Stronger type safety around IPFS CIDs

Cargo.toml

@@ -1,3 +1,44 @@
 [workspace]
 resolver = "2"
 members = ["crates/*"]
+
+[workspace.dependencies]
+anyhow = "1"
+async-graphql = "7"
+async-graphql-axum = "7"
+async-trait = "0.1.52"
+axum = "0.7"
+bigdecimal = "0.4"
+chrono = "0.4"
+cid = "0.11"
+clap = "4"
+derive_more = "0.99"
+diesel = "2"
+diesel-async = "0.4"
+diesel_async_migrations = "0.12"
+futures = "0.3.18"
+graphql_client = "0.13"
+hex = "0.4.3"
+itertools = "0.12"
+nanoid = "0.4.0"
+num-traits = "0.2"
+once_cell = "1.9.0"
+prometheus = { version = "0.13", default-features = false }
+prometheus_exporter = "0.8.5"
+quickcheck = "1"
+quickcheck_macros = "1"
+rand = "0.8.4"
+reqwest = "0.11"
+schemars = "0.8"
+serde = "1"
+serde_json = "1"
+serde_yaml = "0.9"
+testcontainers = "0.15"
+testcontainers-modules = "0.3"
+thiserror = "1"
+tokio = "1.14.0"
+tracing = "0.1.29"
+tracing-subscriber = "0.3.2"
+tracing-test = "0.2.1"
+url = "2.5"
+uuid = "1"

crates/autogen_config_json_schema/Cargo.toml

@@ -5,5 +5,5 @@ edition = "2021"
 
 [build-dependencies]
 graphix_lib = { path = "../graphix_lib" }
-schemars = "0.8"
-serde_json = "1"
+schemars = { workspace = true }
+serde_json = { workspace = true }

crates/autogen_config_json_schema/build.rs

@@ -11,13 +11,11 @@ fn main() -> std::io::Result<()> {
     // See <https://doc.rust-lang.org/cargo/reference/build-scripts.html#change-detection>.
     println!("cargo:rerun-if-changed=build.rs");
 
-    let path = env::current_dir()?.join("schema.json");
-
-    let mut f = File::create(&path)?;
+    let out_path = env::current_dir()?.join("schema.json");
+    let mut f = File::create(&out_path)?;
 
     let schema = schema_for!(Config);
     f.write_all(serde_json::to_string_pretty(&schema).unwrap().as_bytes())?;
 
-    println!("Updated: {}", path.display());
     Ok(())
 }

crates/autogen_graphql_schema/Cargo.toml

@@ -4,5 +4,5 @@ version = "0.1.0"
 edition = "2021"
 
 [build-dependencies]
-async-graphql = "7"
+async-graphql = { workspace = true }
 graphix_lib = { path = "../graphix_lib" }

crates/autogen_graphql_schema/api_schema.graphql

@@ -204,6 +204,8 @@ type IndexerNetworkSubgraphMetadata {
 }
 
 
+scalar IpfsCid
+
 type MutationRoot {
 	"""
 	Launches a divergence investigation, which is a process of comparing
@@ -313,10 +315,9 @@ input PoisQuery {
 	"""
 	network: String
 	"""
-	Restricts the query to PoIs for these given subgraph deployments (by
-	hex-encoded IPFS CID with '0x' prefix).
+	Restricts the query to PoIs for these given subgraph deployment IDs.
 	"""
-	deployments: [String!]!
+	deployments: [IpfsCid!]!
 	"""
 	Restricts the query to PoIs that were collected in the given block
 	range.
@@ -356,7 +357,7 @@ type QueryRoot {
 		"""
 		The network name of the subgraph deployments to fetch
 		"""
-		networkName: String,		name: String,		ipfsCid: String,
+		networkName: String,		name: String,		ipfsCid: IpfsCid,
 		"""
 		Upper limit on the number of shown results.
 		"""
@@ -388,7 +389,7 @@ type QueryRoot {
 		"""
 		Restricts the query to PoIs for these given subgraph deployments (by hex-encoded IPFS CID with '0x' prefix).
 		"""
-		deployments: [String!]! = [],
+		deployments: [IpfsCid!]! = [],
 		"""
 		Restricts the query to PoIs that were collected in the given block range.
 		"""
@@ -424,7 +425,7 @@ type SubgraphDeployment {
 	"""
 	IPFS CID of the subgraph deployment.
 	"""
-	cid: String!
+	cid: IpfsCid!
 	"""
 	Human-readable name of the subgraph deployment, if present.
 	"""

crates/common_types/Cargo.toml

@@ -4,16 +4,17 @@ version = "0.1.0"
 edition = "2021"
 
