download + extract pipeline

- artifact.Download: resumable HTTP with optional SHA256 check + progress cb
- artifact.ExtractZstdTar: streamed zstd+tar with tar-slip defense
- aggregator client matches real API shape (digests/immutables/ancillary blocks
  with URIHolder polymorphism for templated immutable URIs)
- cmd: show + download subcommands wired up
- end-to-end verified against preprod: digests archive pulls cleanly, yields
  16836-entry SHA manifest ready for verification sprint

deps: github.com/klauspost/compress (pure-go zstd)
This commit is contained in:
Sulkta 2026-04-23 15:16:48 -07:00
parent ddf1ec3491
commit e4012bf79a
6 changed files with 483 additions and 84 deletions

View file

@ -1,7 +1,7 @@
// Package aggregator is a thin HTTP client for the Mithril aggregator REST API.
//
// Only the handful of endpoints needed for client-side snapshot workflows are
// exposed. Authentication is not required for the read paths used here.
// Only the endpoints needed for client-side snapshot workflows are exposed.
// Authentication is not required for the read paths used here.
package aggregator
import (
@ -23,57 +23,98 @@ type Client struct {
func New(baseURL string) *Client {
return &Client{
baseURL: strings.TrimRight(baseURL, "/"),
http: &http.Client{Timeout: 60 * time.Second},
http: &http.Client{Timeout: 120 * time.Second},
}
}
// CardanoDBSnapshot is the server-reported shape for /artifact/cardano-database/{hash}.
// Field set is trimmed to what the client actually consumes — full schema documented
// at https://mithril.network/doc/aggregator-api/.
// CardanoDBSnapshot is the server shape for /artifact/cardano-database and its
// /{hash} detail endpoint. The list response omits {digests,immutables,ancillary};
// only the detail endpoint populates them.
type CardanoDBSnapshot struct {
Hash string `json:"hash"`
MerkleRoot string `json:"merkle_root"`
Network string `json:"network"`
Beacon Beacon `json:"beacon"`
CertificateHash string `json:"certificate_hash"`
TotalDBSizeUncompressed uint64 `json:"total_db_size_uncompressed"`
Digests LocationList `json:"digests"`
ImmutablesAncillary LocationList `json:"immutables"`
ImmutablesIncremental *IncrementalImmutables `json:"immutables_incremental,omitempty"`
CreatedAt time.Time `json:"created_at"`
Hash string `json:"hash"`
MerkleRoot string `json:"merkle_root"`
Network string `json:"network"`
Beacon Beacon `json:"beacon"`
CertificateHash string `json:"certificate_hash"`
TotalDBSizeUncompressed uint64 `json:"total_db_size_uncompressed"`
CardanoNodeVersion string `json:"cardano_node_version"`
CreatedAt time.Time `json:"created_at"`
Digests DigestsBlock `json:"digests"`
Immutables ImmutsBlock `json:"immutables"`
Ancillary AncillaryBlock `json:"ancillary"`
}
type Beacon struct {
Epoch uint64 `json:"epoch"`
Epoch uint64 `json:"epoch"`
ImmutableFileNumber uint64 `json:"immutable_file_number"`
}
type LocationList struct {
Size uint64 `json:"size"`
Locations []LocationAlt `json:"locations"`
type DigestsBlock struct {
SizeUncompressed uint64 `json:"size_uncompressed"`
Locations []Location `json:"locations"`
}
// LocationAlt is a best-of alternative; Mithril returns a typed-discriminated object.
type LocationAlt struct {
Type string `json:"type"` // e.g. "cloud_storage", "ipfs"
URI string `json:"uri"`
type ImmutsBlock struct {
AverageSizeUncompressed uint64 `json:"average_size_uncompressed"`
Locations []Location `json:"locations"`
}
type IncrementalImmutables struct {
AverageSize uint64 `json:"average_size"`
Locations []LocationAlt `json:"locations"`
type AncillaryBlock struct {
SizeUncompressed uint64 `json:"size_uncompressed"`
Locations []Location `json:"locations"`
}
// Location is a polymorphic URI holder. The Mithril API ships URI as either
