dbagent-cli 0.11.3__tar.gz → 0.11.4__tar.gz

{dbagent_cli-0.11.3 → dbagent_cli-0.11.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dbagent-cli
-Version: 0.11.3
+Version: 0.11.4
 Summary: The Privacy-First AI Database Copilot, Schema Graph Introspector & Post-Commit Recovery Engine (CLI)
 Author: Santhosh Gupta
 License: MIT

{dbagent_cli-0.11.3 → dbagent_cli-0.11.4}/dbagent/__init__.py

@@ -2,5 +2,5 @@
 DB-Agent: Universal Database Introspection and Script Generation AI Agent (CLI).
 """
 
-__version__ = "0.11.3"
+__version__ = "0.11.4"
 __author__ = "Santhosh Gupta"

{dbagent_cli-0.11.3 → dbagent_cli-0.11.4}/dbagent/agent/generator.py

@@ -60,7 +60,7 @@ Generate ONLY the raw executable SQL query for the given database. No markdown f
    - If the column is VARCHAR(1) or CHAR(1) or text, check if values use 'Y'/'N', 'YES'/'NO', '1'/'0', 'ACTIVE'/'INACTIVE', 'TRUE'/'FALSE'.
    - If sample values show 'Y'/'N' (or if it's a single-character VARCHAR/CHAR column), use `active = 'Y'` or `active IN ('Y', '1', 'true', 'TRUE')` for active/enabled status.
 10. For Deletions & Cascading Mapping Cleanup (when deleting a record/user without foreign key constraint violations):
-   - For PostgreSQL: Use a single Common Table Expression (CTE) with data-modifying statements deleting from all child, mapping, activity, profile, and preference tables first, and finally deleting from the parent table (e.g. `WITH target AS (SELECT pk_user_id FROM ivis_user WHERE ...), del_act AS (DELETE FROM user_activity WHERE user_id IN (SELECT pk_user_id FROM target)), del_map AS (DELETE FROM ivis_user_access WHERE fk_user_id IN (SELECT pk_user_id FROM target)) DELETE FROM ivis_user WHERE pk_user_id IN (SELECT pk_user_id FROM target);`).
+   - For PostgreSQL: Use a single Common Table Expression (CTE) with data-modifying statements deleting from all child, mapping, activity, profile, and preference tables first, and finally deleting from the parent table (e.g. `WITH target AS (SELECT id FROM users WHERE status = 'inactive'), del_orders AS (DELETE FROM orders WHERE user_id IN (SELECT id FROM target)), del_logs AS (DELETE FROM user_logs WHERE user_id IN (SELECT id FROM target)) DELETE FROM users WHERE id IN (SELECT id FROM target);`).
    - For MySQL / SQLite: Generate clean, ordered DELETE statements deleting referencing child records before the parent record.
    - Always delete from all referencing tables present in the schema to avoid foreign key violations.
 11. Multi-Turn Session Memory & Query Refinement:
@@ -201,7 +201,7 @@ class ScriptGenerator:
         if re.search(r'\b(?:is|are)\b.*\b(?:same|different|difference|related|equal|linked)\b', prompt_lower):
             return "analysis"
 
-        # Contextual clarification for analysis (e.g. "i mean ivis_scope", "i mean table_x")
+        # Contextual clarification for analysis (e.g. "i mean user_profiles", "i mean table_x")
         if re.search(r'\bi mean\b', prompt_lower) and not re.search(r'\b(insert|update|delete|drop)\b', prompt_lower):
             return "analysis"
 

{dbagent_cli-0.11.3 → dbagent_cli-0.11.4}/dbagent/agent/pipeline.py

@@ -176,7 +176,7 @@ class QueryPipeline:
                 all_table_names = self.connector.get_table_names()
                 stems = set()
                 for t in target_names:
-                    unprefixed = re.sub(r'^(?:ivis_|tbl_|table_|app_)', '', t.lower())
+                    unprefixed = re.sub(r'^(?:tbl_|table_|app_|dim_|fact_|rel_|raw_)', '', t.lower())
                     unprefixed = re.sub(r'(?:s|es)$', '', unprefixed)
                     if len(unprefixed) >= 3:
                         stems.add(unprefixed)

