Search⌘ K
AI Features
Home
Courses
Building Robust Object-Oriented Python Applications and Libraries

Reading CSV Files

Explore how to read CSV files in Python by using from_dict methods and a load function that partitions data. Learn to handle errors like InvalidSampleError for better user feedback and avoid repeating code by validating data with superclass and subclasses.

We'll cover the following...

Using form_dict() method

We’ll provide a common template for creating objects from CSV source data. The idea is to leverage the from_dict() methods of the various classes to create the objects our application uses:

Python 3.10.4
class TrainingData:
    def __init__(self, name: str) -> None:
        self.name = name
        self.uploaded: datetime.datetime
        self.tested: datetime.datetime
        self.training: list[TrainingKnownSample] = []
        self.testing: list[TestingKnownSample] = []
        self.tuning: list[Hyperparameter] = []
        
    def load(self, raw_data_iter) -> None:
        for n, row in enumerate(raw_data_iter):
            try:
                if n % 5 == 0:
                    test = TestingKnownSample.from_dict(row)
                    self.testing.append(test)
                else:
                    train = TrainingKnownSample.from_dict(row)
                    self.training.append(train)
            except InvalidSampleError as ex:
                print(f"Row {n+1}: {ex}")
                return
        self.uploaded = datetime.datetime.now(tz = datetime.timezone.utc)

Constructing the load() method

The load() method is partitioning the samples into testing and training subsets. It expects an iterable source of dict[str, str] objects, which are produced by a csv.DictReader object.

The user experience implemented here is to report the first failure and return. This might lead to an error message like the following:

Python 3.10.4
text Row 2: invalid species in {'sepal_length': 7.9, 'sepal_width':
3.2, 'petal_length': 4.7, 'petal_width': 1.4, 'species': 'Buttercup'}

This message has all the required information, but may not be as helpful as desired. We ...