Giters

jordanlibrande / SFMovies

Front-end & RESTful Back-end to display locations in SF that have been in movies

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

#SFMovies UberFarmer's application for the Uber Coding Challenge.

Project: SF Movies (this may be obvious...)

Check out the example front-end in action at http://uberfarmer.github.io/SFMovies/

Usage

Before you start, set these two environment variables: DJANGO_SFMOVIES_SECRET_KEY and GOOGLE_API_KEY. Use Django to generate you a new random secret key, and use your own Google API Key.

Get your database into shape with the command python manage.py syncdb , run sfmovies.updatedb.update_db_if_old() to fill the database with data, and then run the webserver locally using the command python manage.py runserver. Once it's up and running, you can point the front-end to your new SFMovies API Endpoint.

Enjoy exploring San Francisco's rich motion picture history with SFMovies!

SFMovies RESTful API Documentation

SFMovies has three API calls, all of which return JSON[P] data. They are located under /sfmovies/services/.

###GET locations

  • Required parameter: bounds=latlo,lnglo,lathi,lnghi where latlo is the West boundary, lathi is the East boundary, lnglo is the South boundary, and lnghi is the North boundary.
  • Optional parameter: title=x where x is a movie title.
  • Returns information about up to 50 movie locations within the specified region. If the title option is set, locations from that movie will be returned. Example API call:
GET examplehost/sfmovies/services/locations?bounds=37.702607%2C-122.516062%2C37.835453%2C-122.338049&movie_title=The+Assassination+of+Richard+Nixon

Example API response:

[
  {"latitude": 37.7976, "longitude": -122.4059, "location_id": 727, "movie_title": "The Assassination of Richard Nixon"},
  {"latitude": 37.7889, "longitude": -122.4131, "location_id": 728, "movie_title": "The Assassination of Richard Nixon"}
]

###GET location_detail

  • Required parameter: location_id=id where id is the unique id of a location. Location ids can be discovered through the GET sfmovies/services/locations call.
  • Returns information about that location Example API call:
GET examplehost/sfmovies/services/location_detail?location_id=728

Example API response:

{"Actor 1": "Sean Penn", "Longitude": -122.4131, "Writer": "Niels Mueller & Kevin Kennedy", "Actor 3": null, "Production Company": "Anhelo Productions", "Fun Facts": null, "Actor 2": "Naomi Watts", "Title": "The Assassination of Richard Nixon", "Locations": "766 Sutter Street", "Release Year": 2004, "Distributor": "THINKFilm", "Director": "Niels Mueller", "Latitude": 37.7889, "location_id": 728}

###GET title_autocomplete

  • Required parameter: term=x where x is the beginning of a movie title. term was chosen as the parameter name for easy compatability with the jQueryUI autocomplete widget.
  • Returns up to 5 movie titles matching the provided beginning of title. Example API call:
GET examplehost/sfmovies/services/title_autocomplete?term=f

Example API response:

[  
  {  "label":"Family Plot",  "value":"Family Plot" },
  {  "label":"Fat Man and Little Boy", "value":"Fat Man and Little Boy" },
  {  "label":"Fearless", "value":"Fearless" },
  {  "label":"Final Analysis", "value":"Final Analysis" },
  {  "label":"Flower Drum Song", "value":"Flower Drum Song" }
]

Project Discussion

####Technical Track: Primarily back-end. (Although I also created a reasonable but not super-polished front-end).

####Assumptions:

  • The user wants to pan and zoom the map around San Francisco and see where movie scenes have been shot
  • The user wants to be able to tap on a location and get more information about the movie and scene that was shot there.
  • Returning 50 locations at a time is a reasonable upper limit -- A front-end could always choose to display fewer of them.
  • Returning 5 autocomplete suggestions at a time is a reasonable number.
  • Movie title will be the most common parameter users will want to filter by.

####Technical Choices + Experience: Note: Prior to this project, I have had almost no experience with web development. While I tried to choose technologies that would let me accomplish the job, I leaned toward widely-adopted technologies that I felt I should have experience with.