{dbagent_cli-0.11.3 → dbagent_cli-0.11.4}/dbagent/cli.py

@@ -808,7 +808,7 @@ def chat_command(
 
                 if cmd_name == ":table":
                     if not cmd_arg:
-                        print_warning("Usage: :table <table_name> (e.g. :table ivis_users)")
+                        print_warning("Usage: :table <table_name> (e.g. :table users)")
                         continue
                     exact, fuzzy = connector.resolve_tables(cmd_arg)
                     target = exact[0] if exact else (fuzzy[0] if fuzzy else cmd_arg)
@@ -823,7 +823,7 @@ def chat_command(
 
                 if cmd_name in [":fk", ":relations", ":mapped"]:
                     if not cmd_arg:
-                        print_warning("Usage: :fk <table_name> (e.g. :fk ivis_users)")
+                        print_warning("Usage: :fk <table_name> (e.g. :fk users)")
                         continue
                     exact, fuzzy = connector.resolve_tables(cmd_arg)
                     target = exact[0] if exact else (fuzzy[0] if fuzzy else cmd_arg)
@@ -931,7 +931,7 @@ def chat_command(
 
                 if cmd_name in [":learn", ":remember"]:
                     if not cmd_arg or "=" not in cmd_arg:
-                        print_warning("Usage: :learn <topic> = <description> (e.g. :learn site_user_map = ivis_site maps to ivis_user via site_group -> group)")
+                        print_warning("Usage: :learn <topic> = <description> (e.g. :learn user_roles = users map to roles via user_role_map)")
                         continue
                     parts = cmd_arg.split("=", 1)
                     topic = parts[0].strip()

{dbagent_cli-0.11.3 → dbagent_cli-0.11.4}/dbagent/connectors/base.py

@@ -87,8 +87,8 @@ class BaseConnector(ABC):
                 exact.append(t)
                 continue
 
-            # 2. Match without common prefixes (e.g. "ivis_scope" -> "scope", "scopes")
-            unprefixed = re.sub(r'^(?:ivis_|tbl_|table_|app_)', '', t_lower)
+            # 2. Match without common database table prefixes (e.g. "tbl_orders" -> "orders")
+            unprefixed = re.sub(r'^(?:tbl_|table_|app_|dim_|fact_|rel_|raw_)', '', t_lower)
             if unprefixed and len(unprefixed) >= 3 and unprefixed not in stop_words:
                 if (unprefixed in tokens or 
                     f"{unprefixed}s" in tokens or 

{dbagent_cli-0.11.3 → dbagent_cli-0.11.4}/dbagent_cli.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dbagent-cli
-Version: 0.11.3
+Version: 0.11.4
 Summary: The Privacy-First AI Database Copilot, Schema Graph Introspector & Post-Commit Recovery Engine (CLI)
 Author: Santhosh Gupta
 License: MIT

{dbagent_cli-0.11.3 → dbagent_cli-0.11.4}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "dbagent-cli"
-version = "0.11.3"
+version = "0.11.4"
 description = "The Privacy-First AI Database Copilot, Schema Graph Introspector & Post-Commit Recovery Engine (CLI)"
 readme = "README.md"
 requires-python = ">=3.9"

{dbagent_cli-0.11.3 → dbagent_cli-0.11.4}/tests/test_direct_sql.py

@@ -11,7 +11,7 @@ from dbagent.llm.mock_provider import MockProvider
 def test_is_direct_sql_detection():
     # Direct SQL queries
     assert SQLValidator.is_direct_sql("SELECT * FROM users;") is True
-    assert SQLValidator.is_direct_sql("select pk_user_id, username from ivis_user where active = 'Y'") is True
+    assert SQLValidator.is_direct_sql("select id, username from users where active = 'Y'") is True
     assert SQLValidator.is_direct_sql("SELECT count(*) FROM orders") is True
     assert SQLValidator.is_direct_sql("SELECT 1;") is True
     assert SQLValidator.is_direct_sql("WITH cte AS (SELECT 1) SELECT * FROM cte") is True

