reveal.js

# Pause for Breath

---

CS 65 // 2021-04-08

# Weekend Review Sessions

## Assignment 7
- Can be completed individually or in a group of your choosing

# Questions
## ...about anything?

# Case Study
## Procedural Text Generation

## Procedural Text Generation
- As a case study, we were implementing to randomly generate English sentences
- 
 The idea we were employing was using the **two previous words** to determine which word should come next
- 
 The probability of choosing a word is based on what words commonly follow the previous words

## Procedural Text Generation
- Consider the text from *Eric, the Half a Bee*:

> Half a bee, philosophically,<br>
> Must, ipso facto, half not be.<br>
> But half the bee has got to be<br>
> Vis a vis, its entity. D’you see?
> 
> But can a bee be said to be<br>
> Or not to be an entire bee<br>
> When half the bee is not a bee<br>
> Due to some ancient injury?

</div>

## Procedural Text Generation
- We decided to use a **dictionary** `d` like this:
    + The keys are pairs of words: `(w1,w2)`
    + 
 The value  `d[w1,w2]` is a list of all the words that commonly follow those two words

## Parsing Words

```py
def parse_words(filename):
    f = open(filename)
    words = []
    for line in f:
        words.extend(line.split())
    
    f.close()
    return words
```

## Markov Analysis

```py
def markov(words, order=2):
    suffixes = {}
    prefix = tuple(words[:order])
    for next_word in words[order:]:
        if prefix not in suffixes:
            suffixes[prefix] = [next_word]
        else:
            suffixes[prefix].append(next_word)
            
        prefix = prefix[1:] + (next_word,)
            
    return suffixes
```

## Random Text

```py
def random_text(suffixes, n=100):
    text = ""
    prefix = random.choice(list(suffixes.keys()))
    for i in range(n):
        next_word = random.choice(suffixes[prefix])
        text = text + " " + next_word
        prefix = prefix[1:] + (next_word,)
    
    return text
```

# Assignment 7 Time