CineMate

---

Table of Contents

Table of Contents

• Project Background
• Features
  • Roadmap
• Code Structure & Logic
  • Data Processing
  • Data Model
    • Movie Class
    • Movies Class
    • UML Class Diagram
    • Flowchart
• Technologies Used
  • Libraries & Modules
• Testing
  • Manual Test Process
  • Error Handling
  • PEP8 Validation
  • Bugs
    • Resolved
    • Unresolved
• Deployment
  • Pre-deployment (Optional)
  • Deployment Steps
• Credits
• Resources

---

Project Background

CineMate is a Python command line application designed to make it easy for users to browse and explore movies. Powered by an extensive dataset from IMDB, CineMate focuses on the top 1000 rated movies since the year 2000, offering users a curated selection of films.

The primary objective of CineMate is to provide users with an effortless movie discovery process. With CineMate, users can search for movies, navigate through them by year or genre, and uncover hidden cinematic gems. By harnessing the capabilities of Python, CineMate simplifies the task of finding and selecting movies, catering to both casual viewers and avid film enthusiasts.

CineMate offers features that enrich the user experience. Users can add their favorite movies to a personalized favorites list, ensuring easy access to beloved films. Furthermore, CineMate enables users to maintain a watched list, allowing them to keep track of their movie-watching history and receive personalized recommendations based on their preferences.

With CineMate, finding your next movie masterpiece has never been easier.

---

Features

• On initialisation, CineMate shows users a home menu which clearly communicates the intent of the application and provides the user with options for navigation.

  Screenshot - Homescreen

  

• Favourites/Watch List: CineMate allows users add movies to a favourites list for movies they particularly liked. There is also a 'watch list' feature which lets users easily save movies for later viewing. Movies can be easily added or removed from these lists.

  Screenshot - Favourites

  

• Browsing: CineMate lets user browse movies by search, genre or release year.

  • Search: Users can enter a search query which will compare the movie titles against the query and show any relevant results.
  • Genre: Users can select a genre they are interested in and CineMate will show all relevant movies within that genre.
  • Year: Users can select a specific year and CineMate will show relevant movies released in that year.
  Screenshot - Browse by Genre

  

  Screenshot - Browse by Search

  

• UI: The UI has some basic styling applied which makes information easily readable:

  • Movies are printed with a blue background while the list headers are printed in green.
  • User prompts are coloured yellow so they stand out.
  • Menu options are green to differentiate them from other information on screen.
  • Error messages are red to indicate a problem.

Roadmap

• Improve browsing experience by letting users select multiple genres or enter a range of years to browse by.
• Paginate movie lists to avoid having long lists print on screen.

---

Code Structure & Logic

Data Processing

• Two .tsv datasets from IMDB were used in this project - one file containing movie titles, release dates and genres; and the other containing movie ratings. The dataset had over 9 million rows of data and included documentaries, shorts, TV shows etc. For the purposes of this project, the dataset needed to be filtered down to a usable set of movies only.
• Before development of the main application (run.py), a Python script was written to process the datasets.
• Utilising the pandas library, the script data_cleanup.py reads the two .tsv files, removes irrelevant columns, applies filters, merges the two datasets and finally exports a new .tsv file containing the top 1000 movies.
• The filters applied are as follows:
  • Type: Movies
  • Runtime: 90 minutes or longer
  • Release Year: 2000 onwards
  • Avg. Rating: 7.0 with 50,000+ votes
  • Sorted by Number of Votes and trimmed to 1000 rows
• This process could have been carried out in the main application, but would have led to significant wait times on initial loading of the application due to the amount of data in the datasets.

Code Block

Data Model

Movie Class

• A Movie Class was created to contain the data for each movie instance from the dataset.
• The Movie Class has the following attributes:
  • index: List index number when printing list of movies
  • title: Title of movie
  • date: Release year of movie
  • runtime: Runtime in minutes
  • genres: Genres applicable to movie
  • rating: Average rating from IMDB ratings
  • votes: Number of rating votes from IMDB ratings
• The __str__ method of this Class prints the movie in a tabular format.

