Add configurable schema enforcement and structured JSON validation for extraction pipeline

src/controller.py

@@ -1,11 +1,17 @@
+from src.validation import validate_extracted_data
 from src.file_manipulator import FileManipulator
 
 class Controller:
     def __init__(self):
         self.file_manipulator = FileManipulator()
 
     def fill_form(self, user_input: str, fields: list, pdf_form_path: str):
-        return self.file_manipulator.fill_form(user_input, fields, pdf_form_path)
+    data = self.file_manipulator.fill_form(user_input, fields, pdf_form_path)
+
+    if not validate_extracted_data(data):
+        raise ValueError("Invalid extracted data")
+
+    return data
 
     def create_template(self, pdf_path: str):
-        return self.file_manipulator.create_template(pdf_path)
+        return self.file_manipulator.create_template(pdf_path)

src/schema.py

@@ -0,0 +1,7 @@
+REQUIRED_FIELDS = {
+    "incident_type": str,
+    "location": str,
+    "incident_time": str,
+    "units_involved": list,
+    "summary": str
+}

src/validation.py

@@ -0,0 +1,11 @@
+from src.schema import REQUIRED_FIELDS
+
+def validate_extracted_data(data: dict) -> bool:
+    for field, field_type in REQUIRED_FIELDS.items():
+        if field not in data:
+            return False
+        if not isinstance(data[field], field_type):
+            return False
+        if data[field] in ["", None]:
+            return False
+    return True

tests/src/tests/test_validation.py

@@ -0,0 +1,19 @@
+def test_valid_data():
+    data = {
+        "incident_type": "Fire",
+        "location": "Downtown",
+        "incident_time": "2026-03-20 10:00",
+        "units_involved": ["Unit1", "Unit2"],
+        "summary": "Fire contained"
+    }
+    assert validate_extracted_data(data) == True
+
+def test_wrong_type():
+    data = {
+        "incident_type": "Fire",
+        "location": "Downtown",
+        "incident_time": "2026-03-20 10:00",
+        "units_involved": "Unit1",  # should be list
+        "summary": "Fire contained"
+    }
+    assert validate_extracted_data(data) == False

tests/test_controller.py

@@ -0,0 +1,22 @@
+from src.controller import Controller
+
+
+def test_fill_form_validation_fail(monkeypatch):
+    controller = Controller()
+
+    def mock_fill_form(user_input, fields, pdf_form_path):
+        return {
+            "incident_type": "",  # invalid (empty)
+            "location": "Downtown",
+            "incident_time": "2026-03-20T10:30:00",
+            "units_involved": ["Unit 42"],
+            "summary": "Minor incident reported"
+        }
+
+    monkeypatch.setattr(controller.file_manipulator, "fill_form", mock_fill_form)
+
+    try:
+        controller.fill_form("input", [], "file.pdf")
+        assert False  # Should not reach here
+    except ValueError:
+        assert True

tests/test_vallidation.py

@@ -0,0 +1,19 @@
+def test_valid_data():
+    data = {
+        "incident_type": "Fire",
+        "location": "Downtown",
+        "incident_time": "2026-03-20 10:00",
+        "units_involved": ["Unit1", "Unit2"],
+        "summary": "Fire contained"
+    }
+    assert validate_extracted_data(data) == True
+
+def test_wrong_type():
+    data = {
+        "incident_type": "Fire",
+        "location": "Downtown",
+        "incident_time": "2026-03-20 10:00",
+        "units_involved": "Unit1",  # should be list
+        "summary": "Fire contained"
+    }
+    assert validate_extracted_data(data) == False