{dbagent_cli-0.11.3 → dbagent_cli-0.11.4}/tests/test_generator.py

@@ -74,7 +74,7 @@ def test_validator_destructive_delete_without_where():
 
 
 def test_classify_intent_analysis():
-    prompt = "analyze the database and give me the linkage architecture that how ivis_user and ivis_tenant was related and how they were mapped and how to know that which sites in ivis_site were mapped to which user"
+    prompt = "analyze the database and give me the linkage architecture of how users and accounts are related and how they are mapped and how to know which devices in customer_devices are mapped to which user"
     assert ScriptGenerator.classify_intent(prompt) == "analysis"
 
     assert ScriptGenerator.classify_intent("how are users and orders related?") == "analysis"

dbagent_cli-0.11.4/tests/test_ground_truth_sql_validation.py

@@ -0,0 +1,150 @@
+"""
+Unit tests for Schema Ground-Truth SQL Validation and Auto-Correction.
+Verifies table, column, PK, and FK validation against physical database schema using generic domain models.
+"""
+
+import pytest
+from unittest.mock import MagicMock
+from dbagent.schema.models import DatabaseSchema, TableModel, ColumnModel, ForeignKeyModel
+from dbagent.schema.ground_truth import SchemaGroundTruthValidator
+from dbagent.agent.pipeline import QueryPipeline, PipelineResult
+from dbagent.security.modes import OperationMode
+
+
+@pytest.fixture
+def sample_schema():
+    """Create a sample generic schema with realistic primary keys and foreign keys."""
+    customers = TableModel(
+        name="customers",
+        columns=[
+            ColumnModel(name="customer_id", data_type="INTEGER", is_primary_key=True),
+            ColumnModel(name="customer_name", data_type="VARCHAR(255)"),
+            ColumnModel(name="account_id", data_type="INTEGER"),
+            ColumnModel(name="region_id", data_type="INTEGER"),
+            ColumnModel(name="is_active", data_type="BOOLEAN"),
+        ],
+        foreign_keys=[
+            ForeignKeyModel(
+                constrained_columns=["account_id"],
+                referred_table="accounts",
+                referred_columns=["account_id"],
+            ),
+            ForeignKeyModel(
+                constrained_columns=["region_id"],
+                referred_table="regions",
+                referred_columns=["region_id"],
+            ),
+        ],
+        primary_key=["customer_id"],
+    )
+
+    accounts = TableModel(
+        name="accounts",
+        columns=[
+            ColumnModel(name="account_id", data_type="INTEGER", is_primary_key=True),
+            ColumnModel(name="account_name", data_type="VARCHAR(255)"),
+            ColumnModel(name="billing_code", data_type="VARCHAR(50)"),
+        ],
+        foreign_keys=[],
+        primary_key=["account_id"],
+    )
+
+    regions = TableModel(
+        name="regions",
+        columns=[
+            ColumnModel(name="region_id", data_type="INTEGER", is_primary_key=True),
+            ColumnModel(name="region_name", data_type="VARCHAR(255)"),
+        ],
+        foreign_keys=[],
+        primary_key=["region_id"],
+    )
+
+    return DatabaseSchema(
+        database_name="sample_db",
+        dialect_name="postgresql",
+        tables=[customers, accounts, regions],
+    )
+
+
+def test_valid_select_query_passes_ground_truth(sample_schema):
+    sql = "SELECT c.customer_id, c.customer_name, a.account_name FROM customers c JOIN accounts a ON c.account_id = a.account_id WHERE c.is_active = true;"
+    is_valid, errors = SchemaGroundTruthValidator.validate_sql_against_schema(sql, sample_schema)
+    assert is_valid is True
