Log original submission UID in duplication errors

womboai · Oct 18, 2024 · c92e05a · c92e05a
1 parent dddf9b5
commit c92e05a
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 4 deletions.
diff --git a/validator/weight_setting/deduplication.py b/validator/weight_setting/deduplication.py
@@ -1,4 +1,5 @@
 from dataclasses import dataclass
+from collections.abc import Iterator
 
 from imagehash import ImageHash
 
@@ -12,7 +13,7 @@ class PotentiallyDuplicateSubmissionInfo:
     block: int
 
 
-def find_duplicates(benchmark_info: list[PotentiallyDuplicateSubmissionInfo | None]):
+def find_duplicates(benchmark_info: list[PotentiallyDuplicateSubmissionInfo | None]) -> Iterator[tuple[int, int]]:
     duplicate_buckets: list[set[int]] = []
 
     for uid_a, benchmark_a in enumerate(benchmark_info):
@@ -43,4 +44,4 @@ def find_duplicates(benchmark_info: list[PotentiallyDuplicateSubmissionInfo | No
 
         for uid in bucket:
             if uid != oldest:
-                yield uid
+                yield uid, oldest
diff --git a/validator/weight_setting/validator.py b/validator/weight_setting/validator.py
@@ -846,9 +846,9 @@ async def do_step(self, block: int):
                 for uid, benchmark in enumerate(self.benchmarks)
             ]
 
-            for duplicate_uid in find_duplicates(benchmark_duplicate_info):
+            for duplicate_uid, original_uid in find_duplicates(benchmark_duplicate_info):
                 self.benchmarks[duplicate_uid] = None
-                self.invalid[duplicate_uid] = "Duplicate submission"
+                self.invalid[duplicate_uid] = f"Duplicate submission of UID '{original_uid}'"
 
             self.benchmarking = False
             self.step += 1