RTE Pytest

def compare_first_four_keys(expected_dict, predicted_dict):
    #Compare the first 4 key-value pairs to see if the dictionaries match, 
    #including checking the first 3 values if the value is a list.

    expected_items = list(expected_dict.items())[:3]  # Get the first 4 key-value pairs
    for key, expected_value in expected_items:
        if key in predicted_dict:
            predicted_value = predicted_dict[key]
            # If both values are lists, compare the first 3 elements
            if isinstance(expected_value, list) and isinstance(predicted_value, list):
                if len(expected_value) >= 3 and len(predicted_value) >= 3:
                   if all(pytest.approx(ev) == pv for ev, pv in zip(expected_value, predicted_value)):
                        return True
    return False


# Sample function that generates the output dictionaries (replace with your function)
def get_output_dict(Table_Category):
    # Example function output based on Table_Category
    if Table_Category == 'SLP':
        return {'key1': 'value1', 'key2': 'value2', ...}  # Replace with actual dictionary
    elif Table_Category == 'XYZ':
        return {'keyA': 'valueA', 'keyB': 'valueB', ...}
    return {}


def dict_length_check(expected_dict):
    if expected_dict['Table Category'] == 'RLM' and len(expected_dict) == 46:
        print("Length of this RLM table matches")
        return True
    elif expected_dict['Table Category'] == 'SLP' and len(expected_dict) == 24:
        print("Length of this SLP table matches")
        return True
    elif expected_dict['Table Category'] == 'Metering' and len(expected_dict) == 17:
        print("Length of this Metering table matches")
        return True        


#test with parametrize for two arguments: pdf_name and expected_pdf
@pytest.mark.parametrize("pdf_name, expected_pdf", [
    ("Actual-Avacon_2024", "Actual-Avacon_2024"),
    ("ACTUAL_Erdgas Mittelsachsen GmbH_2024", "Mittelsachsen"),("Actual-bnNETZE_2024","bnNetze"),
    ("Preliminar_Gemeindewerke Peiner Land GmbH_2024","Preliminar_Gemeindewerke")])

def test_actual_avacon(pdf_name,expected_pdf):
    accuracies = []
    pdf_file_path = os.path.join(current_dir, '..', '..', 'docs', 'RTE', pdf_name + ".pdf")
    with open (pdf_file_path,'rb') as file:
        response = client.post('/rte_from_file', files= {'file': file})
        print("GPT Response received")
    assert response.status_code == 200
    predicted_json = response.json()
    expected_output = expected_op_json['expected_output'][expected_pdf]
    for i in range (len(predicted_json['extracted_keyword_list'])):
        expected_dict = expected_output[i]
        predicted_dict = predicted_json['extracted_keyword_list'][i]
        if compare_first_four_keys(expected_dict,predicted_dict) and dict_length_check(expected_dict):
            total_elements = 0
            matched_keys = 0
            # Iterate through the expected dictionary
            for key in list(expected_dict.keys()):
                if key in predicted_dict:
                    total_elements += 1
                    expected_value = expected_dict[key]
                    predicted_value = predicted_dict[key]
                    # Handle list comparisons
                    if isinstance(expected_value, list) and isinstance(predicted_value, list):
                        if len(expected_value) == len(predicted_value):
                            # Use pytest.approx for numerical lists
                            if all(isinstance(i, (int, float)) for i in expected_value):
                                if all(pytest.approx(ev) == pv for ev, pv in zip(expected_value, predicted_value)):
                                    matched_keys += 1
                            # Direct equality for non-numeric lists
                            elif expected_value == predicted_value:
                                matched_keys += 1
                    # Use pytest.approx for single numeric values
                    elif isinstance(expected_value, (int, float)) and isinstance(predicted_value, (int, float)):
                        if pytest.approx(expected_value) == predicted_value:
                            matched_keys += 1
                    # Direct equality for non-numeric values
                    elif expected_value == predicted_value:
                        matched_keys += 1
        accuracy = (matched_keys/total_elements) * 100 if total_elements else 0
        accuracies.append(accuracy)
        print(f"Dictionary {i+1}: Accuracy is {accuracy:.2f}% with {matched_keys}/{total_elements} elements matched.")
        print(f"Dictionary {i+1} has Accuracy: {accuracy:.2f}%.")