Merge branch 'staging' into prod

CHANGELOG.md

@@ -3,6 +3,18 @@
 
 All notable changes to this project will be documented in this file.
 
+## 1.3.0 - 2023-05-03
+### Breaking Changes:
+- No Change.
+
+### New features:
+- No Change.
+
+### Bug fixes:
+- Row update with a bit broken. Users can now set row values equal to null with JSON null value.
+- Users can also now update any individual row value, previously rows with not "null" would set "required" and would mean that updates would also require said values.
+
+
 ## 1.3.0 - 2023-03-09
 ### Breaking Changes:
 - No Change.

Makefile

@@ -22,13 +22,13 @@ build-core:
 # @echo "Now migrating"
 # @docker-compose run api python /home/tapis/manage.py migrate_schemas --shared
 
-local-deploy: build-core
+up: build-core
 	@echo "Running docker-compose up"
 	@docker-compose up -d api
 
 
 # Running the pgrest/test.py file
-test: local-deploy
+test:
 	@docker-compose run api python /home/tapis/manage.py test -v 2
 
 

pgrest/db_transactions/table_data.py

@@ -246,8 +246,10 @@ def update_row_with_pk(table_name, pk_id, data, tenant, primary_key, db_instance
     logger.info(f"Updating row with pk \'{pk_id}\' in {tenant}.{table_name}...")
     command = f"UPDATE {tenant}.{table_name} SET"
     for col in data:
-        if type(data[col]) == str:
+        if isinstance(data[col], str):
             command = f"{command} {col} = (\'{data[col]}\'), "
+        elif isinstance(data[col], type(None)):
+            command = f"{command} {col} = (NULL), "
         else:
             command = f"{command} {col} = ({data[col]}), "
     if type(pk_id) == 'int' or type(pk_id) == 'float':
@@ -277,8 +279,10 @@ def update_rows_with_where(table_name, data, tenant, db_instance, search_params=
     logger.info(f"Updating rows in {tenant}.{table_name}...")
     command = f"UPDATE {tenant}.{table_name} SET"
     for col in data:
-        if type(data[col]) == str:
+        if isinstance(data[col], str):
             command = f"{command} {col} = (\'{data[col]}\'), "
+        elif isinstance(data[col], type(None)):
+            command = f"{command} {col} = (NULL), "
         else:
             command = f"{command} {col} = ({data[col]}), "
 

pgrest/tests.py

@@ -343,6 +343,52 @@ def test_update_row(self):
                                    content_type='application/json')
         self.assertEqual(response.status_code, 200)
 
+    def test_update_row_with_null_value(self):
+        # We should be able to change a row value from filled to null without issue, if it's nullable. 
+        # first, we need to create row
+        root_url = self.init_resp_1["result"]["root_url"]
+        table_name = self.init_resp_1["result"]["table_name"]
+        data = {"col_one": "hello", "col_two": 100, "col_three": 90, "col_four": False, "col_five": "hehe"}
+        response = self.client.post(f'/v3/pgrest/data/{root_url}',
+                                    data=json.dumps({"data": data}),
+                                    **auth_headers,
+                                    content_type='application/json')
+        self.assertEqual(response.status_code, 200)
+        response = self.client.get(f'/v3/pgrest/data/{root_url}', **auth_headers)
+        self.assertEqual(len(response.json()["result"]), 1)
+        self.assertEqual(response.status_code, 200)
+
+        # now, update
+        data = {"col_two": None}
+        row_id = response.json()["result"][0]['_pkid']
+        response = self.client.put(f'/v3/pgrest/data/{root_url}/{row_id}',
+                                   **auth_headers,
+                                   data=json.dumps({"data": data}),
+                                   content_type='application/json')
+        self.assertEqual(response.status_code, 200)
+
+    def test_update_row_with_null_value_on_non_nullable_column_400(self):
+        root_url = self.init_resp_1["result"]["root_url"]
+        table_name = self.init_resp_1["result"]["table_name"]
+        data = {"col_one": "hello", "col_two": 100, "col_three": 90, "col_four": False, "col_five": "hehe"}
+        response = self.client.post(f'/v3/pgrest/data/{root_url}',
+                                    data=json.dumps({"data": data}),
+                                    **auth_headers,
+                                    content_type='application/json')
+        self.assertEqual(response.status_code, 200)
+        response = self.client.get(f'/v3/pgrest/data/{root_url}', **auth_headers)
+        self.assertEqual(len(response.json()["result"]), 1)
+        self.assertEqual(response.status_code, 200)
+
+        # now, update
+        data = {"col_four": None}
+        row_id = response.json()["result"][0]['_pkid']
+        response = self.client.put(f'/v3/pgrest/data/{root_url}/{row_id}',
+                                   **auth_headers,
+                                   data=json.dumps({"data": data}),
+                                   content_type='application/json')
+        self.assertEqual(response.status_code, 400)
+
     def test_update_row_wrong_data_type_400(self):
         # first, we need to create row
         root_url = self.init_resp_1["result"]["root_url"]
@@ -356,6 +402,7 @@ def test_update_row_wrong_data_type_400(self):
         response = self.client.get(f'/v3/pgrest/data/{root_url}', **auth_headers)
         self.assertEqual(len(response.json()["result"]), 1)
         self.assertEqual(response.status_code, 200)
+
         # now, update
         data = {"col_two": "haha"}
         row_id = response.json()["result"][0]['_pkid']
@@ -388,6 +435,7 @@ def test_update_row_nonexistent_column_400(self):
         response = self.client.get(f'/v3/pgrest/data/{root_url}', **auth_headers)
         self.assertEqual(len(response.json()["result"]), 1)
         self.assertEqual(response.status_code, 200)
+
         # now, update
         data = {"col_where": 30}
         row_id = response.json()["result"][0]['_pkid']
@@ -1405,17 +1453,17 @@ def test_all_search_parameters(self):
     # Nonexistence checks
     def test_get_nonexistent_view(self):
         response = self.client.get(f'/v3/pgrest/views/this_view_does_not_exist', **auth_headers)
-        print(response)
+        #print(response)
         self.assertEqual(response.status_code, 404)
 
