Add raw parser for db integrity checks (#3122)

* Parse leveldb entries

* Add VMDomainEdge

* Output as json

---------

Co-authored-by: jouzo <[email protected]>

lib/Cargo.toml

@@ -2,6 +2,7 @@
 members = [
     "ain-*",
     "cli",
+    "parser"
 ]
 
 default-members = [

lib/ain-dftx/Cargo.toml

@@ -11,3 +11,4 @@ bitcoin.workspace = true
 hex.workspace = true
 bitflags = "2.4.1"
 lazy_static.workspace = true
+serde.workspace = true

lib/ain-dftx/src/types/balance.rs

@@ -1,5 +1,6 @@
 use ain_macros::ConsensusEncoding;
 use bitcoin::{io, ScriptBuf, VarInt};
+use serde::Serialize;
 
 use super::common::CompactVec;
 

lib/ain-dftx/src/types/common.rs

@@ -4,6 +4,7 @@ use bitcoin::{
     consensus::{Decodable, Encodable},
     io::{self, ErrorKind},
 };
+use serde::Serialize;
 
 #[derive(Debug, PartialEq, Eq, Clone)]
 pub struct CompactVec<T>(Vec<T>);
@@ -110,7 +111,7 @@ impl Decodable for RawBytes {
 /// In the rust-bitcoin library, variable-length integers are implemented as CompactSize.
 /// See [issue #1016](https://github.com/rust-bitcoin/rust-bitcoin/issues/1016)
 
-#[derive(Debug, PartialEq, Eq, Clone)]
+#[derive(Debug, PartialEq, Eq, Clone, Serialize)]
 pub struct VarInt(pub u64);
 
 impl Encodable for VarInt {

lib/ain-dftx/src/types/price.rs

@@ -1,15 +1,15 @@
+use super::common::CompactVec;
 use ain_macros::ConsensusEncoding;
 use bitcoin::io;
-
-use super::common::CompactVec;
+use serde::Serialize;
 
 #[derive(ConsensusEncoding, Debug, PartialEq, Eq, Clone)]
 pub struct CurrencyPair {
     pub token: String,
     pub currency: String,
 }
 
-#[derive(ConsensusEncoding, Debug, PartialEq, Eq, Clone)]
+#[derive(ConsensusEncoding, Debug, PartialEq, Eq, Clone, Serialize)]
 pub struct TokenAmount {
     pub currency: String,
     pub amount: i64,

lib/ain-macros/src/lib.rs

@@ -115,6 +115,7 @@ pub fn ocean_endpoint(_attr: TokenStream, item: TokenStream) -> TokenStream {
 
     TokenStream::from(expanded)
 }
+
 #[proc_macro_derive(ConsensusEncoding)]
 pub fn consensus_encoding_derive(input: TokenStream) -> TokenStream {
     let input = parse_macro_input!(input as DeriveInput);

lib/parser/Cargo.toml

@@ -0,0 +1,12 @@
+[package]
+name = "parser"
+version = "0.1.0"
+edition = "2021"
+
+[dependencies]
+ain-macros.workspace = true
+ain-dftx.workspace = true
+bitcoin = { workspace = true, features = ["serde"] }
+hex.workspace = true
+serde_json.workspace = true
+serde.workspace = true

