Fix test caching - only cache pagination DB to avoid max models errors

src/mapp/test.py

@@ -8,6 +8,7 @@
 import multiprocessing
 import time
 import shutil
+import jwt
 
 from pathlib import Path
 from copy import deepcopy
@@ -127,6 +128,43 @@ def env_to_string(env:dict) -> str:
             out += f'{key}={value}\n'
     return out
 
+def login_cached_user(cmd:list[str], ctx:dict, user_name:str, email:str, password:str='testpass123') -> dict:
+    """Login to a cached user and return user data with env context"""
+    login_params = {'email': email, 'password': password}
+    login_cmd = cmd + ['auth', 'login-user', 'run', json.dumps(login_params), '--show', '--no-session']
+    result = subprocess.run(login_cmd, capture_output=True, text=True, env=ctx, timeout=10)
+
+    if result.returncode != 0:
+        raise RuntimeError(f'Error logging in cached user {user_name}:\n{result.stdout + result.stderr}')
+
+    login_response = json.loads(result.stdout)['result']
+    access_token = login_response['access_token']
+
+    # Decode JWT token to get user_id (without verification since we generated it locally)
+    # Signature verification is disabled because:
+    # 1. This is for test purposes only with locally generated tokens
+    # 2. We just need to extract the user_id from the payload
+    # 3. The token was just created by our own auth system
+    try:
+        token_payload = jwt.decode(access_token, options={'verify_signature': False})
+        user_id = token_payload['sub']
+    except (jwt.DecodeError, jwt.InvalidTokenError, KeyError) as e:
+        raise RuntimeError(f'Error decoding access token for cached user {user_name}: {e}')
+
+    user_data = {
+        'id': user_id,
+        'name': user_name,
+        'email': email,
+        'password': password,
+        'password_confirm': password
+    }
+
+    user_env = ctx.copy()
+    user_env['MAPP_CLI_ACCESS_TOKEN'] = access_token
+    user_env['Authorization'] = f'Bearer {access_token}'
+
+    return {'user': user_data, 'env': user_env}
+
 
 class TestMTemplateApp(unittest.TestCase):
 
@@ -188,6 +226,11 @@ def setUpClass(cls):
 
         os.makedirs(cls.test_dir, exist_ok=True)
 
+        # Always delete crud DB to avoid max models exceeded errors with cached data
+        # Only cache the expensive pagination DB
+        if cls.crud_db_file.exists():
+            cls.crud_db_file.unlink()
+
         pagination_db_exists = cls.pagination_db_file.exists()
 
         #
@@ -281,13 +324,11 @@ def setUpClass(cls):
         except AssertionError as e:
             raise RuntimeError(f'AssertionError {e} while creating table for crud db {module_name_kebab}.{model_name_kebab}: {crud_result.stdout + crud_result.stderr}')
 
-        # create 2 crud users #
+        # create crud users (always fresh since crud DB is recreated each time) #
 
-        if cls.use_cache:
-            raise RuntimeError('Need to fix cache with new auth system')
-
-        if not cls.use_cache and cls.spec['project']['use_builtin_modules']:
+        if cls.spec['project']['use_builtin_modules']:
             crud_users = ['alice', 'bob', 'charlie', 'david', 'evelyn']
+            print('  :: Creating crud users ::')
             for user_name in crud_users:
 
                 # create #
@@ -364,7 +405,24 @@ def setUpClass(cls):
 
         if cls.use_cache and pagination_db_exists:
             print('  :: Using cached pagination db ::')
-        else:
+
+            # login to cached pagination user #
+
+            if cls.spec['project']['use_builtin_modules']:
+                print('  :: Logging in to cached pagination user ::')
+                try:
+                    cls.pagination_user = login_cached_user(
+                        cls.cmd, 
+                        cls.pagination_ctx, 
+                        'pagination_tester', 
+                        'pagination_tester@example.com'
+                    )
+                except RuntimeError as e:
+                    print(f'  :: Could not login to cached pagination user, will recreate: {e} ::')
+                    # fall through to else block to recreate pagination db
+                    cls.use_cache = False  # disable cache for pagination db
+
+        if not cls.use_cache or not pagination_db_exists:
             print(f'  :: Seeding pagination db ::')
             seed_jobs = []
             for module in cls.spec['modules'].values():