Added integration test for repseudo from stable id to default encrypt…

…ion (#368) * Added integration test for repseudo from stable id to default encryption * Bump new version --------- Co-authored-by: Bjørn-Andre Skaar <[email protected]>
statisticsnorway · Apr 4, 2024 · aa50dc7 · aa50dc7
1 parent a64af93
commit aa50dc7
Show file tree

Hide file tree

Showing 11 changed files with 115 additions and 98 deletions.
diff --git a/pyproject.toml b/pyproject.toml
@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "dapla-toolbelt-pseudo"
-version = "1.4.0"
+version = "1.5.0"
 description = "Pseudonymization extensions for Dapla"
 authors = ["Dapla Developers <[email protected]>"]
 license = "MIT"

diff --git a/tests/conftest.py b/tests/conftest.py
@@ -86,12 +86,10 @@ def df_personer_fnr_ff31_encrypted() -> pl.DataFrame:
 
 
 @pytest.fixture
-def df_personer_daead_encrypted_ssb_common_key_2() -> pl.DataFrame:
-    CSV_FILE = (
-        "tests/data/personer_pseudonymized_default_encryption_ssb_common_key_2.csv"
-    )
-    return pl.read_csv(
-        CSV_FILE,
+def df_personer_daead_encrypted_ssb_common_key_1() -> pl.DataFrame:
+    JSON_FILE = "tests/data/personer_pseudonymized_daead_ssb_common_key_1.json"
+    return pl.read_json(
+        JSON_FILE,
         schema={
             "fnr": pl.String,
             "fornavn": pl.String,
@@ -103,8 +101,8 @@ def df_personer_daead_encrypted_ssb_common_key_2() -> pl.DataFrame:
 
 
 @pytest.fixture
-def df_personer_pseudo_stable_id() -> pl.DataFrame:
-    JSON_FILE = "tests/data/person_3_sid_deid.json"
+def df_personer_daead_encrypted_ssb_common_key_2() -> pl.DataFrame:
+    JSON_FILE = "tests/data/personer_pseudonymized_daead_ssb_common_key_2.json"
     return pl.read_json(
         JSON_FILE,
         schema={
@@ -118,8 +116,8 @@ def df_personer_pseudo_stable_id() -> pl.DataFrame:
 
 
 @pytest.fixture
-def df_personer_depseudo_stable_id() -> pl.DataFrame:
-    JSON_FILE = "tests/data/person_3_sid.json"
+def df_personer_pseudo_stable_id_daead_encrypted_ssb_common_key_2() -> pl.DataFrame:
+    JSON_FILE = "tests/data/personer_pseudonymized_sid_daead_ssb_common_key_2.json"
     return pl.read_json(
         JSON_FILE,
         schema={