Movies Class

• A Movies Class was created to contain the full dataset of movies in a list
• The contructor takes two arguments:
  • top_100: A global variable containing the Top 100 movies list which is created upon initialisation by the create_top_100() function.
  • all_movies: A global variable containing the full list of movies from the dataset which is created upon initialisation by the get_movies() function.
• The Movies Class has the following attributes:
  • top_100: A list of the Top 100 movies; created upon initialisation of the program.
  • all_movies: A list of all movies in the dataset; created upon initialisation of the program.
  • favourites: A list of user-selected favourite movies
  • watchlist: A list of user-selected movies intended to be watched later by the user
• The print_movies method takes either top_100 or all_movies as arguments and prints them in a tabular list.

UML Class Diagram

The below diagram shows the relationship between the Movie and Movies Classes:

Flowchart

The following flowchart shows the logic of the code for the program:

---

Technologies Used

• The application was developed with Python
• A non-commercial freely available dataset from IMDB was used and consists of two subsets: title.basics.tsv; title.ratings.tsv (Available here)
  • Note: These files are not included in the GitHub Repository due to their large size
• Heroku was used for deployment
• Jupyter Notebooks and VS Code were used for development
• mermaid.live was used to generate the UML Class Diagram

Libraries & Modules

• The Pandas library was used for processing the original datasets.
• The built-in csv module was used for reading the .tsv movies data file. The built-in textwrap module was used to shorten long movie titles so they would print within the width of the terminal window.

---

Testing

Manual Test Process

Test	Action	Success Criteria
Navigation	Starting from the home menu, navigate through the application, checking that correct action is taken according to user input	Navigation works as expected. Accepts only inputs shown on menu options. Invalid inputs are handled. Successfully navigate from main menu to sub-menus and vice versa.
User input	Test all instances where user input is required. Enter valid and invalid inputs.	Invalid inputs are handled gracefully. Error message shown. User prompted to try again.
Browse movies by search	Enter search query that matches a known movie in the dataset. Enter random invalid search query.	Valid queries show relevant search results. Invalid queries are handled gracefully.
Browse movies by year	Browse by year at both ends of accepted input limits. Browse by year within accepted input limits. Browse by year outside of accepted input limits.	Relevant movies shown according to year input. Invalid inputs outside of accepted range are handled gracefully.
Browsing movies by genre	Test browsing by 5 different genres.	Genre list prints as expected. Input matches list index. Relevant movies shown according to selected genre.
Movie list presentation	Check: show all; show top 100; show favourites; show watch list; browse by genre/year/search to ensure lists print as intended	Only relevant movie lists print for the given context. Index numbering is correct.
Favourites/watch list	Add 5 random movies to both favourites and watch list. View favourites/watch list. Remove all movies from lists. y. Add to lists from various contexts (e.g. from search results, top 100 etc.)	Correct movies added to list. Lists print correctly. Feedback messages shown on add/remove. Empty list message shown if empty.

Error Handling

• Invalid user inputs are handled gracefully and a feedback message is shown to users on invalid input submission.
• This includes:
  • Only accepting input within the given range for menu selections
  • Not allowing users to add the same movie to favourites/watchlist
  • Removing the 'Remove from favourites/watchlist' option when the list is empty to prevent user error
  • Showing a feedback message if no movies match user search query
  • Prompting user to select a valid option within accepted range (e.g. for movie indexes, genre list, browse by year etc.)
  • Acceptable inputs are not case-sensitive (e.g. for Y/N prompts)

PEP8 Validation

• The code for both run.py and data_cleanup.py were passed through a PEP8 validator with no issues present.

