beacon_chain/utils/shuffling: Introduce test vectors from sigp/shuffling_sandbox

Also:
 - return early if we shuffle an empty list
 - change RAND_MAX as per spec

beacon_chain/utils/shuffling/Cargo.toml

@@ -5,3 +5,6 @@ authors = ["Paul Hauner <paul@paulhauner.com>"]
 
 [dependencies]
 hashing = { path = "../hashing" }
+
+[dev-dependencies]
+yaml-rust = "0.4.2"

beacon_chain/utils/shuffling/src/lib.rs

@@ -25,9 +25,15 @@ pub fn shuffle<T>(
     -> Result<Vec<T>, ShuffleErr>
 {
     let mut rng = ShuffleRng::new(seed);
+
     if list.len() > rng.rand_max as usize {
         return Err(ShuffleErr::ExceedsListLength);
     }
+
+    if list.len() == 0 {
+        return Ok(list);
+    }
+
     for i in 0..(list.len() - 1) {
         let n = list.len() - i;
         let j = rng.rand_range(n as u32) as usize + i;
@@ -39,17 +45,43 @@ pub fn shuffle<T>(
 
 #[cfg(test)]
 mod tests {
+    extern crate yaml_rust;
+
     use super::*;
     use super::hashing::canonical_hash;
+    use std::fs::File;
+    use std::io::prelude::*;
+    use self::yaml_rust::yaml;
 
     #[test]
     fn test_shuffling() {
-        let seed = canonical_hash(b"4kn4driuctg8");
-        let list: Vec<usize> = (0..12).collect();
-        let s = shuffle(&seed, list).unwrap();
-        assert_eq!(
-            s,
-            vec![7, 3, 2, 5, 11, 9, 1, 0, 4, 6, 10, 8],
-        )
+        let mut file = File::open("./src/specs/shuffle_test_vectors.yaml").unwrap();
+        let mut yaml_str = String::new();
+
+        file.read_to_string(&mut yaml_str).unwrap();
+
+        let docs = yaml::YamlLoader::load_from_str(&yaml_str).unwrap();
+        let doc = &docs[0];
+        let test_cases = doc["test_cases"].as_vec();
+
+        for test_case in test_cases.unwrap() {
+            let input = test_case["input"].clone().into_vec().unwrap();
+            let output = test_case["output"].clone().into_vec().unwrap();
+            let seed_bytes = test_case["seed"].as_str().unwrap().as_bytes();
+            let mut seed;
+
+            if seed_bytes.len() > 0 {
+                seed = canonical_hash(seed_bytes);
+            } else {
+                seed = vec![];
+            }
+
+            let mut s = shuffle(&seed, input).unwrap();
+
+            assert_eq!(
+                s,
+                output,
+                );
+        }
     }
 }

beacon_chain/utils/shuffling/src/rng.rs

@@ -2,7 +2,7 @@ use super::hashing::canonical_hash;
 
 const SEED_SIZE_BYTES: usize = 32;
 const RAND_BYTES: usize = 3;      // 24 / 8
-const RAND_MAX: u32 = 16_777_216;   // 2**24
+const RAND_MAX: u32 = 16_777_215;   // 2 ** (rand_bytes * 8) - 1
 
 /// A pseudo-random number generator which given a seed
 /// uses successive blake2s hashing to generate "entropy".

