https://github.com/variable/django-rest-framework-queryset 

I think I will just share my working sample code here to save someone’s time

import requests

# local file
with open('image.jpg', 'rb') as file:
    resp = requests.post('http://localhost:8888/image', data=file.read())
    url = resp.headers['Location']

# remote file
remote_resp = requests.get(image_url, stream=True)
resp = requests.post('http://localhost:8888/image', data=remote_resp.content)
url = resp.headers['Location']

from .your_module import unused_stuff  # noqa

Peace.

Don’t assign different value types to the same variable

Let’s see what dynamic type in the simplest form in Python:

variable_a = True
if variable_a:
    variable_a = 'Result'

While this is perfectly legitimate, but it’s perfectly unacceptable, the rule of thumb is, you never assign a different type of value to a already defined variable.

variable_a = True
if variable_a:
    result = 'Result'

This way, when someone reads your code, one can safely assume result is always a ‘String’ type, seriously this will help a lot.

Don’t return different types of value in a function

Consider the following example, one would need to understand your condition in order to determine the type of result, also makes it difficult to consume this function.

def work():
    result = None
    variable_a = True
    if variable_a:
        result = 'Result'
    else
        result = 123

Define your class members at the top of the class

While this is perfectly legitimate:

class A(object):
    def work(self):
        self.member = 'abc'

But it’s difficult to work out what class members at a glance, please define them.

class A(object):
    def __init__(self):
        self.member = None

    def work(self):
        self.member = 'abc'

Now you know this class has a class member without reading the work function.

Exception is first class citizen

try:
    a = 1 / 0
except:
    print 'Caught'

Prints ‘Caught’

Let’s look at PHP:

try{ 
    $a = 1 / 0; 
}catch (Exception $e){ 
    print "Caught"; 
} 
I cannot believe by default this still gives me PHP Warning:  Division by zero

Better functions

Imagine you can write functions like this

def marry(man, woman):
    ...

marry('John', 'Julie')
marry(woman='Julie', man='John')

def party(*args):
    for person in args:
      print person

party('John')
party('John', 'Ben', ...)

def save(name, *kwargs):
    data = {'name': name}
    data.update(kwargs)
    write_to_db(data)

save('John')
save('John', gender='male', birthday='01/01/1970')

Short-lived vs Long-lived

Almost all PHP scripts are short-lived, including most frameworks, which means to serve a request, your PHP application will run from start to finish. However, Python can run as processes to serve requests, for example, A Django application runs as a process, so initiation code only runs once, serve a request, then wait for the next request to serve.

Imports Properly

Since 5.3, PHP introduced namespace, since they used .”dot” for concatenation, so they had to use the \”slash”, looks like this

namespace NZ\NET\LIN\Test

And use like this

use NZ\NET\LIN\Test

But it doesn’t actually import the Test class! Unless you use __autoload!

I am not a sales person so I don’t know how to sell it, but it’s not hard from a technical perspective to, let’s say consolidate all real estate websites in NZ into a platform.

If we look at all these real estate listings, they basically consisted with 2 models, the “listing” and “agent”. Having those 2 models stores 90% of the information needed. Surely we are not stopping at those 2 models, we will normalise the repetitive fields into models defined such as “location” , “company”(if we e want to have just one installation to serve multiple company websites”, etc etc. I think for implementing this is using Django, creating models is a breeze and using Django Rest Framework serving as APIs.

Slapping on a comprehensive search function using ElasticSearch, easy.

Since now we have a backend and communicating via API calls, then implementing the frontends is rather a freedom of choice for different companies. Lots of options (Django, NodeJS etc etc)

Side note:

These real estate companies will most probably have their own CRM systems, might be a bit of pain the implement the integration and probably having each installation for every company is a better choice.

from

self.execute_command(command, cwd=os.path.dirname(infile))

to

logger.debug(self.execute_command(command, cwd=os.path.dirname(infile)))

One of the team lead pointed out the statement has a problem, and it was a life saver, but I won’t tell you why just yet, see if you can spot it.

import cv2, time

cap = cv2.VideoCapture(0)
time.sleep(1)
cascade = cv2.CascadeClassifier("haarcascade_frontalface_alt.xml")

def detect(image):
    #faces = cascade.detectMultiScale(img, 1.1, 2, cv2.cv.CV_HAAR_SCALE_IMAGE, (20,20), (200,200))
    faces = cascade.detectMultiScale(image)
    for _face in faces:
        cv2.rectangle(image, (_face[0], _face[1]), (_face[0]+_face[2], _face[1]+_face[3]), (255,255,255))

def repeat():
    ret, image = cap.read()
    detect(image)
    cv2.imshow("w1", image)
    cv2.waitKey(1)

while True:
    repeat()

Installation:

1. pip install numpy (sometimes you might need to run as sudo)
2. brew tap homebrew/science
3. brew install opencv
4. export PYTHONPATH=/usr/local/Cellar/opencv/2.4.7.1/lib/python2.7/site-packages/
5. download the xml file https://warai.googlecode.com/files/haarcascade_frontalface_alt.xml and place it in the same folder.

import datetime
from haystack import indexes   
from cms.models.managers import PageManager
from cms.models.pagemodel import Page
from cms.models.pluginmodel import CMSPlugin


class PageIndex(indexes.SearchIndex, indexes.Indexable):
    text = indexes.CharField(document=True, use_template=False)
    pub_date = indexes.DateTimeField(model_attr='publication_date', null=True)
    login_required = indexes.BooleanField(model_attr='login_required')
    url = indexes.CharField(model_attr='get_absolute_url')
    title = indexes.CharField(model_attr='get_title')

    def prepare(self, obj):    
        self.prepared_data = super(PageIndex, self).prepare(obj)
        plugins = CMSPlugin.objects.filter(placeholder__in=obj.placeholders.all())
        text = ''              
        for plugin in plugins: 
            instance, _ = plugin.get_plugin_instance()
            if hasattr(instance, 'search_fields'):
                text += ''.join(getattr(instance, field) for field in instance.search_fields)
        text += obj.get_meta_description() or u''
        text += obj.get_title() or u''
        text += obj.get_meta_keywords() if hasattr(obj, 'get_meta_keywords') and obj.get_meta_keywords() else u''
        self.prepared_data['text'] = text
        return self.prepared_data       

    def get_model(self):
        return Page

    def index_queryset(self, using=None):
        return Page.objects.published().filter(publisher_is_draft=False).distinct()