Diffstat (limited to 'internal/blog')
| -rw-r--r-- | internal/blog/b2.go | 133 | ||||
| -rw-r--r-- | internal/blog/client.go | 27 | ||||
| -rw-r--r-- | internal/blog/index.go | 32 | ||||
| -rw-r--r-- | internal/blog/s3.go | 286 |
4 files changed, 0 insertions, 478 deletions
diff --git a/internal/blog/b2.go b/internal/blog/b2.go deleted file mode 100644 index 0caa0ba..0000000 --- a/internal/blog/b2.go +++ /dev/null @@ -1,133 +0,0 @@ -package blog - -import ( - "context" - "fmt" - "slices" - "strings" - "time" - - "github.com/Backblaze/blazer/b2" - "github.com/SayaAndy/saya-today-web/config" - "github.com/SayaAndy/saya-today-web/internal/frontmatter" -) - -type B2Client struct { - prefix string - bucket *b2.Bucket - b2cl *b2.Client -} - -func NewB2Client(cfg *config.StorageConfig) (Client, error) { - if cfg.Type != "b2" { - return nil, fmt.Errorf("invalid storage type for B2InputClient") - } - b2cfg := cfg.Config.(*config.B2Config) - - b2cl, err := b2.NewClient(context.Background(), b2cfg.KeyID, b2cfg.ApplicationKey) - if err != nil { - return nil, err - } - - bucket, err := b2cl.Bucket(context.Background(), b2cfg.BucketName) - if err != nil { - return nil, err - } - - return &B2Client{b2cl: b2cl, bucket: bucket, prefix: b2cfg.Prefix}, nil -} - -func (c *B2Client) Scan(prefix string) ([]*Page, error) { - filePaths := []*Page{} - - iter := c.bucket.List(context.Background(), b2.ListPrefix(c.prefix+prefix)) - - for iter.Next() { - obj := iter.Object() - if obj == nil { - return nil, fmt.Errorf("failed to reference object in B2 bucket") - } - - attrs, err := obj.Attrs(context.Background()) - if err != nil { - return nil, fmt.Errorf("get attributes for object: %w", err) - } - - if attrs.Status != b2.Uploaded { - continue - } - - if !strings.Contains(attrs.ContentType, "text/markdown") { - continue - } - - if _, ok := attrs.Info["title"]; !ok { - continue - } - - publishedTime, err := time.Parse(time.RFC3339, attrs.Info["published-time"]) - if err != nil { - return nil, fmt.Errorf("failed to parse published time metadata field: %w", err) - } - - linkParts := strings.Split(obj.Name(), "/") - nameParts := strings.Split(linkParts[len(linkParts)-1], ".") - fileName := strings.Join(nameParts[:len(linkParts)-1], ".") - - tags := strings.Split(attrs.Info["tags"], ",") - slices.Sort(tags) - - lang, _ := strings.CutPrefix(linkParts[0], c.prefix) - - filePaths = append(filePaths, &Page{ - Link: obj.Name(), - FileName: fileName, - Lang: lang, - ModifiedTime: attrs.LastModified, - Metadata: &frontmatter.Metadata{ - Title: attrs.Info["title"], - ShortDescription: attrs.Info["short-description"], - ActionDate: attrs.Info["action-date"], - PublishedTime: publishedTime, - Thumbnail: attrs.Info["thumbnail"], - Tags: tags, - Geolocation: attrs.Info["geolocation"], - }, - }) - } - - if err := iter.Err(); err != nil { - return nil, fmt.Errorf("iterate over B2 objects: %w", err) - } - - return filePaths, nil -} - -func (c *B2Client) ReadAll(path string) ([]byte, error) { - obj := c.bucket.Object(c.prefix + path) - if obj == nil { - return nil, fmt.Errorf("failed to reference object in B2 bucket") - } - attrs, err := obj.Attrs(context.Background()) - if err != nil { - return nil, fmt.Errorf("error getting attributes of an object: %w", err) - } - - content := make([]byte, attrs.Size) - reader := obj.NewReader(context.Background()) - - if _, err = reader.Read(content); err != nil { - return nil, fmt.Errorf("failed to read file content: %w", err) - } - - return content, nil -} - -func (c *B2Client) ReadFrontmatter(path string) (metadata *frontmatter.Metadata, markdown []byte, err error) { - contentBytes, err := c.ReadAll(path) - if err != nil { - return nil, nil, fmt.Errorf("failed to read file for frontmatter parsing: %w", err) - } - - return frontmatter.ParseFrontmatter(contentBytes) -} diff --git a/internal/blog/client.go b/internal/blog/client.go deleted file mode 100644 index 56bbee4..0000000 --- a/internal/blog/client.go +++ /dev/null @@ -1,27 +0,0 @@ -package blog - -import ( - "time" - - "github.com/SayaAndy/saya-today-web/config" - "github.com/SayaAndy/saya-today-web/internal/frontmatter" -) - -type Page struct { - Link string - FileName string - Lang string - ModifiedTime time.Time - Metadata *frontmatter.Metadata -} - -type Client interface { - Scan(prefix string) ([]*Page, error) - ReadAll(path string) ([]byte, error) - ReadFrontmatter(path string) (metadata *frontmatter.Metadata, markdown []byte, err error) -} - -var NewClientMap = map[string]func(*config.StorageConfig) (Client, error){ - "b2": NewB2Client, - "s3": NewS3Client, -} diff --git a/internal/blog/index.go b/internal/blog/index.go deleted file mode 100644 index 861a90e..0000000 --- a/internal/blog/index.go +++ /dev/null @@ -1,32 +0,0 @@ -package blog - -import "time" - -const IndexFileName = "index.json" - -const IndexSchemaVersion = 1 - -type IndexEntry struct { - Link string `json:"link"` - ModifiedTime time.Time `json:"modifiedTime"` - Title string `json:"title"` - ShortDescription string `json:"shortDescription"` - ActionDate string `json:"actionDate"` - PublishedTime time.Time `json:"publishedTime"` - Thumbnail string `json:"thumbnail"` - Tags []string `json:"tags"` - Geolocation string `json:"geolocation"` - Medley string `json:"medley,omitempty"` - MedleyPart int `json:"medleyPart,omitempty"` -} - -type IndexCategory struct { - GeneratedAt time.Time `json:"generatedAt"` - Pages []IndexEntry `json:"pages"` -} - -type Index struct { - SchemaVersion int `json:"schemaVersion"` - GeneratedAt time.Time `json:"generatedAt"` - Categories map[string]IndexCategory `json:"categories"` -} diff --git a/internal/blog/s3.go b/internal/blog/s3.go deleted file mode 100644 index bbd5238..0000000 --- a/internal/blog/s3.go +++ /dev/null @@ -1,286 +0,0 @@ -package blog - -import ( - "context" - "encoding/json" - "errors" - "fmt" - "io" - "log/slog" - "net/url" - "slices" - "strings" - "sync" - "time" - - "github.com/SayaAndy/saya-today-web/config" - "github.com/SayaAndy/saya-today-web/internal/frontmatter" - "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" - s3types "github.com/aws/aws-sdk-go-v2/service/s3/types" -) - -const s3ScanConcurrency = 32 - -type S3Client struct { - prefix string - bucketName string - s3cl *s3.Client -} - -func NewS3Client(cfg *config.StorageConfig) (Client, error) { - if cfg.Type != "s3" { - return nil, fmt.Errorf("invalid storage type for S3Client") - } - s3cfg := cfg.Config.(*config.S3Config) - - opts := []func(*awsconfig.LoadOptions) error{ - awsconfig.WithRegion(s3cfg.Region), - } - if s3cfg.AccessKeyID != "" && s3cfg.SecretAccessKey != "" { - opts = append(opts, awsconfig.WithCredentialsProvider( - credentials.NewStaticCredentialsProvider(s3cfg.AccessKeyID, s3cfg.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 s3cfg.Endpoint != "" { - s3Opts = append(s3Opts, func(o *s3.Options) { - o.BaseEndpoint = aws.String(s3cfg.Endpoint) - }) - } - s3Opts = append(s3Opts, func(o *s3.Options) { - o.UsePathStyle = s3cfg.UsePathStyle - o.DisableLogOutputChecksumValidationSkipped = true - }) - - s3cl := s3.NewFromConfig(awsCfg, s3Opts...) - - return &S3Client{s3cfg.Prefix, s3cfg.BucketName, s3cl}, nil -} - -func (c *S3Client) Scan(prefix string) ([]*Page, error) { - pages, err := c.scanFromIndex(prefix) - if err == nil { - return pages, nil - } - - var nsk *s3types.NoSuchKey - if !errors.As(err, &nsk) { - return nil, err - } - - slog.Warn("index.json missing, falling back to listing", slog.String("prefix", c.prefix)) - return c.scanByListing(prefix) -} - -func (c *S3Client) scanFromIndex(prefix string) ([]*Page, error) { - out, err := c.s3cl.GetObject(context.Background(), &s3.GetObjectInput{ - Bucket: aws.String(c.bucketName), - Key: aws.String(IndexFileName), - }) - if err != nil { - return nil, fmt.Errorf("get index.json: %w", err) - } - defer out.Body.Close() - - raw, err := io.ReadAll(out.Body) - if err != nil { - return nil, fmt.Errorf("read index.json: %w", err) - } - - var idx Index - if err := json.Unmarshal(raw, &idx); err != nil { - return nil, fmt.Errorf("unmarshal index.json: %w", err) - } - - wantLang := "" - if i := strings.Index(prefix, "/"); i > 0 { - wantLang = prefix[:i] - } - - fullPrefix := c.prefix + prefix - pages := make([]*Page, 0) - for catKey, cat := range idx.Categories { - lang, ok := strings.CutPrefix(catKey, c.prefix) - if !ok { - continue - } - if wantLang != "" && wantLang != lang { - continue - } - for _, e := range cat.Pages { - if !strings.HasPrefix(e.Link, fullPrefix) { - continue - } - linkParts := strings.Split(e.Link, "/") - nameParts := strings.Split(linkParts[len(linkParts)-1], ".") - fileName := strings.Join(nameParts[:len(nameParts)-1], ".") - pages = append(pages, &Page{ - Link: e.Link, - FileName: fileName, - Lang: lang, - ModifiedTime: e.ModifiedTime, - Metadata: &frontmatter.Metadata{ - Title: e.Title, - ShortDescription: e.ShortDescription, - ActionDate: e.ActionDate, - PublishedTime: e.PublishedTime, - Thumbnail: e.Thumbnail, - Tags: e.Tags, - Geolocation: e.Geolocation, - Medley: e.Medley, - MedleyPart: e.MedleyPart, - }, - }) - } - } - return pages, nil -} - -func (c *S3Client) scanByListing(prefix string) ([]*Page, error) { - fullPrefix := c.prefix + prefix - - type candidate struct { - key string - lastModified time.Time - } - var candidates []candidate - - paginator := s3.NewListObjectsV2Paginator(c.s3cl, &s3.ListObjectsV2Input{ - Bucket: aws.String(c.bucketName), - Prefix: aws.String(fullPrefix), - }) - 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 { - key := aws.ToString(obj.Key) - if !strings.HasSuffix(key, ".md") { - continue - } - candidates = append(candidates, candidate{key, aws.ToTime(obj.LastModified)}) - } - } - - pages := make([]*Page, len(candidates)) - sem := make(chan struct{}, s3ScanConcurrency) - var wg sync.WaitGroup - var firstErr error - var errMu sync.Mutex - - for i, cand := range candidates { - sem <- struct{}{} - wg.Go(func() { - defer func() { <-sem }() - - head, err := c.s3cl.HeadObject(context.Background(), &s3.HeadObjectInput{ - Bucket: aws.String(c.bucketName), - Key: aws.String(cand.key), - }) - if err != nil { - errMu.Lock() - if firstErr == nil { - firstErr = fmt.Errorf("head S3 object %s: %w", cand.key, err) - } - errMu.Unlock() - return - } - - if head.ContentType == nil || !strings.Contains(*head.ContentType, "text/markdown") { - return - } - meta := head.Metadata - if meta["title"] == "" { - return - } - - publishedTime, err := time.Parse(time.RFC3339, meta["published-time"]) - if err != nil { - errMu.Lock() - if firstErr == nil { - firstErr = fmt.Errorf("failed to parse published time metadata field: %w", err) - } - errMu.Unlock() - return - } - - linkParts := strings.Split(cand.key, "/") - nameParts := strings.Split(linkParts[len(linkParts)-1], ".") - fileName := strings.Join(nameParts[:len(nameParts)-1], ".") - lang, _ := strings.CutPrefix(linkParts[0], c.prefix) - - tags := strings.Split(meta["tags"], ",") - slices.Sort(tags) - - title, _ := url.QueryUnescape(meta["title"]) - shortDescription, _ := url.QueryUnescape(meta["short-description"]) - thumbnail, _ := url.QueryUnescape(meta["thumbnail"]) - - pages[i] = &Page{ - Link: cand.key, - FileName: fileName, - Lang: lang, - ModifiedTime: cand.lastModified, - Metadata: &frontmatter.Metadata{ - Title: title, - ShortDescription: shortDescription, - ActionDate: meta["action-date"], - PublishedTime: publishedTime, - Thumbnail: thumbnail, - Tags: tags, - Geolocation: meta["geolocation"], - }, - } - }) - } - wg.Wait() - - if firstErr != nil { - return nil, firstErr - } - - out := pages[:0] - for _, p := range pages { - if p != nil { - out = append(out, p) - } - } - return out, nil -} - -func (c *S3Client) ReadAll(path string) ([]byte, error) { - output, err := c.s3cl.GetObject(context.Background(), &s3.GetObjectInput{ - Bucket: aws.String(c.bucketName), - Key: aws.String(c.prefix + path), - }) - if err != nil { - return nil, fmt.Errorf("get S3 object: %w", err) - } - defer output.Body.Close() - - content, err := io.ReadAll(output.Body) - if err != nil { - return nil, fmt.Errorf("read S3 object body: %w", err) - } - - return content, nil -} - -func (c *S3Client) ReadFrontmatter(path string) (metadata *frontmatter.Metadata, markdown []byte, err error) { - contentBytes, err := c.ReadAll(path) - if err != nil { - return nil, nil, fmt.Errorf("failed to read file for frontmatter parsing: %w", err) - } - - return frontmatter.ParseFrontmatter(contentBytes) -} |