+    assert len(errors) == 0
+
+
+def test_valid_update_query_passes_ground_truth(sample_schema):
+    sql = "UPDATE customers SET account_id = 10, region_id = 20 WHERE customer_id = 100;"
+    is_valid, errors = SchemaGroundTruthValidator.validate_sql_against_schema(sql, sample_schema)
+    assert is_valid is True
+    assert len(errors) == 0
+
+
+def test_hallucinated_column_in_update_set_detected(sample_schema):
+    sql = "UPDATE customers SET project = 5, site_group = 10 WHERE customer_id = 100;"
+    is_valid, errors = SchemaGroundTruthValidator.validate_sql_against_schema(sql, sample_schema)
+    assert is_valid is False
+    assert any("Column 'project' does not exist in table 'customers'" in e for e in errors)
+    assert any("Column 'site_group' does not exist in table 'customers'" in e for e in errors)
+    assert any("Valid foreign keys" in e for e in errors)
+
+
+def test_hallucinated_column_in_where_clause_detected(sample_schema):
+    sql = "UPDATE customers SET account_id = 5 WHERE legacy_user_code = 100;"
+    is_valid, errors = SchemaGroundTruthValidator.validate_sql_against_schema(sql, sample_schema)
+    assert is_valid is False
+    assert any("Column 'legacy_user_code' in WHERE clause does not exist in table 'customers'" in e for e in errors)
+    assert any("Primary key is 'customer_id'" in e for e in errors)
+
+
+def test_hallucinated_table_name_detected(sample_schema):
+    sql = "SELECT * FROM non_existent_table WHERE id = 1;"
+    is_valid, errors = SchemaGroundTruthValidator.validate_sql_against_schema(sql, sample_schema)
+    assert is_valid is False
+    assert any("Table 'non_existent_table' does not exist" in e for e in errors)
+
+
+def test_hallucinated_qualified_column_detected(sample_schema):
+    sql = "SELECT c.customer_id, c.fake_column FROM customers c;"
+    is_valid, errors = SchemaGroundTruthValidator.validate_sql_against_schema(sql, sample_schema)
+    assert is_valid is False
+    assert any("Column 'fake_column' does not exist in table 'customers'" in e for e in errors)
+
+
+def test_multi_statement_validation(sample_schema):
+    sql = """
+    UPDATE customers SET account_id = 2 WHERE customer_id = 1;
+    UPDATE customers SET bad_col = 3 WHERE customer_id = 2;
+    """
+    is_valid, errors = SchemaGroundTruthValidator.validate_sql_against_schema(sql, sample_schema)
+    assert is_valid is False
+    assert any("Column 'bad_col' does not exist in table 'customers'" in e for e in errors)
+
+
+def test_pipeline_auto_corrects_hallucinated_columns(sample_schema):
+    mock_connector = MagicMock()
+    mock_connector.get_table_names.return_value = ["customers", "accounts", "regions"]
+    mock_connector.resolve_tables.return_value = (["customers"], [])
+    mock_connector.inspect_schema.return_value = sample_schema
+    mock_connector.execute_query.return_value = (["customer_id", "account_id"], [{"customer_id": 1, "account_id": 2}], None)
+
+    mock_llm = MagicMock()
+    mock_llm.generate.side_effect = [
+        "UPDATE customers SET account = 2 WHERE cust_id = 1;",
+        "UPDATE customers SET account_id = 2 WHERE customer_id = 1;",
+    ]
+
+    pipeline = QueryPipeline(
+        connector=mock_connector,
+        llm=mock_llm,
+        mode=OperationMode.CONFIRM,
+        auto_execute=True,
+    )
+
+    result = pipeline.run("Change the account for customer 1 to account 2", force=True)
+
+    assert result.success is True
+    assert "account_id = 2" in result.sql
+    assert "customer_id = 1" in result.sql
+    assert result.retries >= 1
+    assert result.ground_truth is not None

