Day 18.  Continuing to write epics (wannabe.py)

This shows an interesting data structure!  Not a simple int, e.g.


handoutDay18.txt - so they can see up close the texts and the
dictionaries built for them.

Version 1 review:

[wannabe_starter2.py]

Look at your handouts.  Sample text and the dictionary that is built.

Talk through build_dict:

   we haven't seen any words yet, so last_word is ""
   the dictionary is empty, to start

   get the words in the line.  

   use split, to create a list of the words in the line

   for each of those words:

      if the last word we saw is already in the dictionary,
      then update its entry:  w, our current word, follows
      the last word; so, add w to the end of that list.

      if it isn't, create a new entry, initializing to the
      list containing only w.

         this means: we just saw last_word for the first time
         in the text.  So far w is the only word we have
         seen following last word.


Talk through write_epic:

   current_word is "" - we don't have any words yet.

   loop:  in each iteration, we add one word to our epic

   just read the code; it is straightforward.

   starting things off?  See:  there is an entry in 
     the dictionary with key "".  What do our epics ALWAYS
     start with?  'Happy'.

Ok, now look at the code and example, and trace through.
Write down what is printed.

   # do it on the board, I guess, since they are on paper

Now, the second version:

  Back in lecture mode.

  Look at the second example on the handout.

  prefix = 2

  we are going to remember what followed every PAIR of words
  in the text.

  use a moving window

  happy day
        day and
            and Happy
                Happy day
                      day but
                          but Sad
                                  etc.

  so, what changes in our code?

  the keys in the dictionary are tuples.

  initial context:  ('','')  prefix == 2
         the last 2 words are nothing, nothing
                    ('','','') prefix == 3
         the last 3 words are nothing, nothing, nothing
                    ('','','','') prefix == 4
         the last 4 words are nothing, nothing, nothing, nothing
  
  the code?

  context = ('',)  you need the "," so python knows it is a tuple.
      just something you need to remember/check in code you saw before

  for i in range(context_length-1):

  suppose context_length = 3.  Then, this is:

  for i in [0,1]:  so, we'll do this twice.  we already have the first
  one.

  context = context + ('',)

  + words for tuples too.

  show in the shell.

  OK?

  We read through the words just the same way.

  instead of last_word, we have context - compare the
  two pieces of code - you have them on your handout.

  the interesting thing:  updating the context.

  suppose:

    context = (but, Sad)

    word = minute

    we just added new entry word_dict[(but, Sad)] = [minute] 

    so, we are HERE in the text - show on the board with arrows

    we want to slide our window over 1 to the right

    want context to be (Sad, minute)

    we want to kick one word off the last, and slot in word to the
    right

    draw a picture

    context[1:] + (word,)

    suppose we are running it with prefix == 3

    context = (day, but, Sad)

    word = minute

    context[1:] is (but, Sad)
  
    word is minute

    so, context[1:] + (minute,) is (but, Sad, minute)