// a plain string (for single artifacts) or as {"Template": "..."} for
// templated per-file URIs (immutables only).
type Location struct {
Type string `json:"type"`
URI URIHolder `json:"uri"`
CompressionAlgorithm string `json:"compression_algorithm,omitempty"`
}
// URIHolder absorbs both string and templated-object URI shapes.
type URIHolder struct {
Plain string
Template string
}
func (h *URIHolder) UnmarshalJSON(b []byte) error {
// Try plain string first
var s string
if err := json.Unmarshal(b, &s); err == nil {
h.Plain = s
return nil
}
// Fall back to {"Template": "..."}
var t struct {
Template string `json:"Template"`
}
if err := json.Unmarshal(b, &t); err == nil {
h.Template = t.Template
return nil
}
return fmt.Errorf("unrecognized URI shape: %s", string(b))
}
// String returns whichever URI form is populated.
func (h URIHolder) String() string {
if h.Template != "" {
return h.Template
}
return h.Plain
}
// Certificate is the server-reported shape for /certificate/{hash}.
// Kept minimal; STM verification reads what it needs from the raw JSON later.
// Kept wide — STM verification consumes raw bytes separately from the decoded view.
type Certificate struct {
Hash string `json:"hash"`
PreviousHash string `json:"previous_hash"`
Epoch uint64 `json:"epoch"`
SignedMessage string `json:"signed_message"`
ProtocolMessage json.RawMessage `json:"protocol_message"`
Multisignature json.RawMessage `json:"multi_signature"`
GenesisSignature string `json:"genesis_signature,omitempty"`
Hash string `json:"hash"`
PreviousHash string `json:"previous_hash"`
Epoch uint64 `json:"epoch"`
SignedMessage string `json:"signed_message"`
ProtocolMessage json.RawMessage `json:"protocol_message"`
Multisignature json.RawMessage `json:"multi_signature"`
GenesisSignature string `json:"genesis_signature,omitempty"`
}
func (c *Client) get(ctx context.Context, path string, out any) error {
@ -84,12 +125,12 @@ func (c *Client) get(ctx context.Context, path string, out any) error {
req.Header.Set("Accept", "application/json")
resp, err := c.http.Do(req)
if err != nil {
return fmt.Errorf("aggregator GET %s: %w", path, err)
return fmt.Errorf("GET %s: %w", path, err)
}
defer resp.Body.Close()
if resp.StatusCode >= 400 {
body, _ := io.ReadAll(io.LimitReader(resp.Body, 4096))
return fmt.Errorf("aggregator GET %s: status %d: %s", path, resp.StatusCode, string(body))
return fmt.Errorf("GET %s: %d: %s", path, resp.StatusCode, string(body))
}
if out == nil {
return nil
@ -97,16 +138,11 @@ func (c *Client) get(ctx context.Context, path string, out any) error {
return json.NewDecoder(resp.Body).Decode(out)
}
// ListCardanoDBSnapshots returns the sorted-newest-first list of cardano-database snapshots.
func (c *Client) ListCardanoDBSnapshots(ctx context.Context) ([]CardanoDBSnapshot, error) {
var out []CardanoDBSnapshot
if err := c.get(ctx, "/artifact/cardano-database", &out); err != nil {
return nil, err
}
return out, nil
return out, c.get(ctx, "/artifact/cardano-database", &out)
}
// GetCardanoDBSnapshot fetches details for a single snapshot by hash (or "latest").
func (c *Client) GetCardanoDBSnapshot(ctx context.Context, hash string) (*CardanoDBSnapshot, error) {
var out CardanoDBSnapshot
if err := c.get(ctx, "/artifact/cardano-database/"+url.PathEscape(hash), &out); err != nil {
@ -115,7 +151,6 @@ func (c *Client) GetCardanoDBSnapshot(ctx context.Context, hash string) (*Cardan
return &out, nil
}
// GetCertificate fetches a certificate by hash for signature verification.
func (c *Client) GetCertificate(ctx context.Context, hash string) (*Certificate, error) {
var out Certificate
if err := c.get(ctx, "/certificate/"+url.PathEscape(hash), &out); err != nil {