I/O

Working with files

Writing programs that run on different operating systems is tricky: home directories are C:\Users\andgein on Windows, /home/andgein on Linux, and /Users/andgein on macOS
Operating systems can use different file systems, such as FAT32, NTFS, ext3/ext4, and APFS
Python provides an abstraction layer, but paths still need care: does open("C:\new\test.txt") look correct?
The classic approach uses open, import os, and related APIs
The modern approach uses import pathlib

Pathlib

from pathlib import Path
from tempfile import gettempdir

# E.g. /tmp on Linux, /var/folders/.../T on macOS,
# C:\Users\<user>\AppData\Local\Temp on Windows
path = Path(gettempdir()) / "notes.txt"

file = path.open("w")
file.write("Operator := is called a walrus operator\n")
file.write("Operator |= merges another dictionary into this dictionary in place\n")
file.close()

You must always close open files. But what happens if an error is raised between open() and close()?

Writing a file

from pathlib import Path
from tempfile import gettempdir

# E.g. /tmp on Linux, /var/folders/.../T on macOS,
# C:\Users\<user>\AppData\Local\Temp on Windows
path = Path(gettempdir()) / "notes.txt"

with path.open("w") as file:
    file.write("Operator := is called a walrus operator\n")
    file.write("Operator |= merges another dictionary into this dictionary in place\n")

Another approach

path.write_text(
    "Operator := is called a walrus operator\n"
    "Operator |= merges another dictionary into this dictionary in place\n"
)

It does not require opening or closing the file, and it replaces its entire contents

Reading a file

with path.open("r") as file:
    line = file.readline()         # Read the next line
    lines = file.readlines()       # Read remaining lines into a list

with path.open("r") as file:
    part = file.read(100)          # Read the next 100 characters
    content = file.read()          # Read all remaining text

with path.open("r") as file:
    for line in file:              # Iterate over lines
        print(line)

content = path.read_text()         # Read all text at once

By default, reading starts at the beginning of a file and advances its position with each read

Text vs bytes

Python opens files in text mode by default
Not every file stores text — so what makes a binary file different?
Use "rb" or "wb" for binary files, or the read_bytes() and write_bytes() shortcuts

with path.open("rb") as file:
    content = file.read()

print(content)

Bytes

A byte is a number from 0 to 255
A sequence of bytes has the type bytes and behaves much like a list of integers from 0 to 255

print(list(b"Operator := is called a walrus operator\n"))

Those numbers are ASCII codes for Latin letters and common characters

ASCII

Code	Character	Code	Character	Code	Character	Code	Character
00	Null	32	Space	64	@	96	`
01	Start heading	33	!	65	A	97	a
02	Start text	34	"	66	B	98	b
03	End text	35	#	67	C	99	c
04	End transmission	36	$	68	D	100	d
05	Enquiry	37	%	69	E	101	e
06	Acknowledge	38	&	70	F	102	f
07	Bell	39	'	71	G	103	g
08	Backspace	40	(	72	H	104	h
09	Tab	41	)	73	I	105	i
10	Line feed	42	*	74	J	106	j
11	Vertical tab	43	+	75	K	107	k
12	Form feed	44	,	76	L	108	l
13	Carriage return	45	-	77	M	109	m
14	Shift out	46	.	78	N	110	n
15	Shift in	47	/	79	O	111	o
16	Data escape	48	0	80	P	112	p
17	Control 1	49	1	81	Q	113	q
18	Control 2	50	2	82	R	114	r
19	Control 3	51	3	83	S	115	s
20	Control 4	52	4	84	T	116	t
21	Negative ack	53	5	85	U	117	u
22	Synchronous idle	54	6	86	V	118	v
23	End block	55	7	87	W	119	w
24	Cancel	56	8	88	X	120	x
25	End medium	57	9	89	Y	121	y
26	Substitute	58	:	90	Z	122	z
27	Escape	59	;	91	[	123	{
28	File separator	60	<	92	\	124	\|
29	Group separator	61	=	93	]	125	}
30	Record separator	62	>	94	^	126	~
31	Unit separator	63	?	95	_	127	Delete

Bytes

But there are many non-Latin letters — ß, δ, я — and other characters, such as ⌘ and ⌀
How do we represent them?
One character can span multiple bytes!

Unicode!

text = "⌘"
print(text.encode())
print(list(text.encode()))

Network

Exchanging data over a network is similar: we read and write bytes
Both sides must agree on protocols though, from TCP or UDP to DNS, TLS, HTTP, and IMAP
Network programming is outside this course’s scope — unfortunately

HTTP

HTTP (and HTTPS) is the most common kind of network communication for Python developers
Use it to download pages or call APIs
Choose your weapon: requests, httpx2, or aiohttp

import requests

response = requests.get("https://swapi.dev/api/planets/")
data = response.json()
for planet in data["results"]:
    print(planet["name"], planet["population"], planet["url"], sep=" — ")

HTTP

Let’s iterate over pages and make some other improvements

import requests

url = "https://swapi.dev/api/planets/"
while url:
    response = requests.get(url, timeout=10)
    response.raise_for_status()
    data = response.json()
    for planet in data["results"]:
        print(planet["name"], planet["population"], planet["url"], sep=" — ")
    url = data["next"]

HTTP Serverbonus

You can create your own HTTP applications with Python too!
We’ll cover that later. For now, useful keywords are Flask, FastAPI, Django, uvicorn, and gunicorn