madebyloosers/favouriteq

What's your favourite question? Follow twitter.com/favouriteq for daily questions, this web app aggregate the answers...

#favouriteQ

What's your favourite question?

Follow @favouriteq for daily questions

Another Gunt London production

---

##Installation

###Database

Create a mysql database called favouriteq

Create tables for apps that aren't using south

./manage.py syncdb

Create tables for apps that are using south. Data will be loaded from fixtures

./manage.py migrate

During development developers should continue to run the ./manage.py migrate command to keep their DB up to date.

###Virtual environment

Create virtual environment

virtualenv --no-site-packages env

(if you don't have virtualenv, install it: sudo pip install virtualenv)

(if you don't have pip, install it: sudo easy_install pip)

Activate virtual environment

source env/bin/activate

###Dependencies

From the root

pip install -r requirements/requirements.txt

###Run

From the root

cd favouriteQ
python manage.py runserver

Open http://127.0.0.1:8000/ in a browser

###Deploy Live

Note you will need to have installed requirements so you have fabric on your machine.

From the root

cd favouriteQ
fab production deploy

If you want less verbose output

fab production deploy --hide=stdout

favouriteq's Issues

Send out a summary tweet at the end of a question?

Consider sending a summary tweet at the end of a question - i.e. just before a new question is posted.

The summary tweet could feature notes about the previous question, e.g. how many people have replied, if there was a most popular answer, maybe a link to the archive page.

Fix 'tweet your answer' link layout for mobile

At the moment it doesn't fit in to a single line, so half the link goes off the page. Add responsive CSS fix to make it display properly...

Use model property for person.name

Call person.name on model instead of adding the names together in the view. Technique is documented here https://docs.djangoproject.com/en/dev/topics/db/models/

Save all tweets to the DB

This would require a DB migration but would be easier to track what's happen with tweets processed by the system.

Get users real name from Twitter

Get users real name from Twitter as well as the handle

Remove production setting from git

This is bad practice. Add them to a private repo (bit bucket?)

Add page for each person

Page for each person showing all their answers

Change production site config to not show debug details

CMS: show list or filter for non-approved questions

Add infographic / pie chart to each page summarising the top answers

Think of some clever SQL to aggregate the answers, pull out the top 10 and how many times each has been answered. Then we can feed these figures into a Chart/Graph API to create infographix

Make question text white on archive list page

SCSS files shouldn't be deployed to the live site

http://favouritequestion.com/static/questions/scss/

Move question app urls to their own file

Tweak nav

Change link to just 'suggest'

Put in a 'home' link (test between archive and home pages)

Catch exception when they are no questions to ask

python post_to_twitter.py
Traceback (most recent call last):
File "post_to_twitter.py", line 16, in
question = Question.objects.get_new_question()
File "/Users/pxg/Sites/favourite_q_django/favouriteQ/questions/models.py", line 15, in get_new_question
return questions[0]
File "/Users/pxg/Sites/favourite_q_django/env/lib/python2.7/site-packages/django/db/models/query.py", line 207, in getitem
return list(qs)[0]
IndexError: list index out of range

Friday Favourite

We should offer of a 6-pack of Guntweisers to the answer we think is best on Friday

Render form with Django forms

Set up 404 error page in Django

Current this looks really bad http://favouritequestion.com/archive/123456

Allow people to answer archive questions from twitter

If people tweet at @favouriteQ with a specific #hashtag allow them to answer an archived question

Combine shared render code in views.py

Copy/Design needed for when no answers have come in yet

Add unit tests to twitter processing scripts

Move twitter handle to settings file.

Favourites Favourite

We can choose a favourite answer of the day. This shows on the site with a star by the
answer (maybe it's moved to the top too?). We could tweet the favourite when it's selected.

Remove console.log when the panels are opened/closed

Redirect for www.favouritequestion.com

Kill the www. @thegingerbloke says it's for loosers.

Add search to archive page

Add incrementing number to current question page

Complete #8 first. As an enhancement it would be better if the number matched the order in which the questions have been asked rather than the DB id.

Next question

Add the ability to select the next question - this should de-prioritise all other questions (unlike the existing priority filter which allows multiple questions to have priority)

Add a 'maxlength' property to the suggest a question textarea

At the moment its absence is causing a JavaScript error. It should be present so we can display the character countdown.

Is there a character limit on the textarea? If not I suggest 140, as that's the number of characters that'll fit in a tweet. Shorter if we're also planning on tweeting who suggested the question.

(e.g. "Who is your favourite ginger? (via @thegingerbloke)"

Add question archive page

List all the asked Questions, then let the user drill-down to see the answers to each question.

Remove duplication of code in view.py

detail and current_question have duplicated code for dealing with the form. How can we have them share this code?

Work in home button to menu

Add question number to current question page

To get working quickly use the DB id.

Do something when there are no active questions left

Tweet a request for more questions instead (with a link that deeplinks to the #submit form)

Maybe also email us to remind us to set a question - although we should really see the tweet and do this ourselves anyway

Search twitter cron job log instead of print

The current search twitter script prints it's output. This is good for debugging, however when it runs on cron the output is emailed to ec2-user this results in 1440 emails a day which is silly.

Research a better way to log this output. Would be nice to have a switch so the output can still be seen when scripts are run from the command line.

Automatic text processing of user-generated questions

Make all questions consistent:

• Add a question mark to the end if there isn't one
• Capitalise the first letter if not done already
• Change "what is" to "what's"

(etc)

Set server time to GMT

Currently on UTC. We're British god damn it.

Add slugs for archived questions

Use slugs like favouritequestion.com/q/whats-your-favourite-question Google will lap this up for SEO

Add favicon

Have gunicorn run with gunicorn.conf.py

http://honza.ca/2011/05/deploying-django-with-nginx-and-gunicorn

Catch exceptions with the Twitter API

Research the best way to log these in Python/Django.

Indexes should be switched off on the live site

http://favouritequestion.com/static/questions/scss/

Change daily tweet time to 2pm (for better global samosa-ing)

Allow deeplinking to http://favouritequestion.com/#suggest to open the suggest panel by default

Allow deeplinking to http://favouritequestion.com/#suggest to open the suggest panel by default.

So we can link straight to the 'suggest a question' form

Allow people to post new questions from twitter

If someone tweets at @favouriteQ using a specific #hashtag then pull it in as a question

Remove "tweet your answer from archive page"

Generate new Twitter API keys

The current keys can be found in the Github history.

1. Generate new API keys
2. Update them in the private config repo
3. Deploy to the live server - test they work (run the cron tasks manually)
4. Delete the old keys.

Create a weekly tweet requesting new questions

This should deeplink to the suggest form:

http://favouritequestion.com/#suggest

Move twitter API access tokens out of repo