Found something useful to do with my enormous collection of movie subtitle files(.srt). Its a bot (like ELIZA or the emacs psychiatrist, not too bright) which is a huge movie fan. Ask him anything, he’ll reply with the closest known movie dialogue I’m attaching a sample conversation here (Try to Identify the movies and also, a well known TV Series from the sample).

With help from python inbuilt difflib, the resulting code was shorter than expected. To start I had to filter out the .srt files from timestamps and dialogue numbers and combine everything to a huge file and with a little approximate string matching .. voila!

you: hello

MBot: its me only

you: then why do you say that?

MBot: because l wanted to hurt you.

you: oh really?

MBot: yes, but… this is a copy.

you: a copy of what?

MBot: they’re watching you, neo.

you: who is watching?

MBot: come on, go!

you: go where?

MBot: to the machine city.


MBot: this lily or that lily?

you: you dont make much sense now

MBot: “much” sense?

you: cant you learn anything?

MBot: l don’t know.

you: what do you know?

MBot: you married a lesbian!

you: I did not marry a lesbian!

MBot: to make you look good!

you: you are an idiot!

MBot: yep, that’s me.

you: finally we agreed on something

MBot: you don’t have to become yakuza

you: who said anything about them?

MBot: anyone hear from phoebe yet?

you: nope

MBot: excuse me.

you: are you leaving?

MBot: i have a small job to do

you: what job?

MBot: you wanted him to invite you

you: invite me where?

MBot: . who, pops?

you: yeah.. invite?


you: ok.. general questions

MBot: shoot.

you: the first rule of fight club

MBot: i’m not supposed to talk about it.

you: what does indiana jones wear?

MBot: so does joy behar.

you: what are you wearing?

MBot: nothing.

you: why bats sir?

MBot: Bats frighten me. It’s time my enemies share my dread.

you: how did you get these scars?

MBot: i just saw you in salem.

you: tell a joke

MBot: just shut up. i’ll try not to

I lost the code used to generate this. But it was surprisingly ~ 30 lines!