws-sdk-go-v2/config" + "github.com/aws/aws-sdk-go-v2/credentials" + "github.com/aws/aws-sdk-go-v2/service/s3" + + "github.com/SayaAndy/saya-today-thumbnail-generator/config" +) + +var _ InputClient = (*S3InputClient)(nil) + +type S3InputClient struct { + prefix string + bucketName string + s3cl *s3.Client + knownExtensions []string +} + +func NewS3InputClient(cfg *config.InputConfig) (InputClient, error) { + if cfg.Storage.Type != "s3" { + return nil, fmt.Errorf("invalid storage type for S3InputClient") + } + s3cfg := cfg.Storage.Config.(*config.S3Config) + + s3cl, err := newS3Client(s3cfg) + if err != nil { + return nil, fmt.Errorf("create S3 client: %w", err) + } + + return &S3InputClient{ + s3cl: s3cl, + bucketName: s3cfg.BucketName, + prefix: s3cfg.Prefix, + knownExtensions: cfg.KnownExtensions, + }, nil +} + +func (c *S3InputClient) Scan() ([]string, error) { + var filePaths []string + + paginator := s3.NewListObjectsV2Paginator(c.s3cl, &s3.ListObjectsV2Input{ + Bucket: aws.String(c.bucketName), + Prefix: aws.String(c.prefix), + }) + + for paginator.HasMorePages() { + output, err := paginator.NextPage(context.Background()) + if err != nil { + return nil, fmt.Errorf("list S3 objects: %w", err) + } + + for _, obj := range output.Contents { + name := aws.ToString(obj.Key) + + if len(c.knownExtensions) != 0 { + nameParts := strings.Split(name, ".") + if len(nameParts) < 2 { + continue + } + ext := strings.ToLower(nameParts[len(nameParts)-1]) + if !slices.Contains(c.knownExtensions, ext) { + continue + } + } + + filePaths = append(filePaths, strings.TrimPrefix(name, c.prefix)) + } + } + + return filePaths, nil +} + +func (c *S3InputClient) ReadMetadata(path string) (*MetadataStruct, error) { + key := c.prefix + path + + head, err := c.s3cl.HeadObject(context.Background(), &s3.HeadObjectInput{ + Bucket: aws.String(c.bucketName), + Key: aws.String(key), + }) + if err != nil { + return nil, fmt.Errorf("head S3 object %s: %w", key, err) + } + + metadata := MetadataStruct{ + Name: key, + StorageType: "s3", + Hash: strings.Trim(aws.ToString(head.ETag), "\""), + ContentType: aws.ToString(head.ContentType), + Misc: head.Metadata, + } + + if head.ContentLength != nil { + metadata.Size = *head.ContentLength + } + if head.LastModified != nil { + metadata.LastModified = *head.LastModified + metadata.FirstCreated = *head.LastModified + } + + return &metadata, nil +} + +func (c *S3InputClient) ID(path string) string { + return fmt.Sprintf("s3://%s/%s%s", c.bucketName, c.prefix, path) +} + +func (c *S3InputClient) GetReader(path string) (io.ReadCloser, error) { + key := c.prefix + path + + output, err := c.s3cl.GetObject(context.Background(), &s3.GetObjectInput{ + Bucket: aws.String(c.bucketName), + Key: aws.String(key), + }) + if err != nil { + return nil, fmt.Errorf("get S3 object %s: %w", key, err) + } + + return output.Body, nil +} + +func newS3Client(cfg *config.S3Config) (*s3.Client, error) { + opts := []func(*awsconfig.LoadOptions) error{ + awsconfig.WithRegion(cfg.Region), + } + + if cfg.AccessKeyID != "" && cfg.SecretAccessKey != "" { + opts = append(opts, awsconfig.WithCredentialsProvider( + credentials.NewStaticCredentialsProvider(cfg.AccessKeyID, cfg.SecretAccessKey, ""), + )) + } + + awsCfg, err := awsconfig.LoadDefaultConfig(context.Background(), opts...) + if err != nil { + return nil, fmt.Errorf("load AWS config: %w", err) + } + + var s3Opts []func(*s3.Options) + if cfg.Endpoint != "" { + s3Opts = append(s3Opts, func(o *s3.Options) { + o.BaseEndpoint = aws.String(cfg.Endpoint) + }) + } + s3Opts = append(s3Opts, func(o *s3.Options) { + o.UsePathStyle = cfg.UsePathStyle + o.DisableLogOutputChecksumValidationSkipped = true + }) + + return s3.NewFromConfig(awsCfg, s3Opts...), nil +} |