diff --git a/tests/data/kontaktopplysninger.json b/tests/data/kontaktopplysninger.json
diff --git a/tests/data/person_3_sid.json b/tests/data/person_3_sid.json
diff --git a/tests/data/person_3_sid_deid.json b/tests/data/person_3_sid_deid.json
diff --git a/tests/data/personer_pseudonymized_daead_ssb_common_key_1.json b/tests/data/personer_pseudonymized_daead_ssb_common_key_1.json
@@ -0,0 +1,23 @@
+[
+  {
+    "fnr": "AWIRfKLSNfR0ID+wBzogEcUT7JQPayk7Gosij6SXr8s=",
+    "fornavn": "AWIRfKKWWRC1hURqsYw4S/h/NitvuP6bO/R7",
+    "etternavn": "AWIRfKIKAGiRoGTd/Cid5gxsIDx4H1ya6w==",
+    "kjonn": "M",
+    "fodselsdato": "020995"
+  },
+  {
+    "fnr": "AWIRfKKLagk0LqYCKpiC4xfPkHqIWGVfc3wg5gUwRNE=",
+    "fornavn": "AWIRfKJuYBaBQIXIprRO9UFDXNLd4YXcHtY=",
+    "etternavn": "AWIRfKLzOfzOw+Bdo9zIa4savOOeAiEr",
+    "kjonn": "M",
+    "fodselsdato": "060970"
+  },
+  {
+    "fnr": "AWIRfKIzL1T9iZqt+pLjNbHMsLa0aKSszsRrLiLSAAg=",
+    "fornavn": "AWIRfKKsIDQgWLnpsSln38z1RSfHjjL8FS4=",
+    "etternavn": "AWIRfKIKAGiRoGTd/Cid5gxsIDx4H1ya6w==",
+    "kjonn": "M",
+    "fodselsdato": "180999"
+  }
+]
diff --git a/tests/data/personer_pseudonymized_daead_ssb_common_key_2.json b/tests/data/personer_pseudonymized_daead_ssb_common_key_2.json
@@ -0,0 +1,23 @@
+[
+  {
+    "fnr": "AXzg5m91ZV4nzWpqYtdo+jfHIE40fLbDxug3Lr8Tamg=",
+    "fornavn": "AXzg5m+lFA0OcH5L1eONOgzZCItnUaXIUshA",
+    "etternavn": "AXzg5m+e0GDK5mhcOknMY71YcP3BxMlnOA==",
+    "kjonn": "M",
+    "fodselsdato": "020995"
+  },
+  {
+    "fnr": "AXzg5m9Ec/KlRWJBHfpwIsksDwmA7F/jhojyhBU8pec=",
+    "fornavn": "AXzg5m/zZww19slv39H+s/wiILYBaeYqPEk=",
+    "etternavn": "AXzg5m9FgnhAZErrWTmEQaRZCm4/t1AG",
+    "kjonn": "M",
+    "fodselsdato": "060970"
+  },
+  {
+    "fnr": "AXzg5m/FALW/hNStYCElfyMR7/9rD398dEolAsISK6A=",
+    "fornavn": "AXzg5m92Bx8qcV7AEXa7m4IAbvUyk2iPaD4=",
+    "etternavn": "AXzg5m+e0GDK5mhcOknMY71YcP3BxMlnOA==",
+    "kjonn": "M",
+    "fodselsdato": "180999"
+  }
+]
diff --git a/tests/data/personer_pseudonymized_sid_daead_ssb_common_key_2.json b/tests/data/personer_pseudonymized_sid_daead_ssb_common_key_2.json
@@ -0,0 +1,23 @@
+[
+  {
+    "fnr": "jJuuj0i",
+    "fornavn": "AXzg5m+lFA0OcH5L1eONOgzZCItnUaXIUshA",
+    "etternavn": "AXzg5m+e0GDK5mhcOknMY71YcP3BxMlnOA==",
+    "kjonn": "M",
+    "fodselsdato": "020995"
+  },
+  {
+    "fnr": "ylc9488",
+    "fornavn": "AXzg5m/zZww19slv39H+s/wiILYBaeYqPEk=",
+    "etternavn": "AXzg5m9FgnhAZErrWTmEQaRZCm4/t1AG",
+    "kjonn": "M",
+    "fodselsdato": "060970"
+  },
+  {
+    "fnr": "yeLfkaL",
+    "fornavn": "AXzg5m92Bx8qcV7AEXa7m4IAbvUyk2iPaD4=",
+    "etternavn": "AXzg5m+e0GDK5mhcOknMY71YcP3BxMlnOA==",
+    "kjonn": "M",
+    "fodselsdato": "180999"
+  }
+]
diff --git a/tests/integration/test_integration_deseudonymize.py b/tests/integration/test_integration_deseudonymize.py
@@ -26,15 +26,18 @@ def test_depseudonymize_default_encryption(
 @integration_test()
 def test_depseudonymize_sid(
     setup: Generator[None, None, None],
+    df_personer_pseudo_stable_id_daead_encrypted_ssb_common_key_2: pl.DataFrame,
     df_personer: pl.DataFrame,
-    df_personer_pseudo_stable_id: pl.DataFrame,
-    df_personer_depseudo_stable_id: pl.DataFrame,
 ) -> None:
     result = (
-        Depseudonymize.from_polars(df_personer_pseudo_stable_id)
+        Depseudonymize.from_polars(
+            df_personer_pseudo_stable_id_daead_encrypted_ssb_common_key_2
+        )
         .on_fields("fnr")
         .with_stable_id()
+        .on_fields("fornavn", "etternavn")
+        .with_default_encryption(custom_key="ssb-common-key-2")
         .run()
         .to_polars()
     )
-    assert result.equals(df_personer_depseudo_stable_id)
+    assert result.equals(df_personer)
diff --git a/tests/integration/test_integration_pseudonymize.py b/tests/integration/test_integration_pseudonymize.py
@@ -1,14 +1,6 @@
-import inspect
-import json
-import typing as t
 from collections.abc import Generator
-from functools import partial
-from pathlib import Path
-from unittest.mock import ANY
 
-import pandas as pd
 import polars as pl
-import pytest
 
 from dapla_pseudo import Pseudonymize
 from tests.integration.utils import get_calling_function_name

diff --git a/tests/integration/test_integration_repseudonymize.py b/tests/integration/test_integration_repseudonymize.py
@@ -21,21 +21,48 @@ def test_repseudonymize_from_default_encryption_to_fpe(
         .run()
         .to_polars()
     )
+
     assert result_df.equals(df_personer_fnr_ff31_encrypted)
 
 
 @integration_test()
 def test_repseudonymize_change_keys(
     setup: Generator[None, None, None],
     df_personer_daead_encrypted_ssb_common_key_2: pl.DataFrame,
-    df_personer_fnr_daead_encrypted: pl.DataFrame,
+    df_personer_daead_encrypted_ssb_common_key_1: pl.DataFrame,
 ) -> None:
     result = (
-        Repseudonymize.from_polars(df_personer_fnr_daead_encrypted)
-        .on_fields("fnr")
+        Repseudonymize.from_polars(df_personer_daead_encrypted_ssb_common_key_1)
+        .on_fields("fnr", "fornavn", "etternavn")
         .from_default_encryption()
         .to_default_encryption(custom_key="ssb-common-key-2")
         .run()
         .to_polars()
     )
+
     assert result.equals(df_personer_daead_encrypted_ssb_common_key_2)
+
+
+@integration_test()
+def test_repseudonymize_from_sid_to_daead(
+    setup: Generator[None, None, None],
+    df_personer: pl.DataFrame,
+    df_personer_daead_encrypted_ssb_common_key_1: pl.DataFrame,
+    df_personer_daead_encrypted_ssb_common_key_2: pl.DataFrame,
+    df_personer_pseudo_stable_id_daead_encrypted_ssb_common_key_2: pl.DataFrame,
+) -> None:
+    result = (
+        Repseudonymize.from_polars(
+            df_personer_pseudo_stable_id_daead_encrypted_ssb_common_key_2
+        )
+        .on_fields("fnr")
+        .from_stable_id()
+        .to_default_encryption()
+        .on_fields("fornavn", "etternavn")
+        .from_default_encryption(custom_key="ssb-common-key-2")
+        .to_default_encryption(custom_key="ssb-common-key-1")
+        .run()
+        .to_polars()
+    )
+
+    assert result.equals(df_personer_daead_encrypted_ssb_common_key_1)