Validation Screenshot

Bugs

Resolved

1. When initially testing the 'Add to favourites' feature, I found that when trying to add a favourite from a movie list other than "Show all movies" (e.g. adding a favourite from a list of search results), an incorrect movie would be added. This was due to the add-to-favourites function referencing the main list of movies rather than e.g. the search results list. This was fixed by passing an argument to the function to check where the user is coming from i.e. check whether they are trying to add a favourite from a list of search results, top 100 list etc.
2. During development, I noticed a bug where the first 100 movies in "Show all movies" would print with incorrect index numbers, only after "Showing top 100". This was due to the Top 100 movies being assigned a new index when printing, and not being reassigned to their original value afterwards. This was fixed by using enumerate() in the Movies.print_movies() method to reassign the correct index when printing.

Index bug screenshot

Unresolved

• No unresolved bugs

---

Deployment

Pre-deployment (Optional)

Note: This step is only required if you want to change the filters applied to the processed dataset, or if you want to download the latest data from IMBD (IMDB updates datasets daily)

1. Download the datasets: title.basics.tsv; title.ratings.tsv (Available here)
2. Download the repository (or just download data_cleanup.py)
3. Ensure you have the Pandas library installed with the command pip install pandas
4. Open data_cleanup.py. Specify filepaths to read the datasets and your desired OUTPUT_PATH on lines 4, 7 and 8.

# CHOOSE PATH FOR NEW TSV OUTPUT
OUTPUT_PATH = 'SPECIFY OUTPUT PATH HERE' 

# OPEN IMDB TSV FILES
df_basics = pd.read_table('YOUR_PATH/title.basics.tsv', low_memory=False)
df_ratings = pd.read_table('YOUR_PATH/title.ratings.tsv', low_memory=False)

5. If desired, change the filters applied to the data to your liking as noted in the docstrings. The default filters are specified at Data Processing.
6. Run the script - if successful, a new .tsv file will be found at your specified OUTPUT_PATH

Deployment Steps

Cloning the Repo

1. Click on the "Code" button near the top right corner of the page.
2. Copy the HTTPS or SSH URL that appears in the box. Open your terminal (or Git Bash on Windows) and navigate to the directory where you want to clone the repository.
3. Type "git clone" followed by a space, and then paste the URL you copied in step 3.
4. Press enter to run the command. This will clone the repository onto your local machine.
5. You should now have a local copy of the GitHub repository on your machine.

Forking the Repo

1. Click the "Fork" button near the top right corner of the page. This will create a copy of the repository in your own GitHub account.
2. Once the fork is complete, you will be redirected to the forked repository in your account.
3. If you haven't already, clone the forked repository to your local machine using the steps outlined in the previous answer.
4. Make any changes or additions you want to the code in your local copy of the repository.
5. Commit your changes to your local repository using the "git commit" command.
6. Push your changes to the forked repository on GitHub using the "git push" command.
7. If you want to contribute your changes back to the original repository, create a pull request by going to the original repository's page and clicking the "New pull request" button. From there, you can compare your changes to the original repository and request that they be merged.
8. You should now have a forked copy of the GitHub repository in your account, and you can make changes to it and contribute back to the original repository if desired.

Deployment to Heroku

Note: A Heroku account is required for deployment to Heroku.

1. From the Dashboard, click "New" - "Create new app".

2. Enter a name for the app. Click "Create App".

3. Connect your GitHub account and select the repository and branch to deploy.

4. When you create the app, you will need to add two buildpacks from the Settings tab. The ordering is as follows:

   • heroku/python
   • heroku/nodejs

   

5. You must then create a Config Var called PORT. Set this to 8000.

---

Credits

• Code Institute repository template used for deployment
• IMDB for the datasets utilised (Available here)
• The Pandas library was used for processing the original datasets.

Resources

• Harvard CS50’s Introduction to Computer Science (CS50x)
• Harvard CS50’s Introduction to Programming with Python (CS50P)
• The Complete Python Bootcamp From Zero to Hero in Python, Jose Portilla (Udemy)