{dbagent_cli-0.11.3 → dbagent_cli-0.11.4}/tests/test_knowledge.py

@@ -17,27 +17,27 @@ def temp_kb(tmp_path):
 def test_knowledge_base_learn_and_find(temp_kb):
     rule = temp_kb.learn(
         db_name="test_db",
-        topic="Site to User Group Mapping",
-        rule_description="ivis_site maps to ivis_user through ivis_site_group -> ivis_group -> ivis_user_group_map",
-        relevant_tables=["ivis_site", "ivis_site_group", "ivis_group", "ivis_user_group_map", "ivis_user"],
+        topic="User to Role Permission Mapping",
+        rule_description="users map to permissions through user_roles -> roles -> role_permissions",
+        relevant_tables=["users", "user_roles", "roles", "role_permissions", "permissions"],
         source="user_correction",
     )
     assert rule.rule_id.startswith("rule_")
-    assert rule.topic == "Site to User Group Mapping"
+    assert rule.topic == "User to Role Permission Mapping"
 
     # List all
     all_rules = temp_kb.list_all("test_db")
     assert len(all_rules) == 1
 
     # Find relevant by prompt
-    matched = temp_kb.find_relevant("test_db", "how to map sites to user")
+    matched = temp_kb.find_relevant("test_db", "how to map users to permissions")
     assert len(matched) == 1
-    assert matched[0].topic == "Site to User Group Mapping"
+    assert matched[0].topic == "User to Role Permission Mapping"
 
     # Context formatting
-    ctx = temp_kb.format_prompt_context("test_db", "how to map sites to user")
+    ctx = temp_kb.format_prompt_context("test_db", "how to map users to permissions")
     assert "Persistent Database Domain Knowledge" in ctx
-    assert "Site to User Group Mapping" in ctx
+    assert "User to Role Permission Mapping" in ctx
 
     # Delete
     deleted = temp_kb.delete("test_db", rule.rule_id)

{dbagent_cli-0.11.3 → dbagent_cli-0.11.4}/tests/test_prompt_queue.py

@@ -26,10 +26,10 @@ def test_prompt_queue_fifo():
 
 def test_split_and_enqueue_chained():
     qm = PromptQueueManager()
-    raw = "how ivis_user and ivis_site were linked ; give the flow diagram ; count total users"
+    raw = "how users and orders were linked ; give the flow diagram ; count total users"
     first = qm.split_and_enqueue_chained(raw)
 
-    assert first == "how ivis_user and ivis_site were linked"
+    assert first == "how users and orders were linked"
     assert qm.count() == 2
     assert qm.list_all() == ["give the flow diagram", "count total users"]
 

{dbagent_cli-0.11.3 → dbagent_cli-0.11.4}/tests/test_prompt_queue_and_sql.py

@@ -11,10 +11,10 @@ def test_queue_does_not_split_multiline_sql_with_semicolons():
     queue_mgr = PromptQueueManager()
 
     sql_input = """
-    DELETE FROM ivis_customer_user_map WHERE fk_user_id = 5201;
-    DELETE FROM ivis_user_access WHERE fk_user_id = 5201;
+    DELETE FROM customer_user_map WHERE user_id = 5201;
+    DELETE FROM user_access WHERE user_id = 5201;
     DELETE FROM user_activity WHERE user_id = 5201;
-    DELETE FROM ivis_user WHERE pk_user_id = 5201;
+    DELETE FROM users WHERE user_id = 5201;
     COMMIT; correct this
     """
 
@@ -27,12 +27,12 @@ def test_classify_query_with_cte_delete():
     """Verify that data-modifying WITH CTE statements are classified as 'write'."""
     cte_delete = """
     WITH target_user AS (
-        SELECT pk_user_id FROM ivis_user WHERE username = 'target_user'
+        SELECT id FROM users WHERE username = 'target_user'
     ),
     del_access AS (
-        DELETE FROM ivis_user_access WHERE fk_user_id IN (SELECT pk_user_id FROM target_user)
+        DELETE FROM user_access WHERE user_id IN (SELECT id FROM target_user)
     )
-    DELETE FROM ivis_user WHERE pk_user_id IN (SELECT pk_user_id FROM target_user);
+    DELETE FROM users WHERE id IN (SELECT id FROM target_user);
     """
     assert SQLValidator.classify_query(cte_delete) == "write"
 
