diff --git a/tests/transfer_analysis_common.py b/tests/transfer_analysis_common.py
index a3d8a3e..ff6fc6f 100644
--- a/tests/transfer_analysis_common.py
+++ b/tests/transfer_analysis_common.py
@@ -27,15 +27,23 @@
 
 @dataclass(frozen=True, eq=True)
 class _Test:
+    # English description of the test.
     description: str
+    # List of universal-application, input transactions.
     input: List[AbstractTransactionDescriptor]
+    # Dictionary of expected per-wallet, input transactions: this is the wanted output of transfer analysis.
     want: Dict[Account, List[AbstractTransactionDescriptor]]
+    # Dictionary of expected actual amounts: this contains the actual crypto amounts of the per-wallet in-transactions. It is useful to understand
+    # where the funds are, because in per-wallet application there can be multiple in-transactions covering the same funds in different exchanges:
+    # artificial in-transactions are created when processing intra-transactions to model the reception of funds).
     want_amounts: Dict[Account, Dict[str, int]]
+    # If not empty, the test is expected to raise an error and this is the expected error message.
     want_error: str
 
 
 class AbstractTransferAnalysis(AbstractTestTransactionProcessing):
 
+    # Run a transfer analysis test.
     def _run_test(self, test: _Test, transfer_semantics: AbstractAccountingMethod) -> None:
         print(f"\nDescription: {test.description:}\nTransfer:    {transfer_semantics}\nWant error:  {'yes' if test.want_error else 'no'}")
         configuration = Configuration("./config/test_data.ini", US())