README_old.md

chatbot

##Potential Resources

• A detailed description of creating an IRC chatbot
• documentation for Natural Language Toolkit
• Parsing English with 500 lines of Python
  • Redshift Github This is the program described by the above link
• Any tutorials for developing chatbots?
• Therapist.py ("a cheezy little Eliza knock-off by Joe Strout")
• Howie the Chatterbot
• Computational Linguistics & Psycholinguistics Research Center

##Similar projects

• thebot 0.1.1
• Videogrep (not that similar, but uses CLiPS packages to search for sentence structure)

##Possible packages:

• [Textblob] (https://textblob.readthedocs.org/en/dev/) is a python package that provides an api for the NLTK.

##Sections of Project:

#Mission Statement:

We are trying to create a simple chatbot with a basic sense of self. It will have the following capabilities:

• Interact on live website

• Generates original grammatically correct sentences

• Can perform basic semantic analysis on user input

• Can appropriately answer basic questions regardless of form:

  • What is your name?, Who are you?, What are you called?

• Generates original grammatically correct sentences

  • Knowledge of parts of speech
  • Understanding of what makes a valid sentence
    • Syntactically valid sentence
    • Semantically valid sentence
  • Be able to generate stylistically appropriate sentence based on training materialshould I refer to you?
  • Where do you live?, Where is your home?, What city do you reside in?, Where are you?, Where are you from?
  • How are you?, How are you feeling?, How's it going? How are things?, What's up?

• Respond to statements (non-questions)

##Workflow (Stages):

###Stage 1

• Generates original grammatically correct sentences
  • Knowledge of parts of speech
  • Understanding of what makes a valid sentence
    • Syntactically valid sentence
    • Semantically valid sentence
  • Be able to generate stylistically appropriate sentence based on training materials

##Milestones for Tuesday 6/17:

• Research and identify POS tagger (Nathan)
• Draft plan for web interface (Josh)
• Outline next steps and research potential packages (Corinne)

###Stage 2 (Ephemeral Stage)

• Can perform basic semantic analysis on user input
  • Identify question versus statement. Hard code vs soft code?

###Stage 3 (Nebulous Stage)

• Can appropriately answer basic questions regardless of form:

###Stage 4 (Profit)

---

###User Stories

• 1. As an anonymous user, when I enter a single sentence into the text field the bot generates a single sentence response.
• 2. As an anonymous user, when I select a filter in the checklist the filter is applied to the next response.
• 3. As an anonymous user, when I click the "show me" button I see a log of the steps to generate the output.
• 4. As an anonymous user, when I visit the page I see a "Tweet" button to share the response.
• 5. An an anonymous user, when I visit the website I see the chatbot page.
• 6. As an anonymous user, when I visit the website I see a "Contact Us" link that takes me to the project email address and a link to the GitHub page.
• 7. As an anonymous user, when I have a noteworthy interaction with the bot I can click on a button to add that interaction to a "Best Of" page.

here is the bot!

UPENN TagSet