@@ -41,7 +41,7 @@ def test_classify_query_with_cte_read():
     """Verify that regular SELECT WITH CTE statements remain classified as 'read'."""
     cte_read = """
     WITH active_users AS (
-        SELECT pk_user_id, username FROM ivis_user WHERE active = 'Y'
+        SELECT id, username FROM users WHERE active = 'Y'
     )
     SELECT * FROM active_users LIMIT 10;
     """

dbagent_cli-0.11.3/tests/test_ground_truth_sql_validation.py

@@ -1,150 +0,0 @@
-"""
-Unit tests for Schema Ground-Truth SQL Validation and Auto-Correction.
-Verifies table, column, PK, and FK validation against physical database schema.
-"""
-
-import pytest
-from unittest.mock import MagicMock
-from dbagent.schema.models import DatabaseSchema, TableModel, ColumnModel, ForeignKeyModel
-from dbagent.schema.ground_truth import SchemaGroundTruthValidator
-from dbagent.agent.pipeline import QueryPipeline, PipelineResult
-from dbagent.security.modes import OperationMode
-
-
-@pytest.fixture
-def sample_schema():
-    """Create a sample schema with realistic primary keys and foreign keys."""
-    ivis_site = TableModel(
-        name="ivis_site",
-        columns=[
-            ColumnModel(name="pk_site_id", data_type="INTEGER", is_primary_key=True),
-            ColumnModel(name="site_name", data_type="VARCHAR(255)"),
-            ColumnModel(name="fk_project_id", data_type="INTEGER"),
-            ColumnModel(name="fk_site_group_id", data_type="INTEGER"),
-            ColumnModel(name="is_active", data_type="BOOLEAN"),
-        ],
-        foreign_keys=[
-            ForeignKeyModel(
-                constrained_columns=["fk_project_id"],
-                referred_table="ivis_project",
-                referred_columns=["pk_project_id"],
-            ),
-            ForeignKeyModel(
-                constrained_columns=["fk_site_group_id"],
-                referred_table="ivis_site_group",
-                referred_columns=["pk_site_group_id"],
-            ),
-        ],
-        primary_key=["pk_site_id"],
-    )
-
-    ivis_project = TableModel(
-        name="ivis_project",
-        columns=[
-            ColumnModel(name="pk_project_id", data_type="INTEGER", is_primary_key=True),
-            ColumnModel(name="project_name", data_type="VARCHAR(255)"),
-            ColumnModel(name="customer_code", data_type="VARCHAR(50)"),
-        ],
-        foreign_keys=[],
-        primary_key=["pk_project_id"],
-    )
-
-    ivis_site_group = TableModel(
-        name="ivis_site_group",
-        columns=[
-            ColumnModel(name="pk_site_group_id", data_type="INTEGER", is_primary_key=True),
-            ColumnModel(name="group_name", data_type="VARCHAR(255)"),
-        ],
-        foreign_keys=[],
-        primary_key=["pk_site_group_id"],
-    )
-
-    return DatabaseSchema(
-        database_name="iportal",
-        dialect_name="postgresql",
-        tables=[ivis_site, ivis_project, ivis_site_group],
-    )
-
-
-def test_valid_select_query_passes_ground_truth(sample_schema):
-    sql = "SELECT s.pk_site_id, s.site_name, p.project_name FROM ivis_site s JOIN ivis_project p ON s.fk_project_id = p.pk_project_id WHERE s.is_active = true;"
-    is_valid, errors = SchemaGroundTruthValidator.validate_sql_against_schema(sql, sample_schema)
-    assert is_valid is True
-    assert len(errors) == 0
-
-
-def test_valid_update_query_passes_ground_truth(sample_schema):
-    sql = "UPDATE ivis_site SET fk_project_id = 10, fk_site_group_id = 20 WHERE pk_site_id = 100;"
