optimization csv from url analyzer

Question

I was messing around with the csv package and I tried to implement a class that returns csv objects whenever I need to open a csv file directly from a url(and occasionally prints out something).I am still very green and I would like to know if there is something I can do to optimize this code.You are going to find the description of every method in each docstring.There are still some functionalities that I want to add, but I would like to know if the design of this very simple class is acceptable and if there is something that I could make better or faster. Thank you for your attention and time.

import csv, urllib.request

class read_CSV_URL:

    def __init__(self, url):
        self.request = urllib.request.urlopen(url)

    def read_csv_from_url(self, delimiter=",", num_of_lines="all", print_lines=False) -> object:
        """
        Read a csv file form a url.
        Three default values:
            1. Delimiter is a standard coma.
            2. Number of lines that are going to be printed
               In case you wanted just the first three lines of
               your file you can just set num_of_lines=3.
            3. In case you want to print the lines set, print_lines=True.

        """
        lines = [l.decode('utf-8') for l in self.request.readlines()]  # decode response
        csv_object = csv.reader(lines, delimiter=delimiter)
        if print_lines:
            if num_of_lines == "all":
                for line in csv_object:
                    print(line)
            else:
                for number in range(num_of_lines):
                    row = next(csv_object)
                    print(row)

        return csv_object

    def read_csv_from_url_as_dict(self, delimiter=",", num_of_lines="all", print_lines=False) -> object:
        """
        Read a csv file form a url as a dictionary.
        Three default values:
            1. Delimiter is a standard coma.
            2. Number of lines that are going to be printed
               In case you wanted just the first three lines of
               your file you can just set num_of_lines=3.
            3. In case you want to print the lines set, print_lines=True.

        """
        lines = [l.decode('utf-8') for l in self.request.readlines()]  # decode response
        csv_object = csv.DictReader(lines, delimiter=delimiter)
        if print_lines:
            if num_of_lines == 'all':
                for dicty in csv_object:
                    print(dicty)
            else:
                for number in range(num_of_lines):
                    print(next(csv_object))

        return csv_object

    def get_key(self, key : str, csv_object):
        """
        Get a single key given a csv_dictionary_object.
        """
        listy = []
        for row in csv_object:
            listy.append(row[key])
        return listy

    def get_keys(self, csv_object, **kwargs) -> list:
        listy = []
        for row in csv_object:
            sub_list = []
            for key in kwargs.values():
                sub_list.append(row[key])
            listy.append(sub_list)
        return listy


csv_1 = read_CSV_URL('http://winterolympicsmedals.com/medals.csv')
print(csv_1.read_csv_from_url_as_dict(print_lines=True))

Answer 1

This is doing too much. You're trying to support reading rows as both a list and a dict; both printing and not; and reading all lines or some lines.

Change the class name to follow PEP 8.

Drop most of your configuration options.

• Don't print lines as you go. This was debugging code and can be removed in the final version.
• Either return list or dict items but not both--write two classes if you really need both options.
• Always read all lines. Let the user directly iterate over the class. As a feature, support streaming access--I assume the reason you're avoiding reading all lines is so you can print faster.
• Drop get_key and get_keys entirely, make them a static method, or move them out of the class.

A piece of example code, the way I would want to use this as a user:

for row in read_CSV_URL('http://winterolympicsmedals.com/medals.csv'):
​    print(row)

Answer 2

Building off of what Zachary Vance said: since you haven't shown much usage of this class, I can only assume that some of your features (like supporting partial body retrieval) are not needed; and others should be moved like printing, which really should not be integrated.

Generally, don't use urllib.request when requests exists. A basic implementation of line-buffered, streamed CSV retrieval can look like

from csv import DictReader
from pprint import pprint
from typing import Iterator, Dict

import requests


def read_csv_url(url: str, *args, **kwargs) -> Iterator[Dict[str, str]]:
    with requests.get(url, stream=True) as resp:
        resp.raise_for_status()
        file_like = resp.iter_lines(decode_unicode=True)
        yield from DictReader(f=file_like, *args, **kwargs)


def test() -> None:
    for record in read_csv_url('http://winterolympicsmedals.com/medals.csv'):
        pprint(record)


if __name__ == '__main__':
    test()