Uli Köhler

If you have an URL like

https://logodix.com/logo/1667872.jpg?param=value

and you want to find just the filename extension (.jpg for this example), use the following code:

from urllib.parse import urlsplit
import os.path

url = "https://logodix.com/logo/1667872.jpg?param=value"

path = urlsplit(url).path
extension = os.path.splitext(path)[-1] # e.g. ".jpg"

Problem:

When trying to urlsplit an URL in Python 3:

from urllib.parse import urlparse

path = urlparse.urlsplit(remote_image_url).path

you see the following error message:

AttributeError                            Traceback (most recent call last)
Input In [30], in <cell line: 3>()
      1 from urllib.parse import urlparse
----> 3 path = urlparse.urlsplit(remote_image_url).path
      4 filename = posixpath.basename(path)

AttributeError: 'function' object has no attribute 'urlsplit'

Solution:

The equivalent of urlparse.urlsplit() in Python 3 is urllib.parse.urlsplit().

Therefore, a working code example is

from urllib.parse import urlsplit

path = urlsplit(remote_image_url).path

Problem:

While trying to import urlparse in Python 3 using

import urlparse

you see the following error message:

---------------------------------------------------------------------------
ModuleNotFoundError                       Traceback (most recent call last)
Input In [29], in <cell line: 1>()
----> 1 import urlparse

ModuleNotFoundError: No module named 'urlparse'

Solution:

In Python 3, urlparse has been renamed to urllib.urlparse. In order to import it, use this line instead:

from urllib.parse import urlparse

Assuming you have a Python script server.py that you want to auto-reload every time the file changes, use the following script using nodemon:

nodemon -w . --exec python server.py

The -w . tells  nodemon files to watch for changes of all files in the current directory (.)

I generally recommend creating a script start.sh to automatically run this command:

#!/bin/sh
nodemon -w . --exec python server.py

and then make it executable using

chmod a+x start.sh

Now you can run it using

./start.sh

Returning a JSON array in bottle involves setting the response content type to application/json and then manually dumping the array using json.dumps():

#!/usr/bin/env python3
from bottle import route, run, response
import json

@route('/')
def json_array():
    response.content_type = 'application/json'
    return json.dumps(["a", "b", "c"])

run(host='localhost', port=9000)

Note: I do not recommend using entr any more since for me it doesn’t work reliably. I recommend using nodemon instead, see How to auto-restart bottle server when Python file changes using nodemon

Assuming you have a Python script server.py that you want to auto-reload every time the file changes, use the following script using entr:

ls *.py | entr python server.py

The ls *.py tells entr which files to watch for changes.

This script uses a dedicated class MyServer to encapsulate the bottle server appropriately. Compared to our previous post Python bottle minimal example it provides better encapsulation of the server at the cost of slightly more lines of code.

#!/usr/bin/env python3
from bottle import Bottle, run

class MyServer(object):
    def __init__(self):
        self.app = Bottle()
        self.init_routes()
        
    def init_routes(self):
        """Initialize all routes"""
        @self.app.route('/hello')
        def hello():
            return "Hello World!"

    def run(self):
        run(self.app, host='0.0.0.0', port=8080)
        
# Example usage
if __name__ == "__main__":
    server = MyServer()
    server.run()

How to use

Run the server

python bottle-server.py

and open

http://localhost:8080/hello

Now you should see the Hello World! message.

PyInvoke provides a much easier interface to running CLI processes than subprocess.