     def test_get_nonexistent_manage_view(self):
         response = self.client.get(f'/v3/pgrest/manage/views/22', **auth_headers)
-        print(response)
+        #print(response)
         self.assertEqual(response.status_code, 404)
 
     def test_delete_nonexistent_view(self):
         response = self.client.delete(f'/v3/pgrest/manage/views/22', **auth_headers)
-        print(response)
+        #print(response)
         self.assertEqual(response.status_code, 404)
 
     def test_raw_sql_view_creation(self):

pgrest/utils.py

@@ -73,31 +73,9 @@ def create_validate_schema(columns, tenant, existing_enum_names):
     schema_update = dict()
     schema_create = dict()
 
-    for key, key_info in columns.items():
-        key_type = key_info["data_type"].lower()
-        info_dict = dict()
-        if key_type in ["varchar", "char"]:
-            try:
-                val_length = int(key_info["char_len"])
-                info_dict.update({"type": "string", "maxlength": val_length})
-            except KeyError:
-                raise KeyError(f"Unable to create table. {key_type} data types requires char_len field.")
-        elif key_type == "text":
-            info_dict.update({"type": "string"})
-        elif key_type == "serial":
-            info_dict.update({"type": "integer"})
-        elif key_type == "date":
-            info_dict.update({"type": "string"})
-        elif key_type == "timestamp":
-            info_dict.update({"type": "string"})
-        elif '[]' in key_type:
-            info_dict.update({"type": "list"})
-        elif key_type in existing_enum_names or f'{tenant}.{key_type}' in existing_enum_names:
-            info_dict.update({"type": "string"})
-        else:
-            info_dict.update({"type": key_type})
-        schema_update[key] = info_dict
-
+    # We once had different update and create schemas. I couldn't figure out a valid reason for this.
+    # Only difference is schema_update should never have "required" field. As users should always be able
+    # to update a row's value. 
     for key, key_info in columns.items():
         key_type = key_info["data_type"]
         info_dict = dict()
@@ -120,13 +98,17 @@ def create_validate_schema(columns, tenant, existing_enum_names):
             info_dict.update({"type": key_type})
 
         if "null" in key_info.keys():
-            if not key_info["null"]:
-                info_dict.update({"required": True,
-                                  "nullable": True})
-            else:
+            if key_info["null"]: #true
                 info_dict.update({"required": False,
                                   "nullable": True})
-        schema_create[key] = info_dict
+            else: #false
+                info_dict.update({"required": True,
+                                  "nullable": False})
+        schema_create[key] = info_dict.copy()
+
+        # Update schema should never have required fields. Users should always be able to update a row's value.
+        info_dict.update({"required": False})
+        schema_update[key] = info_dict.copy()
 
     return schema_create, schema_update