A Yoda-speaking challenge, this is

Question

The challenge

Well, I think it is quite clear, isn't it? You have to make a function or a program that takes a string as an argument and outputs the corrispondent Yoda-speaking.

This is code-golf, so least number of bytes wins.

The input

The input could be any string without linefeed. You can translate a string if it is composed like this:

Subject + Verb + Something else.

Where Subject is a personal pronoun (I, you, he, she, it, we, they). You don't have to recognize that the second word is a verb.

If the first word isn't a pronoun, the input is valid but the output will be Too difficult, this sentence is.

You have to win this code challenge -> Translatable
Luke has to win this code challenge -> Untranslatable

Input can end with a letter, a . or a !, not with a ?.

Furthermore, strings can contain Non-Ascii, Parenthesis, commas, colons ...

The output

In case of a translatable sentence, the output is the same sentence, with the subject and the verb at the end of the sentence.

You have to win this code challenge -> To win this code challenge, you have.

Comma, period and lower-case on the pronoun are mandatory. (Except, of course, if the pronoun is I).

As stated before, if it is untranslatable you have to output the string Too difficult, this sentence is.

Examples

You have to win this code challenge -> To win this code challenge, you have.
He won this code challenge -> This code challenge, he won.  (I know this is not yoda-speaking, but I don’t want to overcomplicate this challenge)
I ate an ice cream earlier! -> An ice cream earlier, I ate.
I liked the Star Wars original trilogy more than the prequel’s one. -> The Star Wars original trilogy more than the prequel’s one, I liked.
I find your lack of faith disturbing -> Your lack of faith disturbing, I find.
I think we are done with the examples -> We are done with examples, I think.
He is your son, Vader -> Your son, Vader, he is.
I think they’ll add new features -> They’ll add new features, I think.
I made a fantastic code challenge (I hope) -> A fantastic code challenge (I hope), I made.
I love constants like π -> Constants like π, I love.

I’ll be the next President, I swear! -> Too difficult, this sentence is.
This challenge is great! -> Too difficult, this sentence is.
Is this challenge great? -> Too difficult, this sentence is.
Luke is not ready for this task -> Too difficult, this sentence is.
Somebody loves constants like π -> Too difficult, this sentence is.
[Empty string] -> Too difficult, this sentence is.

Answer 1

Retina, 162 137 145 134 131 129

Multiple punctuation is now handled correctly. Test case: You will not win this challenge...! -> Not win this challenge..., you will.

Try it Online!

[.!]$

T`A-Z`a-z`^\w\w
G`^(I|we|you|s?he|it|they) \S+ \S
(\S+ \S+) (.*)
$2, $1.
T`a-z`A-Z`^.
^$
Too difficult, this sentence is.

Description:

If there is punctuation in the end, delete it.

[.!]$

^empty line

Transform the pronoun to lower case, except it is I.

T`A-Z`a-z`^\w\w

Filter out any line that does not match <Pronoun><space><word><space><word>...

G`^(I|we|you|s?he|it|they) \S+ \S

Split into pronoun + verb, remainder. Then rearrange.

(\S+ \S+) (.*)
$2, $1.

Transform the first character to upper case.

T`a-z`A-Z`^.

If the result is empty, it didn't pass the filter above. Print an error message.

^$
Too difficult, this sentence is.

Answer 2

ES6, 212

This can probably a bit further golfed down:

i=>(r=/(you|s?he|it?|we|they)( \S+) ([^\.!?]+)/i.exec(i))?(c=(a,b)=>a[`to${['Low','Upp'][b]}erCase`]())(r[3].slice(0,1),1)+r[3].slice(1)+", "+c(r[1],+(/i/i.test(r[1])))+r[2]+".":"Too difficult, this sentence is."

