** CHALLENGE: Simple almost human speakbot **

Question

The goal of the challenge is to use the 500 first lines of http://www.gutenberg.org/cache/epub/1952/pg1952.txt to build a "Markov table" ( containing how many times any word was followed by any of its potential "following words"). Dot "." need to be considered as a word). This table can be quite big so you need to think about the right way to do it.
Once you have this table you can generate almost human sentences : start with a dot, print randomly a following word & iterate.
Make the best looking code !

Answer 1

New version that can run on Sololearn Code Playground (doesn't pull text from Project Gutenberg directly):

https://code.sololearn.com/c80p4MH1NdIm/?ref=app

Answer 2

nice, this did reminds my hinter editor https://code.sololearn.com/WoUjEYWDOir8/?ref=app

I'll research about for give a try :)

Answer 3

I will need my word extractor :) https://code.sololearn.com/WqtvZsVN8AUH/?ref=app

Answer 4

https://code.sololearn.com/cMvAGN9YGqj6/?ref=app gave it a limit so it wont get stuck in a loop

Answer 5

mine, I also give a limit because words by itself don't get stop unless it has more words so one stablish when occurrence 1.

I like this bot :)

note: I know it isn't the best looking code 😁

https://code.sololearn.com/WaFXqRzaA6aw/?ref=app

Answer 6

Wow ... like the challenge!!! I'll give it a try!

BTW:
BE AWARE of using the link (gutenberg.org) for URL opening if you're not a US-citizen! They block your IP and might charge you for it (due to copyrights).

(at least that's what they wrote on their home-page)

I'm not from US therefore I just copied the text into my source-code!

Answer 7

Andre wins this one. Very concise code and not so many Rubystes here ;-)

Answer 8

ya but in here...in python
sololearn dnt support urllib.request...
max limit exceeded....

Answer 9

eh, for example: "hello world! hello world."
if I start with hello, the probability of the follower be "world" is 1, and now with world the probability of the follower be "!" is 0.5 and be "." 0.5. .
so we have to do something as it for every word ?

Answer 10

My response in Ruby:

https://code.sololearn.com/cFZ3aQKjcJrS/?ref=app

As a warning, it DOES NOT work in Sololearn's Code Playground - the script will need to be copy-and-pasted and tried on a computer running Ruby. It takes the first 500 lines directly from the Project Gutenberg website, and Sololearn denies the permission to access the site from Code Playground.

It is a base 1 Markov model. The output text is sometimes accidentally almost poetic. Some examples:

Great makes by as bushes convolutions -- but undertaking I.

No my stimulating to baby a long, and greater before fancy. It can kept you tell and shapes dear much figure, or weir.

Old-paper wall-somersault just not I.

Answer 11

@VcC re: "Can you put the first line directly in a variable so that we can test it ?"

Sorry, not sure what you mean - the first line requiring the 'open-uri' module?

I can make a second version that has the text as a variable instead of accessing it directly from the page if that's what you mean - give me a minute, it will just (unfortunately) no longer be only 26 lines or so of actual code :)

Answer 12

Here is my version - php.
https://code.sololearn.com/wmE1uVRK3Eqq/?ref=app

Answer 13

راسل

Answer 14

ؤاسلثا

Answer 15

يرسبسزبةزقظل

Answer 16

ؤرسبى

Answer 17

how many times a word was followed by any of its potential "followimg words"...

please explain i didnt get...

Answer 18

Your markov chain tells you for each word in the original text what the possible followers are and the number for each. Using this you can pick a follower for any word. If you start with . (dot), pick a follower, the a follower's follower and so on your program will generate an almost human looking sentence ! You can feed this programm with any base ( could be the Us president's tweets or a programmers joke database).
Good challenge to practice URL opening and file analysis....

Answer 19

Indeed urllib does not work... Ok i'll accept answers puting the 500 lines in a variable ! For non python you can use url related functions ( google them)

Answer 20

just copied that string😂😂😂😂