feat(x/data)!: enable off-chain coordination of supported algorithms

proto/regen/data/v1/types.proto

@@ -5,21 +5,39 @@ package regen.data.v1;
 option go_package = "github.com/regen-network/regen-ledger/x/data";
 
 // ContentHash specifies a hash-based content identifier for a piece of data.
+// Exactly one of its fields must be set so this message behaves like a oneof.
+// A protobuf oneof was not used because this caused compatibility issues with
+// amino signing.
 message ContentHash {
-
+  // Deprecated: use RawV2 instead.
   // Raw specifies "raw" data which does not specify a deterministic, canonical
   // encoding. Users of these hashes MUST maintain a copy of the hashed data
   // which is preserved bit by bit. All other content encodings specify a
   // deterministic, canonical encoding allowing implementations to choose from a
   // variety of alternative formats for transport and encoding while maintaining
   // the guarantee that the canonical hash will not change. The media type for
   // "raw" data is defined by the MediaType enum.
-  Raw raw = 1;
+  Raw raw = 1 [deprecated = true];
 
+  // Deprecated: use GraphV2 instead.
   // Graph specifies graph data that conforms to the RDF data model.
   // The canonicalization algorithm used for an RDF graph is specified by
   // GraphCanonicalizationAlgorithm.
-  Graph graph = 2;
+  Graph graph = 2 [deprecated = true];
+
+  // raw_v2 specifies "raw" data which does not specify a deterministic, canonical
+  // encoding. Users of these hashes MUST maintain a copy of the hashed data
+  // which is preserved bit by bit. All other content encodings specify a
+  // deterministic, canonical encoding allowing implementations to choose from a
+  // variety of alternative formats for transport and encoding while maintaining
+  // the guarantee that the canonical hash will not change. The media type for
+  // "raw" data is defined by the MediaType enum.
+  RawV2 raw_v2 = 3;
+
+  // graph_v2 specifies graph data that conforms to the RDF data model.
+  // The canonicalization algorithm used for an RDF graph is specified by
+  // GraphCanonicalizationAlgorithm.
+  GraphV2 graph_v2 = 4;
 
   // Raw is the content hash type used for raw data.
   message Raw {
@@ -50,21 +68,64 @@ message ContentHash {
     // merkle_tree is the merkle tree type used for the graph hash, if any.
     GraphMerkleTree merkle_tree = 4;
   }
+
+  // RawV2 is the content hash type used for raw data.
+  message RawV2 {
+    // hash represents the hash of the data based on the specified
+    // digest_algorithm. It must be at least 20 bytes long and at most 64 bytes long.
+    bytes hash = 1;
+
+    // digest_algorithm represents the hash digest algorithm and should be a non-zero value from the DigestAlgorithm enum.
+    uint32 digest_algorithm = 2;
+
+    // file_extension represents the file extension for raw data. It can be
+    // must be between 2-6 characters long, must be all lower-case and should represent
+    // the canonical extension for the media type.
+    //
+    // A list of canonical extensions which should be used is provided here
+    // and SHOULD be used by implementations: txt, json, csv, xml, pdf, tiff,
+    // jpg, png, svg, webp, avif, gif, apng, mpeg, mp4, webm, ogg, heic, raw.
+    //
+    // The above list should be updated as new media types come into common usage
+    // especially when there are two or more possible extensions (i.e. jpg vs jpeg or tif vs tiff).
+    string file_extension = 3;
+  }
+
+  // GraphV2 is the content hash type used for RDF graph data.
+  message GraphV2 {
+    // hash represents the hash of the data based on the specified
+    // digest_algorithm. It must be at least 20 bytes long and at most 64 bytes long.
+    bytes hash = 1;
+
+    // digest_algorithm represents the hash digest algorithm and should be a non-zero value from the DigestAlgorithm enum.
+    uint32 digest_algorithm = 2;
+
+    // graph_canonicalization_algorithm represents the RDF graph
+    // canonicalization algorithm and should be a non-zero value from the GraphCanonicalizationAlgorithm enum.
+    uint32 canonicalization_algorithm = 3;
+
+    // merkle_tree is the merkle tree type used for the graph hash, if any and should be a value from the GraphMerkleTree enum
+    // or left unspecified.
+    uint32 merkle_tree = 4;
+  }
 }
 