lib/parser/src/main.rs

@@ -0,0 +1,160 @@
+use ain_dftx::balance::TokenBalanceVarInt;
+use ain_dftx::common::VarInt;
+use ain_dftx::price::TokenAmount;
+use ain_macros::ConsensusEncoding;
+use bitcoin::{consensus::Decodable, ScriptBuf};
+use bitcoin::{io, Txid};
+use serde::Serialize;
+use std::io::BufRead;
+
+#[derive(Debug)]
+pub struct RawDbEntry {
+    pub prefix: u8,
+    pub key: Vec<u8>,
+    pub value: Vec<u8>,
+}
+
+impl RawDbEntry {
+    pub fn parse(input: &str) -> Result<Self, Box<dyn std::error::Error>> {
+        let parts: Vec<&str> = input.split_whitespace().collect();
+        if parts.len() != 3 {
+            return Err("Invalid input format: expected 3 space-separated parts".into());
+        }
+
+        let prefix = u8::from_str_radix(parts[0], 16)?;
+
+        let key = hex::decode(parts[1])?;
+        let value = hex::decode(parts[2])?;
+
+        Ok(RawDbEntry { prefix, key, value })
+    }
+}
+
+#[derive(ConsensusEncoding, Debug, Clone, PartialEq, Eq, Serialize)]
+pub struct BalanceKey {
+    pub owner: ScriptBuf,
+    pub token_id: VarInt,
+}
+
+#[derive(Debug, Clone, PartialEq, Eq, Serialize)]
+pub struct PrefixedData<K, V> {
+    pub key: K,
+    pub value: V,
+}
+
+#[derive(Debug, Clone, PartialEq, Eq, Serialize)]
+pub enum VMDomainEdge {
+    DVMToEVM,
+    EVMToDVM,
+}
+
+impl Decodable for VMDomainEdge {
+    fn consensus_decode<R: io::Read + ?Sized>(
+        reader: &mut R,
+    ) -> Result<Self, bitcoin::consensus::encode::Error> {
+        let edge = u8::consensus_decode(reader)?;
+        match edge {
+            1 => Ok(Self::DVMToEVM),
+            2 => Ok(Self::EVMToDVM),
+            _ => Err(bitcoin::consensus::encode::Error::ParseFailed(
+                "Unsupported VMDomainEdge",
+            )),
+        }
+    }
+}
+
+#[derive(Debug, Clone, PartialEq, Eq, Serialize)]
+struct UndoKey {
+    height: u32,
+    id: Txid,
+}
+
+#[derive(Debug, Clone, PartialEq, Eq, Serialize)]
+struct Undo {
+    data: Vec<u8>,
+}
+
+#[derive(Debug, Clone, PartialEq, Eq, Serialize)]
+pub enum Prefix {
+    ByBalance(PrefixedData<BalanceKey, i64>),
+    ByHeight(PrefixedData<ScriptBuf, u32>),
+    VMDomainTxEdge(PrefixedData<(VMDomainEdge, String), String>),
+    VMDomainBlockEdge(PrefixedData<(VMDomainEdge, String), String>),
+    Undo(PrefixedData<UndoKey, Undo>),
+}
+
+impl TryFrom<RawDbEntry> for Prefix {
+    type Error = bitcoin::consensus::encode::Error;
+
+    fn try_from(value: RawDbEntry) -> Result<Self, Self::Error> {
+        match value.prefix {
+            b'a' => Ok(Prefix::ByBalance(PrefixedData::try_from(value)?)),
+            b'b' => Ok(Prefix::ByHeight(PrefixedData::try_from(value)?)),
+            b'e' => Ok(Prefix::VMDomainTxEdge(PrefixedData::try_from(value)?)),
+            b'N' => Ok(Prefix::VMDomainBlockEdge(PrefixedData::try_from(value)?)),
+            _ => Err(bitcoin::consensus::encode::Error::ParseFailed(
+                "Unknown prefix",
+            )),
+        }
+    }
+}
+
+impl<K, V> TryFrom<RawDbEntry> for PrefixedData<K, V>
+where
+    K: Decodable,
+    V: Decodable,
+{
+    type Error = bitcoin::consensus::encode::Error;
+
+    fn try_from(value: RawDbEntry) -> Result<Self, Self::Error> {
+        let mut key_slice = value.key.as_slice();
+        let mut value_slice = value.value.as_slice();
+
+        let key = K::consensus_decode(&mut key_slice)?;
+        let value = V::consensus_decode(&mut value_slice)?;
+        Ok(Self { key, value })
+    }
+}
+
+fn process_line(line: &str) -> Result<(), Box<dyn std::error::Error>> {
+    let raw_entry = RawDbEntry::parse(line)?;
+
+    match Prefix::try_from(raw_entry) {
+        Ok(entry) => println!("{}", serde_json::to_string(&entry)?),
+        Err(_) => {}
+    }
+
+    Ok(())
+}
+
+fn main() -> Result<(), Box<dyn std::error::Error>> {
+    let stdin = std::io::stdin();
+    let reader = stdin.lock();
+
+    for line in reader.lines() {
+        let line = line?;
+        if !line.trim().is_empty() {
+            if let Err(e) = process_line(&line) {
+                eprintln!("Error processing line: {}", e);
+            }
+        }
+    }
+
+    Ok(())
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_parse() {
+        let input = "65 014031386566363362646661313837366264383735333933636363626133306430356465356238306261643161646536616131383931376261313939343532353163 4063343432376563623137373563333038346637313439383235313965616161323665636536643561653462303534626239393561326434666130346330353065";
+
+        let raw_entry = RawDbEntry::parse(input).unwrap();
+        println!("raw_entry : {:?}", raw_entry);
+        let entry = Prefix::try_from(raw_entry).unwrap();
+
+        println!("entry : {:?}", entry);
+    }
+}