JSFiddle (requires up-to-date browser like Chrome) or run it via node.js

Ungolfed:

i=>{
    r=/(you|s?he|it?|we|they)( \S+) ([^\.!?]+)/i.exec(i); //Main RegExp
    c=(a,b)=>a[`to${['Low','Upp'][b]}erCase`](); //shortcut for toUpperCase/toLowerCase
    if (r)
        return c(r[3].slice(0,1),1)+r[3].slice(1) + ", " //the "something else", properly formated
             + c(r[1], +(/i/i.test(r[1]))) //The subject lowercased except for "i" or "I"
             + r[2] + "."; //The End
    else //no match, no sentence to translate
        return "Too difficult, this sentence is.";
}

Answer 3

JavaScript (ES6), 164 bytes

s=>([,p,,i,r]=s.match`^(((I)|You|He|She|It|We|They) \\S+) (.*?)[.!]?$`)?r[0].toUpperCase()+r.slice(1)+`, ${i?p:p.toLowerCase()}.`:"Too difficult, this sentence is."

Explanation

It's almost painful how many bytes capitalising the first letter of a string takes in JavaScript...

s=>

  // Match and get specific parts of the input string
  ([,p,,i,r]=s.match`^(((I)|You|He|She|It|We|They) \\S+) (.*?)[.!]?$`)?

    r[0].toUpperCase()+r.slice(1)     // capitalise first letter
      +`, ${i?p:p.toLowerCase()}.`    // lower-case the pronoun (unless it is I)
  :"Too difficult, this sentence is."

Test

Test does not use destructuring assignment to make it more cross-browser compatible.

var solution = s=>(p=s.match`^(((I)|You|He|She|It|We|They) \\S+) (.*?)[.!]?$`)?p[4][0].toUpperCase()+p[4].slice(1)+`, ${p[3]?p[1]:p[1].toLowerCase()}.`:"Too difficult, this sentence is."

<input type="text" id="input" value="I love constants like π...!" />
<button onclick="result.textContent=solution(input.value)">Go</button>
<pre id="result"></pre>

Answer 4

Python, 261 bytes

import re
def a(b):
 if b[-1:]in".!":
  b=b[:-1]
 if re.match('(I|YOU|S?HE|IT|WE|THEY) \w+ \S+',b.upper()):
  b=(b+',').split()
  if b[0]!="I":
   b[0]=b[0].lower()
  b=" ".join(b[2:]+b[:2])
  return b[0].upper()+b[1:]
 return "Too difficult, this sentence is."

Turns out Python doesn't mind things like b[-1]in".!" Next I'll play with the match object more, if no-one beats me to it :)

Answer 5

Python, 218 217 204 bytes

Not sure if this can be further golfed down.

def f(s):t=s.split();u=' '.join(t[2:]).rstrip('!.');return['Too difficult, this sentence is.','%s, %s %s.'%(u[0].upper()+u[1:],['I',t[0].lower()][t[0]!='I'],t[1])][t[0]in'I We You He She It They'.split()]

Ungolfed:

def f(s):
    t = s.split()
    u = ' '.join(t[2:]).rstrip('!.')
    return [
        'Too difficult, this sentence is.',
        '%s, %s %s.' % (u[0].upper() + u[1:],
                        ['I', t[0].lower()][t[0] != 'I'],
                        t[1])
    ][t[0] in 'I We You He She It They'.split()]

Answer 6

GNU sed, 129 bytes

I'm including +1 byte for the -r flag.

#!/bin/sed -rf

/^I /b
s/^(You|[HW]e|She|It|They) /\L&/
t
s/.*/Too difficult, this sentence is./p
d
:
s/[.!]$//
s/^([^ ]+ [^ ]+) (.*)/\u\2, \1./

Explanation

If we match a leading I we jump to the label:

/^I /b

If we match one of the other pronouns, we downcase it, then jump to the label:

s/^(You|[HW]e|She|It|They) /\L&/
t

Otherwise we print the failure message and move to the next input line:

s/.*/Too difficult, this sentence is./p
d

If we jumped to the label, we remove any final punctuation:

:
s/[.!]$//

and swap the first two words with the rest of the sentence, capitalising the new first word and adding the required punctuation as we do so:

s/^([^ ]+ [^ ]+) (.*)/\u\2, \1./