The Python REPL, scripts and the -m switch

There are three ways to give Python some code: type it into the interactive shell, save it in a file and run the file, or pass it on the command line. Each has its place, and knowing which one you are in explains a lot of “it worked here but not there” surprises.

The interactive shell

Run python3 (on Windows, py) with nothing after it. Python prints a short banner and the >>> prompt: you are in the interactive shell, also called the REPL, for read, evaluate, print, loop. Type an expression and press Enter; Python evaluates it and prints its value:

>>> 2 ** 100
1267650600228229401496703205376
>>> _ * 2
2535301200456458802993406410752
>>> name = "Asha"
>>> name.upper()
'ASHA'
>>> "upper" in dir(name)
True

Lines starting with >>> are what was typed; the others are what Python printed. Four things to notice:

• ** raises a number to a power, so 2 ** 100 is $2^{100}$. Python’s whole numbers have no size limit, so the shell prints all 31 digits instead of rounding the result.
• The shell prints the value of an expression by itself. A script does not: there you need print().
• _ (an underscore) holds the last value the shell printed, which saves retyping it. It only works in the shell.
• dir(name) lists the names an object has, its methods and attributes; help(name.upper) shows the documentation of one of them. In a help page, press q to get back to the prompt.

To leave the shell, type exit (in the older basic shell, exit()), or press Ctrl-D on macOS and Linux, or Ctrl-Z and then Enter on Windows.

The shell is for trying things out. Anything you want to keep belongs in a file.

Running a script

A script is a file of Python code. python3 hello.py runs the file hello.py from the current folder, from the first line to the last, and exits. Add -i (python3 -i hello.py) to stay in the interactive shell after the script has finished, with all its variables still there to inspect.

One-liners with -c

For a quick calculation, pass the code itself with -c: python3 -c "print(2 ** 100)". Put the code in quotes so the terminal hands it to Python as one piece.

Running modules with -m

python3 -m name finds the module called name, the way import name would, and runs it as a program. Many standard-library modules are useful this way:

• python3 -m venv .venv creates a virtual environment (a later lesson).
• python3 -m pip install … installs packages with pip.
• python3 -m json data.json checks a JSON file and prints it neatly indented (new in 3.14; python3 -m json.tool also works).
• python3 -m http.server serves the files of the current folder at http://localhost:8000/ until you press Ctrl-C. Other devices on your network can reach it too, so run it only in a folder you are happy to share.

The module comes from the same interpreter as the python3 you typed. That is why python3 -m pip is the safe way to install packages: it installs them for that Python, even when a pip command on your PATH belongs to another one.

Command-line arguments with sys.argv

Words you type after the script’s name reach the program in sys.argv, a list of strings. Item 0 is the script’s name as you typed it; the arguments follow:

import sys

# A terminal fills sys.argv from the command line:  python3 args_demo.py Asha 3
# A browser has no command line, so when no arguments were given we set the same list here.
if len(sys.argv) <= 1:
    sys.argv = ["args_demo.py", "Asha", "3"]

print("Script name:", sys.argv[0])
print("Arguments:", sys.argv[1:])
print("Number of arguments:", len(sys.argv) - 1)

Script name: args_demo.py
Arguments: ['Asha', '3']
Number of arguments: 2

Run it as python3 args_demo.py Ravi 5 extra and it prints those three arguments instead. A browser has no command line, so when the program gets no arguments, as in a browser runner, it sets the list itself, the way a terminal would have.

Key takeaways

• python3 alone starts the interactive shell, which prints each value; leave it with exit or Ctrl-D (Ctrl-Z and Enter on Windows).
• python3 file.py runs a script, -i keeps the shell open afterwards, and -c "code" runs a one-liner.
• python3 -m module runs a module with the same interpreter, which is why python3 -m pip installs to the right Python.
• sys.argv holds the script name and the arguments, all as strings.