CSC111 Homework 7

From dftwiki3
Revision as of 20:20, 10 March 2010 by Thiebaut (talk | contribs) (Memory)
Jump to: navigation, search

This assignment is currently under construction...



Eliza

Please, meet Eliza

Please read this introduction to Eliza in Wikipedia. The goal of this assignment is to write our own version of Eliza.

For a weird implementation of Eliza, check this place out...

Setup

  1. Copy the following program in a file called eliza1.py
  2. Copy the next program in a file called logging111.py. Logging111.py is not the main program, but a simple library created for this class that will allow us to capture the dialog between the user and Eliza in a text file.
  3. Play with the program
  4. Every so often, try answering "no" or "never" and see how the program reacts.
  5. Also, from time to time, use "I", "you", "me", or "your" in your answer, as in "I like you" for example. See how the program responds. We call this a reflection.

Reflection

  • Start your Eliza program and tell it: I like you
  • Observe that it responds: you like me?
This is a reflection!
  • Unfortunately, your current version of Eliza is not very clever.
  • Try: you do remind me of my mother
  • Observe its output.
  • See the problem?

Part #1: Improved Reflections

Improve the way your program performs reflections.

When doing a reflection, your program should transform the string "you" at the beginnning of a sentence in "I". However, if the word "you" is not at the beginning of a sentence, your program will translate it into "me".

The effect of this is that if you type:


you do not understand


Your Eliza program will respond with


I do not understand?


But if you type:


I told you so


your Eliza program will respond with


you told me so?


Modify your program so that it will perform reflections correctly (at least for I, you, and me).

Submit it as follows:

  submit hw7 eliza1.py

Part 2: Picking up on keywords

Modify your program so that if you type the words mother or father anywhere in your input, the program will respond with a random sentence whose topic should include the word family.

For example, if you type My mother always calls me at the wrong time, your Eliza program could respond with something like Is that typical in your family? or Is that a family trait?

Make sure your program recognizes different versions of the keywords, such as (but not limited to) mother, Mother, or MOTHER.

Remember that we have seen several examples of generation of random strings, in particular randomName2.py and the Solution programs for Homework #3.

Part 3: Let's be polite

Make your program scold the user if he/she does not start the first sentence that contains the word hello or Hello. Feel free to find an appropriate sentence for this faux pas.

Part 4: Ending the conversation

Your program should stop not only if the user enters bye, but it should also recognize goodbye, ciao.

Part 5: Aggressive user

If the user enters a sentence of the form I xxxx you, we can imagine that the word xxxx is a verb, and that the user is being aggressive. In this case Eliza should respond with a sentence of the form Why do you xxxx me?


Part 6: Dealing with exclamation marks and question marks

If the user enters a sentence of the form I lost you!, then the last word, "you" will not be reflected in "me" because the string.split() function will have kept the exclamation mark attached to the word "you." The last word is thus "you!" and will not be reflected by the current program. Figure out a way for your program to be able to reflect the last words of a sentence, even when they are terminated by a "!", "?", or "."

Be efficient in your coding. The fewer changes you need to make to your program to make it behave that way, the better.

Part 7: Memory

Whenever the user enters a sentence of the form "I xxxx you", make your program remember it so that whenever it has to output a canned response, it will sometimes output "Is that why you xxxx me?". As an example, imagine that the user has already typed in her conversation with Eliza "I dislike you" and "I trust you", whenever the user enters a sentence that is not reflected and not negative, the program will sometimes, but not always, respond with "Is that why you dislike me?", or "Is that why you trust me?". This should really enhance the illusion that the program is intelligent and remembers the conversation.

You do not need to submit a separate version of your program if you implement the extra-credit questions. Put the enhanced version in hw7.py and submit it as mentioned above.

Recommendations

Use functions! What we have built so far is a very linear program, but it should use functions. For example, returning a canned sentence from a list of canned sentences can easily be done by a function. Only PG13 programs will be accepted! Be imaginative in creating your dialogs and responses, and make your programs enjoyable and fun to read and play with!