Diffstat (limited to 'internal/blog')
| -rw-r--r-- | internal/blog/b2.go | 150 | ||||
| -rw-r--r-- | internal/blog/client.go | 28 | ||||
| -rw-r--r-- | internal/blog/index.go | 103 | ||||
| -rw-r--r-- | internal/blog/s3.go | 241 |
4 files changed, 522 insertions, 0 deletions
diff --git a/internal/blog/b2.go b/internal/blog/b2.go new file mode 100644 index 0000000..f2ac890 --- /dev/null +++ b/internal/blog/b2.go @@ -0,0 +1,150 @@ +package blog + +import ( + "context" + "encoding/json" + "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) GetMedleys() ([]MedleyEntry, error) { + idxRaw, err := c.readAll(MedleysIndexFileName) + if err != nil { + return nil, fmt.Errorf("read %s: %w", MedleysIndexFileName, err) + } + + var idx []MedleyEntry + if err := json.Unmarshal(idxRaw, &idx); err != nil { + return nil, fmt.Errorf("unmarshal %s: %w", MedleysIndexFileName, err) + } + + return idx, 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) + } + + link := obj.Name() + fileName := link[strings.LastIndex(link, "/")+1 : strings.LastIndex(link, ".")] + tags := strings.Split(attrs.Info["tags"], ",") + slices.Sort(tags) + + lang, _ := strings.CutPrefix(link[0:strings.Index(link, "/")], c.prefix) + + filePaths = append(filePaths, &Page{ + Link: link, + 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) { + return c.readAll(c.prefix + path) +} + +func (c *B2Client) readAll(path string) ([]byte, error) { + obj := c.bucket.Object(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 new file mode 100644 index 0000000..53eb0fd --- /dev/null +++ b/internal/blog/client.go @@ -0,0 +1,28 @@ +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) + GetMedleys() ([]MedleyEntry, 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 new file mode 100644 index 0000000..a09b59e --- /dev/null +++ b/internal/blog/index.go @@ -0,0 +1,103 @@ +package blog + +import ( + "encoding/json" + "fmt" + "time" + + "github.com/SayaAndy/saya-today-web/internal/frontmatter" +) + +const IndexFileName = "index.json" +const MedleysIndexFileName = "medleys.json" + +const IndexSchemaVersion = 2 + +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"` +} + +func (e IndexEntry) Metadata() *frontmatter.Metadata { + return &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, + } +} + +type IndexV2Category struct { + GeneratedAt time.Time `json:"generatedAt"` + Pages map[string]IndexEntry `json:"pages"` +} + +type IndexV1Category struct { + GeneratedAt time.Time `json:"generatedAt"` + Pages []IndexEntry `json:"pages"` +} + +type Index struct { + SchemaVersion int `json:"schemaVersion"` + GeneratedAt time.Time `json:"generatedAt"` + Categories any `json:"categories"` +} + +func (idx *Index) UnmarshalJSON(data []byte) error { + var tmp struct { + SchemaVersion int `json:"schemaVersion"` + GeneratedAt time.Time `json:"generatedAt"` + Categories json.RawMessage `json:"categories"` + } + + if err := json.Unmarshal(data, &tmp); err != nil { + return err + } + + idx.SchemaVersion = tmp.SchemaVersion + idx.GeneratedAt = tmp.GeneratedAt + + switch tmp.SchemaVersion { + case 1: + var categories map[string]*IndexV1Category + if err := json.Unmarshal(tmp.Categories, &categories); err != nil { + return fmt.Errorf("unmarshal map[string]*IndexV1Category: %w", err) + } + idx.Categories = &categories + case 2: + var categories map[string]*IndexV2Category + if err := json.Unmarshal(tmp.Categories, &categories); err != nil { + return fmt.Errorf("unmarshal map[string]*IndexV2Category: %w", err) + } + idx.Categories = &categories + default: + return fmt.Errorf("unsupported index version: %d", tmp.SchemaVersion) + } + + return nil +} + +type MedleyEntry struct { + Codename string `json:"codename"` + Localnames map[string]string `json:"localnames"` + Content []string `json:"content"` +} + +type MedleyPageEntry struct { + Codename