 // DigestAlgorithm is the hash digest algorithm
+//
+// With V2 of raw and graph hash, this enum is no longer validated on-chain.
+// However, this enum SHOULD still be used and updated as a registry of known digest
+// algorithms and all implementations should coordinate on these values.
 enum DigestAlgorithm {
-
   // unspecified and invalid
   DIGEST_ALGORITHM_UNSPECIFIED = 0;
 
   // BLAKE2b-256
   DIGEST_ALGORITHM_BLAKE2B_256 = 1;
 }
 
+// Deprecated: use RawV2 instead.
 // RawMediaType defines MIME media types to be used with a ContentHash.Raw hash.
 enum RawMediaType {
-
   // RAW_MEDIA_TYPE_UNSPECIFIED can be used for raw binary data
   RAW_MEDIA_TYPE_UNSPECIFIED = 0;
 
@@ -127,25 +188,30 @@ enum RawMediaType {
 }
 
 // GraphCanonicalizationAlgorithm is the graph canonicalization algorithm
+//
+// With V2 of the graph hash, this enum is no longer validated on-chain.
+// However, this enum SHOULD still be used and updated as a registry of known canonicalization
+// algorithms and all implementations should coordinate on these values.
 enum GraphCanonicalizationAlgorithm {
-
   // unspecified and invalid
   GRAPH_CANONICALIZATION_ALGORITHM_UNSPECIFIED = 0;
 
   // URDNA2015 graph hashing
   GRAPH_CANONICALIZATION_ALGORITHM_URDNA2015 = 1;
 }
 
-// GraphMerkleTree is the graph merkle tree type used for hashing, if any
+// GraphMerkleTree is the graph merkle tree type used for hashing, if any.
+//
+// With V2 of the graph hash, this enum is no longer validated on-chain.
+// However, this enum SHOULD still be used and updated as a registry of known merkle tree
+// types and all implementations should coordinate on these values.
 enum GraphMerkleTree {
-
   // unspecified and valid
   GRAPH_MERKLE_TREE_NONE_UNSPECIFIED = 0;
 }
 
 // ContentHashes contains list of content ContentHash.
 message ContentHashes {
-
   // data is a list of content hashes which the resolver claims to serve.
   repeated ContentHash content_hashes = 1;
-}
+}

scripts/protocgen.sh

@@ -30,6 +30,4 @@ cd ..
 cp -r github.com/regen-network/regen-ledger/* ./
 rm -rf github.com
 
-go mod tidy
-
 ./scripts/protocgen2.sh

scripts/protocgen2.sh

@@ -5,7 +5,7 @@ set -eo pipefail
 protoc_gen_install() {
   go install github.com/cosmos/cosmos-proto/cmd/protoc-gen-go-pulsar@latest #2>/dev/null
   go install google.golang.org/grpc/cmd/protoc-gen-go-grpc@latest #2>/dev/null
-  go install github.com/cosmos/cosmos-sdk/orm/cmd/protoc-gen-go-cosmos-orm@latest #2>/dev/null
+  go install github.com/cosmos/cosmos-sdk/orm/cmd/protoc-gen-go-cosmos-orm@v1.0.0-alpha.12 #2>/dev/null
 }
 
 protoc_gen_install

x/data/features/types_content_hash.feature

@@ -35,7 +35,7 @@ Feature: Types
     {}
     """
     When the content hash is validated
-    Then expect the error "content hash must be one of raw type or graph type: invalid request"
+    Then expect the error "exactly one of ContentHash's fields should be set: invalid request"
 
   Scenario: an error is returned if content hash includes both raw type and graph type
     Given the content hash
@@ -46,7 +46,7 @@ Feature: Types
     }
     """
     When the content hash is validated
-    Then expect the error "content hash must be one of raw type or graph type: invalid request"
+    Then expect the error "exactly one of ContentHash's fields should be set: invalid request"
 
   Scenario: an error is returned if raw content hash is empty
     Given the content hash