Task $1$: Parallel Corpora

Parallel corpora contain a collection of texts in a given language and their translation to one or more other

languages. In this task, you will build a small parallel corpus using data from OpenSubtitles.org, a

database that allows you to search and download subtitles for various languages. It was previously used to

build the OpenSubtitles corpus, which consists of around $2\mathrm{.}6$ billion sentences and covers $60$ languages.

Search for the film Monty Python and the Holy Grail $(1975)$ on OpenSubtitles.org and download subtitles

for English, German, and a third language of your choosing. Open the files using a text editor $($e$.$g$.$ VS

Code$)$ and familiarise yourself with the format. Your corpus will include sentences from a famous scene

that starts at $00$:$17$:$48($first English sentence is : $$Quiet$!$ There are ways of telling whether she is a witch.$),$

and ends at $00$:$20$:$31($last English sentence is: $...$knight of the Round Table.$).$ Your goal is to clean up the

data, match subtitles in different languages and put the lines together, transforming them into the following

format:

line $1$ in English

line $1$ in German

line $1$ in chosen language

line $2$ in English

line $2$ in German

line $2$ in chosen language

$...$

You will see that this manual process is not feasible for greater amounts of data, and you will learn how to

automate a process like this later on in the course.

Save the created corpus as grail$\_$corpus.txt and submit the file together with the assignment.