string `json:"codename"` + Position int `json:"position"` +} diff --git a/internal/blog/s3.go b/internal/blog/s3.go new file mode 100644 index 0000000..3c1a932 --- /dev/null +++ b/internal/blog/s3.go @@ -0,0 +1,241 @@ +package blog + +import ( + "bytes" + "context" + "encoding/json" + "fmt" + "io" + "strings" + + "github.com/SayaAndy/saya-today-web/config" + "github.com/SayaAndy/saya-today-web/internal/frontmatter" + "github.com/SayaAndy/saya-today-web/l10n" + "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" +) + +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) GetMedleys() ([]MedleyEntry, error) { + idxRaw, err := c.readAll(MedleysIndexFileName) + if err != nil { + return nil, fmt.Errorf("read %s: %w", MedleysIndexFileName, err) + } + + var idx []MedleyEntry + if err := json.Unmarshal(idxRaw, &idx); err != nil { + return nil, fmt.Errorf("unmarshal %s: %w", MedleysIndexFileName, err) + } + + return idx, nil +} + +func (c *S3Client) Scan(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 %s: %w", IndexFileName, err) + } + defer out.Body.Close() + + raw, err := io.ReadAll(out.Body) + if err != nil { + return nil, fmt.Errorf("read %s: %w", IndexFileName, err) + } + + var idx Index + if err := json.Unmarshal(raw, &idx); err != nil { + return nil, fmt.Errorf("unmarshal %s: %w", IndexFileName, err) + } + + wantLang := "" + if i := strings.Index(prefix, "/"); i > 0 { + wantLang = prefix[:i] + } + + fullPrefix := c.prefix + prefix + pages := make([]*Page, 0) + + switch idx.SchemaVersion { + case 1: + for catKey, cat := range *idx.Categories.(*map[string]*IndexV1Category) { + 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 + } + fileName := e.Link[strings.LastIndex(e.Link, "/")+1 : strings.LastIndex(e.Link, ".")] + 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, + }, + }) + } + } + case 2: + for catKey, cat := range *idx.Categories.(*map[string]*IndexV2Category) { + lang, ok := strings.CutPrefix(catKey, c.prefix) + if !ok { + continue + } + if wantLang != "" && wantLang != lang { + continue + } + for codename, e := range cat.Pages { + if !strings.HasPrefix(e.Link, fullPrefix) { + continue + } + pages = append(pages, &Page{ + Link: e.Link, + FileName: codename, + 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, + }, + }) + } + } + } + + medleys, _ := c.GetMedleys() + for _, medley := range medleys { + for locale, localname := range medley.Localnames { + l10n.T.SetPath(localname, true, locale, "Medleys", medley.Codename) + } + } + + return pages, nil +} + +func (c *S3Client) ReadAll(path string) ([]byte, error) { + return c.readAll(c.prefix + path) +} + +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(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) { + idxRaw, err := c.readAll(IndexFileName) + if err != nil { + return nil, nil, fmt.Errorf("read %s: %w", IndexFileName, err) + } + + var idx Index + if err := json.Unmarshal(idxRaw, &idx); err != nil { + return nil, nil, fmt.Errorf("unmarshal %s: %w", IndexFileName, err) + } + + contentBytes, err := c.ReadAll(path) + if err != nil { + return nil, nil, fmt.Errorf("failed to read file for frontmatter parsing: %w", err) + } + + switch idx.SchemaVersion { + case 1: + return frontmatter.ParseFrontmatter(contentBytes) + case 2: + fullPath := c.prefix + path + page := (*idx.Categories.(*map[string]*IndexV2Category))[fullPath[:strings.LastIndex(fullPath, "/")]].Pages[fullPath[strings.LastIndex(fullPath, "/")+1:strings.LastIndex(fullPath, ".")]] + metadata = page.Metadata() + + if !bytes.HasPrefix(contentBytes, []byte("---\n")) { + return metadata, contentBytes, nil + } + + end := bytes.Index(contentBytes[4:], []byte("\n---\n")) + if end == -1 { + return metadata, contentBytes, nil + } + + return metadata, contentBytes[end+9:], nil + } + + return frontmatter.ParseFrontmatter(contentBytes) +} |