-    is_valid, errors = SchemaGroundTruthValidator.validate_sql_against_schema(sql, sample_schema)
-    assert is_valid is True
-    assert len(errors) == 0
-
-
-def test_hallucinated_column_in_update_set_detected(sample_schema):
-    sql = "UPDATE ivis_site SET project = 5, site_group = 10 WHERE pk_site_id = 100;"
-    is_valid, errors = SchemaGroundTruthValidator.validate_sql_against_schema(sql, sample_schema)
-    assert is_valid is False
-    assert any("Column 'project' does not exist in table 'ivis_site'" in e for e in errors)
-    assert any("Column 'site_group' does not exist in table 'ivis_site'" in e for e in errors)
-    assert any("Valid foreign keys" in e for e in errors)
-
-
-def test_hallucinated_column_in_where_clause_detected(sample_schema):
-    sql = "UPDATE ivis_site SET fk_project_id = 5 WHERE customer_site_id = 100;"
-    is_valid, errors = SchemaGroundTruthValidator.validate_sql_against_schema(sql, sample_schema)
-    assert is_valid is False
-    assert any("Column 'customer_site_id' in WHERE clause does not exist in table 'ivis_site'" in e for e in errors)
-    assert any("Primary key is 'pk_site_id'" in e for e in errors)
-
-
-def test_hallucinated_table_name_detected(sample_schema):
-    sql = "SELECT * FROM non_existent_table WHERE id = 1;"
-    is_valid, errors = SchemaGroundTruthValidator.validate_sql_against_schema(sql, sample_schema)
-    assert is_valid is False
-    assert any("Table 'non_existent_table' does not exist" in e for e in errors)
-
-
-def test_hallucinated_qualified_column_detected(sample_schema):
-    sql = "SELECT s.pk_site_id, s.fake_column FROM ivis_site s;"
-    is_valid, errors = SchemaGroundTruthValidator.validate_sql_against_schema(sql, sample_schema)
-    assert is_valid is False
-    assert any("Column 'fake_column' does not exist in table 'ivis_site'" in e for e in errors)
-
-
-def test_multi_statement_validation(sample_schema):
-    sql = """
-    UPDATE ivis_site SET fk_project_id = 2 WHERE pk_site_id = 1;
-    UPDATE ivis_site SET bad_col = 3 WHERE pk_site_id = 2;
-    """
-    is_valid, errors = SchemaGroundTruthValidator.validate_sql_against_schema(sql, sample_schema)
-    assert is_valid is False
-    assert any("Column 'bad_col' does not exist in table 'ivis_site'" in e for e in errors)
-
-
-def test_pipeline_auto_corrects_hallucinated_columns(sample_schema):
-    mock_connector = MagicMock()
-    mock_connector.get_table_names.return_value = ["ivis_site", "ivis_project", "ivis_site_group"]
-    mock_connector.resolve_tables.return_value = (["ivis_site"], [])
-    mock_connector.inspect_schema.return_value = sample_schema
-    mock_connector.execute_query.return_value = (["pk_site_id", "fk_project_id"], [{"pk_site_id": 1, "fk_project_id": 2}], None)
-
-    mock_llm = MagicMock()
-    mock_llm.generate.side_effect = [
-        "UPDATE ivis_site SET project = 2 WHERE site_id = 1;",
-        "UPDATE ivis_site SET fk_project_id = 2 WHERE pk_site_id = 1;",
-    ]
-
-    pipeline = QueryPipeline(
-        connector=mock_connector,
-        llm=mock_llm,
-        mode=OperationMode.CONFIRM,
-        auto_execute=True,
-    )
-
-    result = pipeline.run("Change the project for site 1 to project 2", force=True)
-
-    assert result.success is True
-    assert "fk_project_id = 2" in result.sql
-    assert "pk_site_id = 1" in result.sql
-    assert result.retries >= 1
-    assert result.ground_truth is not None