######Back-end:

  • Python 3.4: I chose Python because I was proficient with it and wanted to use Django as my web framework. I prefer Python 3 when I don't have to care about backwards-compatibility (though I didn't end up using very many Python 3 features).
    • Previous experience: I was already fluent in Python before this project.
  • Django: I chose Django as my back-end web framework because it was well-developed, well-documented, and widely used. In addition, it was recommended by one of my friends and it has a built-in testing framework. Lastly, I wanted to learn it anyway. I knew from the start there was a chance it would be more complicated and heavy than was necessary.
    • Previous experience: None (no experience with any web framework)
  • SQLite3: Default database for Django. Didn't see any reason to switch.
    • Previous experience: Some with other databases, none with SQLite.
  • Requests: I used this to simplify HTTP requests from Python.
    • Previous experience: None
  • Google Maps Geocoding API: The obvious choice to convert from locations into GPS coordinates.
    • Previous experience: None
  • SFData SODA API: I chose to collect the data through their JSON web API so it would be easy to update my database if SFData's movies database was updated.
    • Previous experience: None
  • Amazon AWS: I chose this for my back-end hosting because I was sure it would be sufficient and I wanted experience using it.
    • Previous experience: None
  • Apache HTTP Server: Seemed like a common server choice with Django.
    • Previous experience: None

######Front-end:

  • HTML/Javascript/CSS: The only choice for modern web development. As I was planning to make a bare-bones front-end experience (and I wanted more experience with these technologies) I wrote these straight (using some snippets from the web) instead of using a front-end framework.
    • Previous experience: Minimal experience with HTML, Minimal experience with Javascript (no experience with Javascript on the web), Minimal experience with CSS.
  • jQuery: I chose to use jQuery to simplify my client-side scripting (in particular AJAX calls). It looked to be widely-adopted and well-supported.
    • Previous experience: None
  • jQuery UI: I used this library for their autocomplete widget, which did all of the hard work for me. I chose to get my widget from jQuery UI because it seemed widely-adopted and well-supported.
    • Previous experience: None
  • Google Maps Javascript API: I wanted to plot things on a map. This seemed like the choice.
    • Previous experience: None
  • GitHub: front-end hosting
    • Previous experience: None

######Misc:

  • GitHub: source code hosting + version control
    • Previous experience: None
  • Zero to minimal experience with: REST/RESTful APIs, HTTP requests

####Trade-Offs and Future Work:

  • As suspected, Django was too complicated in areas I didn't care about (e.g. templating and formatting HTML, site administration, authentication), and didn't provide enough support for some things I wanted to do (sending HTTP GET requests, processing JSON, sending JSON and JSONP responses). If I was doing this project over again, I would try a different back-end framework or use a Django extension designed with REST/JSON in mind.
  • I could have devoted more time cleaning up the front-end, adding more styles, formatting the location detail text, and using a more intelligent algorithm for marker updates.
  • I would add more options to the back-end API. One example: letting users choose the number of locations they want info about and specifying data offsets (like the SODA API).
  • I would spend more time on the back-end tests.
  • I would come up with a smarter way to update my database using the DataSF site than by wiping my local database and starting over
  • Some of the back-end code could be modularized and refactored. For example, optionally turning JSON responses into JSONP could be less repetitive.
  • Stretch goals/features:
    • Links to imdb pages of the movies
    • Which scene of the movie took place in that location (with start+end times)? (super-strech goal: show the actual scene from the movie)
    • Fancier search filtering
    • Expand beyond SF
    • Geolocalization for better mobile app use
    • Monetization (Movie ads!??!)
    • API Keys

####Boilerplate Code: Files without [much] contribution from me:

mysite/
    manage.py
    mysite/
        __init__.py
        settings.py
        urls.py
        wsgi.py
    sfmovies/
        admin.py
    templates/
        admin/
            base_site.html

####LinkedIn Profile

About

Front-end & RESTful Back-end to display locations in SF that have been in movies

Languages

Language:Python 100.0%