From 1fde6f90a3b4f44ba3fafa86ed973efa7e1c24da Mon Sep 17 00:00:00 2001
From: Lev Gorodetskiy <github@droserasprout.space>
Date: Tue, 1 Jun 2021 16:56:23 +0300
Subject: [PATCH] Fix matching in OperationCache

---
 src/dipdup/datasources/tzkt/cache.py          | 16 ++++++++++------
 src/dipdup/datasources/tzkt/datasource.py     | 10 ++++++++--
 .../test_tzkt/test_datasource.py              | 19 ++++++++++++++++++-
 3 files changed, 36 insertions(+), 9 deletions(-)

diff --git a/src/dipdup/datasources/tzkt/cache.py b/src/dipdup/datasources/tzkt/cache.py
index c5a82385e..24951fbcb 100644
--- a/src/dipdup/datasources/tzkt/cache.py
+++ b/src/dipdup/datasources/tzkt/cache.py
@@ -13,6 +13,7 @@
     OperationHandlerTransactionPatternConfig,
     OperationIndexConfig,
 )
+from dipdup.exceptions import ConfigurationError
 from dipdup.models import BigMapData, OperationData
 
 OperationGroup = namedtuple('OperationGroup', ('hash', 'counter'))
@@ -24,15 +25,18 @@ def __init__(self) -> None:
         super().__init__()
         self._logger = logging.getLogger(__name__)
         self._level: Optional[int] = None
-        self._indexes: Dict[str, OperationIndexConfig] = {}
+        self._indexes: List[OperationIndexConfig] = []
+        self._addresses: List[str] = []
         self._operations: Dict[OperationGroup, List[OperationData]] = {}
 
     async def add_index(self, index_config: OperationIndexConfig) -> None:
-        self._logger.debug('Adding index %s to cache', index_config)
         for contract in index_config.contract_configs:
-            if contract.address in self._indexes:
-                raise RuntimeError(f'Address `{contract.address}` used in multiple indexes')
-            self._indexes[contract.address] = index_config
+            if contract.address in self._addresses:
+                raise ConfigurationError(f'Address `{contract.address}` used in multiple indexes')
+            self._addresses.append(contract.address)
+
+        self._logger.debug('Adding index %s to cache', index_config)
+        self._indexes.append(index_config)
 
     async def add(self, operation: OperationData):
         self._logger.debug('Adding operation %s to cache (%s, %s)', operation.id, operation.hash, operation.counter)
@@ -91,7 +95,7 @@ async def on_match(
             self._logger.debug('Matching %s', key)
             matched = False
 
-            for index_config in self._indexes.values():
+            for index_config in self._indexes:
                 for handler_config in index_config.handlers:
                     operation_idx = 0
                     pattern_idx = 0
diff --git a/src/dipdup/datasources/tzkt/datasource.py b/src/dipdup/datasources/tzkt/datasource.py
index 2c5e4f64c..0ec90e392 100644
--- a/src/dipdup/datasources/tzkt/datasource.py
+++ b/src/dipdup/datasources/tzkt/datasource.py
@@ -110,6 +110,13 @@ async def run(self):
                 return
 
 
+def dedup_operations(operations: List[Dict[str, Any]]) -> List[Dict[str, Any]]:
+    return sorted(
+        list(({op['id']: op for op in operations}).values()),
+        key=lambda op: op['id'],
+    )
+
+
 class OperationFetcher:
     def __init__(
         self,
@@ -250,8 +257,7 @@ async def fetch_operations_by_level(self):
             while self._head <= head:
                 if self._head in self._operations:
                     operations = self._operations.pop(self._head)
-                    operations = sorted(list(({op['id']: op for op in operations}).values()), key=lambda op: op['id'])
-                    yield self._head, operations
+                    yield self._head, dedup_operations(operations)
                 self._head += 1
 
             if all(list(self._fetched.values())):
diff --git a/tests/test_dipdup/test_datasources/test_tzkt/test_datasource.py b/tests/test_dipdup/test_datasources/test_tzkt/test_datasource.py
index 1e9e4ba61..6807b0bb0 100644
--- a/tests/test_dipdup/test_datasources/test_tzkt/test_datasource.py
+++ b/tests/test_dipdup/test_datasources/test_tzkt/test_datasource.py
@@ -17,7 +17,7 @@
     OperationIndexConfig,
     OperationType,
 )
-from dipdup.datasources.tzkt.datasource import TzktDatasource
+from dipdup.datasources.tzkt.datasource import TzktDatasource, dedup_operations
 from dipdup.models import IndexType, OperationData, OperationHandlerContext, State, TransactionContext
 from dipdup.utils import tortoise_wrapper
 
@@ -195,3 +195,20 @@ async def test_on_operation_match_with_storage(self):
                 callback_mock.await_args[0][1].storage.proposals['e710c1a066bbbf73692168e783607996785260cec4d60930579827298493b8b9'],
                 Proposals,
             )
+
+    async def test_dedup_operations(self) -> None:
+        operations = [
+            {'id': 5},
+            {'id': 3},
+            {'id': 3},
+            {'id': 1},
+        ]
+        operations = dedup_operations(operations)
+        self.assertEqual(
+            [
+                {'id': 1},
+                {'id': 3},
+                {'id': 5},
+            ],
+            operations,
+        )