beacon_chain/utils/shuffling/src/specs/shuffle_test_vectors.yaml

@@ -0,0 +1,169 @@
+# This file was generated with a modified version of shuffling_sandbox
+# python3 sandbox.py test_vectors
+#diff --git a/sandbox.py b/sandbox.py
+#index 99b0ba7..cd7cafe 100644
+#--- a/sandbox.py
+#+++ b/sandbox.py
+#@@ -126,13 +126,13 @@ elif args.method == "test_vectors":
+#     results = []
+# 
+#     seeds = [
+#-        b"",
+#-        blake("4kn4driuctg8".encode()),     # known to cause conflicts with old shuffler
+#-        blake("ytre1p".encode()),
+#-        blake("mytobcffnkvj".encode()),
+#-        blake("myzu3g7evxp5nkvj".encode()),
+#-        blake("xdpli1jsx5xb".encode()),
+#-        blake("oab3mbb3xe8qsx5xb".encode()),
+#+        "",
+#+        "4kn4driuctg8",     # known to cause conflicts with old shuffler
+#+        "ytre1p",
+#+        "mytobcffnkvj",
+#+        "myzu3g7evxp5nkvj",
+#+        "xdpli1jsx5xb",
+#+        "oab3mbb3xe8qsx5xb",
+#     ]
+#     lists = [
+#         [],
+#@@ -147,7 +147,8 @@ elif args.method == "test_vectors":
+# 
+#     for seed in seeds:
+#         for lst in lists:
+#-            output = shuffler(lst, seed)
+#+            blake_seed = blake(seed.encode()) if len(seed) > 0 else b""
+#+            output = shuffler(lst, blake_seed)
+#             results.append({"seed": seed, "input": lst, "output": output})
+# 
+#     body = {
+
+title: Shuffling Algorithm Tests
+summary: Test vectors for shuffling a list based upon a seed.
+test_suite: Shuffling
+
+test_cases:
+- input: []
+  output: []
+  seed: ''
+- input: [0]
+  output: [0]
+  seed: ''
+- input: [255]
+  output: [255]
+  seed: ''
+- input: [4, 6, 2, 6, 1, 4, 6, 2, 1, 5]
+  output: [1, 6, 4, 1, 6, 6, 2, 2, 4, 5]
+  seed: ''
+- input: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13]
+  output: [4, 7, 10, 13, 3, 1, 2, 9, 12, 6, 11, 8, 5]
+  seed: ''
+- input: [65, 6, 2, 6, 1, 4, 6, 2, 1, 5]
+  output: [1, 6, 65, 1, 6, 6, 2, 2, 4, 5]
+  seed: ''
+- input: []
+  output: []
+  seed: 4kn4driuctg8
+- input: [0]
+  output: [0]
+  seed: 4kn4driuctg8
+- input: [255]
+  output: [255]
+  seed: 4kn4driuctg8
+- input: [4, 6, 2, 6, 1, 4, 6, 2, 1, 5]
+  output: [6, 4, 2, 5, 4, 2, 6, 6, 1, 1]
+  seed: 4kn4driuctg8
+- input: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13]
+  output: [13, 1, 9, 8, 3, 10, 6, 2, 5, 12, 11, 4, 7]
+  seed: 4kn4driuctg8
+- input: [65, 6, 2, 6, 1, 4, 6, 2, 1, 5]
+  output: [6, 65, 2, 5, 4, 2, 6, 6, 1, 1]
+  seed: 4kn4driuctg8
+- input: []
+  output: []
+  seed: ytre1p
+- input: [0]
+  output: [0]
+  seed: ytre1p
+- input: [255]
+  output: [255]
+  seed: ytre1p
+- input: [4, 6, 2, 6, 1, 4, 6, 2, 1, 5]
+  output: [6, 2, 5, 1, 6, 4, 1, 2, 4, 6]
+  seed: ytre1p
+- input: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13]
+  output: [3, 8, 10, 4, 7, 11, 6, 1, 2, 5, 13, 9, 12]
+  seed: ytre1p
+- input: [65, 6, 2, 6, 1, 4, 6, 2, 1, 5]
+  output: [6, 2, 5, 1, 6, 4, 1, 2, 65, 6]
+  seed: ytre1p
+- input: []
+  output: []
+  seed: mytobcffnkvj
+- input: [0]
+  output: [0]
+  seed: mytobcffnkvj
+- input: [255]
+  output: [255]
+  seed: mytobcffnkvj
+- input: [4, 6, 2, 6, 1, 4, 6, 2, 1, 5]
+  output: [5, 6, 2, 1, 6, 4, 6, 4, 1, 2]
+  seed: mytobcffnkvj
+- input: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13]
+  output: [12, 4, 11, 6, 13, 10, 9, 2, 3, 7, 8, 1, 5]
+  seed: mytobcffnkvj
+- input: [65, 6, 2, 6, 1, 4, 6, 2, 1, 5]
+  output: [5, 6, 2, 1, 6, 65, 6, 4, 1, 2]
+  seed: mytobcffnkvj
+- input: []
+  output: []
+  seed: myzu3g7evxp5nkvj
+- input: [0]
+  output: [0]
+  seed: myzu3g7evxp5nkvj
+- input: [255]
+  output: [255]
+  seed: myzu3g7evxp5nkvj
+- input: [4, 6, 2, 6, 1, 4, 6, 2, 1, 5]
+  output: [6, 2, 6, 5, 4, 4, 1, 6, 2, 1]
+  seed: myzu3g7evxp5nkvj
+- input: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13]
+  output: [10, 12, 13, 3, 7, 11, 2, 4, 9, 8, 6, 5, 1]
+  seed: myzu3g7evxp5nkvj
+- input: [65, 6, 2, 6, 1, 4, 6, 2, 1, 5]
+  output: [6, 2, 6, 5, 65, 4, 1, 6, 2, 1]
+  seed: myzu3g7evxp5nkvj
+- input: []
+  output: []
+  seed: xdpli1jsx5xb
+- input: [0]
+  output: [0]
+  seed: xdpli1jsx5xb
+- input: [255]
+  output: [255]
+  seed: xdpli1jsx5xb
+- input: [4, 6, 2, 6, 1, 4, 6, 2, 1, 5]
+  output: [6, 2, 4, 1, 2, 6, 5, 1, 6, 4]
+  seed: xdpli1jsx5xb
+- input: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13]
+  output: [11, 8, 12, 9, 2, 1, 10, 4, 13, 5, 7, 3, 6]
+  seed: xdpli1jsx5xb
+- input: [65, 6, 2, 6, 1, 4, 6, 2, 1, 5]
+  output: [6, 2, 65, 1, 2, 6, 5, 1, 6, 4]
+  seed: xdpli1jsx5xb
+- input: []
+  output: []
+  seed: oab3mbb3xe8qsx5xb
+- input: [0]
+  output: [0]
+  seed: oab3mbb3xe8qsx5xb
+- input: [255]
+  output: [255]
+  seed: oab3mbb3xe8qsx5xb
+- input: [4, 6, 2, 6, 1, 4, 6, 2, 1, 5]
+  output: [2, 5, 1, 6, 1, 2, 6, 6, 4, 4]
+  seed: oab3mbb3xe8qsx5xb
+- input: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13]
+  output: [5, 13, 9, 7, 11, 10, 12, 2, 6, 8, 3, 1, 4]
+  seed: oab3mbb3xe8qsx5xb
+- input: [65, 6, 2, 6, 1, 4, 6, 2, 1, 5]
+  output: [2, 5, 1, 6, 1, 2, 6, 6, 65, 4]
+  seed: oab3mbb3xe8qsx5xb