 [dependencies]
-anyhow = "1"
-async-graphql = { version = "7", features = ["chrono", "uuid"] }
-chrono = "0.4"
-diesel = "2"
-hex = "0.4"
-quickcheck = "1"
-schemars = "0.8"
-serde = { version = "1", features = ["derive"] }
-serde_json = "1"
-uuid = "1"
+anyhow = { workspace = true }
+async-graphql = { workspace = true, features = ["chrono", "uuid"] }
+chrono = { workspace = true }
+cid = { workspace = true, features = ["serde", "arb"] }
+diesel = { workspace = true }
+hex = { workspace = true }
+quickcheck = { workspace = true }
+schemars = { workspace = true }
+serde = { workspace = true, features = ["derive"] }
+serde_json = { workspace = true }
+uuid = { workspace = true }
 
 [dev-dependencies]
-quickcheck_macros = "1"
+quickcheck_macros = { workspace = true }

crates/common_types/src/hex_string.rs

@@ -5,6 +5,7 @@ use std::str::FromStr;
 use diesel::backend::Backend;
 use diesel::deserialize::{FromSql, FromSqlRow};
 use diesel::expression::AsExpression;
+use diesel::pg::Pg;
 use diesel::serialize::ToSql;
 use diesel::sql_types;
 use hex::FromHex;
@@ -24,7 +25,7 @@ use serde::{Deserialize, Serialize};
 // TODO: The fact that we SQL-encode all kinds of hex strings, even fixed-length
 // ones, as variable-length byte sequences is a bit of a wart. Not that big of a
 // deal, but maybe there's a better way to do this.
-#[diesel(sql_type = ::diesel::sql_types::Binary)]
+#[diesel(sql_type = sql_types::Binary)]
 pub struct HexString<T>(pub T);
 
 impl<T: ToOwned> HexString<T> {
@@ -102,29 +103,25 @@ impl<T> schemars::JsonSchema for HexString<T> {
     }
 }
 
-impl<T, Db> ToSql<sql_types::Binary, Db> for HexString<T>
+impl<T> ToSql<sql_types::Binary, Pg> for HexString<T>
 where
     T: AsRef<[u8]> + Debug,
-    Db: Backend,
-    [u8]: ToSql<sql_types::Binary, Db>,
 {
     fn to_sql<'b>(
         &'b self,
-        out: &mut diesel::serialize::Output<'b, '_, Db>,
+        out: &mut diesel::serialize::Output<'b, '_, Pg>,
     ) -> diesel::serialize::Result {
-        ToSql::<sql_types::Binary, Db>::to_sql(self.0.as_ref(), out)
+        ToSql::<sql_types::Binary, Pg>::to_sql(self.0.as_ref(), out)
     }
 }
 
-impl<T, Db> FromSql<sql_types::Binary, Db> for HexString<T>
+impl<T> FromSql<sql_types::Binary, Pg> for HexString<T>
 where
     T: TryFrom<Vec<u8>>,
     T::Error: Debug,
-    Db: Backend,
-    Vec<u8>: FromSql<sql_types::Binary, Db>,
 {
-    fn from_sql(bytes: <Db as Backend>::RawValue<'_>) -> diesel::deserialize::Result<Self> {
-        T::try_from(Vec::from_sql(bytes)?)
+    fn from_sql(bytes: <Pg as Backend>::RawValue<'_>) -> diesel::deserialize::Result<Self> {
+        T::try_from(FromSql::<sql_types::Binary, Pg>::from_sql(bytes)?)
             .map(HexString)
             .map_err(|e| anyhow::anyhow!("{:?}", e).into())
     }