Updating pathology gx expectations

Sage-Bionetworks · Oct 29, 2024 · 7099261 · 7099261
1 parent 38447dd
commit 7099261
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 10 deletions.
diff --git a/gx_suite_definitions/pathology.ipynb b/gx_suite_definitions/pathology.ipynb
@@ -24,7 +24,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Create Expectation Suite for Biomarkers Data"
+    "# Create Expectation Suite for Pathology Data"
    ]
   },
   {
@@ -43,7 +43,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "biomarkers_data_file = syn.get(\"syn63540269\").path"
+    "pathology_data_file = syn.get(\"syn63644533\").path"
    ]
   },
   {
@@ -59,11 +59,11 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "df = pd.read_json(biomarkers_data_file)\n",
+    "df = pd.read_json(pathology_data_file)\n",
     "nested_columns = ['points']\n",
     "df = GreatExpectationsRunner.convert_nested_columns_to_json(df, nested_columns)\n",
     "validator = context.sources.pandas_default.read_dataframe(df)\n",
-    "validator.expectation_suite_name = \"biomarkers\""
+    "validator.expectation_suite_name = \"pathology\""
    ]
   },
   {
@@ -96,7 +96,7 @@
     "validator.expect_column_values_to_be_of_type(\"model\", \"str\")\n",
     "validator.expect_column_values_to_not_be_null(\"model\")\n",
     "# allows all alphanumeric characters, underscores, periods, and dashes\n",
-    "validator.expect_column_values_to_match_regex(\"type\", \"^[A-Za-z0-9\\s\\*_.-]+$\")\n"
+    "validator.expect_column_values_to_match_regex(\"type\", \"^[A-Za-z0-9\\s\\(\\)\\*_.-]+$\")\n"
    ]
   },
   {
@@ -109,7 +109,7 @@
     "validator.expect_column_values_to_be_of_type(\"type\", \"str\")\n",
     "validator.expect_column_values_to_not_be_null(\"type\")\n",
     "# allows all alphanumeric characters, underscores, periods, and dashes\n",
-    "validator.expect_column_values_to_match_regex(\"type\", \"^[A-Za-z0-9\\s_.-]+$\")"
+    "validator.expect_column_values_to_match_regex(\"type\", \"^[A-Za-z0-9\\s\\(\\)\\*_.-]+$\")"
    ]
   },
   {
@@ -122,7 +122,7 @@
     "validator.expect_column_values_to_be_of_type(\"units\", \"str\")\n",
     "validator.expect_column_values_to_not_be_null(\"units\")\n",
     "# allows all alphanumeric characters, underscores, periods, and dashes\n",
-    "validator.expect_column_values_to_match_regex(\"units\", \"^[A-Za-z0-9\\/\\s\\*_.-]+$\")"
+    "validator.expect_column_values_to_match_regex(\"units\", \"^[A-Za-z0-9\\s\\(\\)\\#\\/\\\\%*_.-]+$\")"
    ]
   },
   {

diff --git a/src/agoradatatools/great_expectations/gx/expectations/pathology.json b/src/agoradatatools/great_expectations/gx/expectations/pathology.json
@@ -55,7 +55,7 @@
       "expectation_type": "expect_column_values_to_match_regex",
       "kwargs": {
         "column": "type",
-        "regex": "^[A-Za-z0-9\\s_.-]+$"
+        "regex": "^[A-Za-z0-9\\s\\(\\)\\*_.-]+$"
       },
       "meta": {}
     },
@@ -93,7 +93,7 @@
       "expectation_type": "expect_column_values_to_match_regex",
       "kwargs": {
         "column": "units",
-        "regex": "^[A-Za-z0-9\\/\\s\\*_.-]+$"
+        "regex": "^[A-Za-z0-9\\s\\(\\)\\#\\/\\%*_.-]+$"
       },
       "meta": {}
     },
@@ -162,4 +162,4 @@
   "meta": {
     "great_expectations_version": "0.18.1"
   }
-}
+}