drussellmrichie

# Natural Language Toolkit: code_baseline_tagger
# functions from http://www.nltk.org/book/ch05.html

from nltk.corpus import brown
import nltk

def performance(cfd, wordlist):
    lt = dict((word, cfd[word].max()) for word in wordlist)
    baseline_tagger = nltk.UnigramTagger(model=lt, backoff=nltk.DefaultTagger('NN'))
    return baseline_tagger.evaluate(brown.tagged_sents(categories='news'))

drussellmrichie

"""
With 93/120/9 FFN with sigmoid hidden layer, and fixing the activations on 
the output layer (make them sum to one, put on interval 0-1), and train for 
5 epochs, only get score of 1.28
"""

import pandas as pd
import os
from pybrain.tools.shortcuts import buildNetwork
from pybrain.supervised.trainers import BackpropTrainer

drussellmrichie

"""
This is a script I wrote to help me select office hours for my spring 2016 TAship of 
COGS 2201: Foundations in Cognitive Science at the University of Connecticut.
"""

# import a few packages that we'll need to structure the doodle data, change the working directory,
# and compute the pairs of dates/times.

import pandas as pd
import os, itertools

drussellmrichie

def cohortModel(word, EnglishWords):
    soFar = ''                                 # when we start, we haven't heard anything yet, so we'll represent that as an empty string
    candidates = set(EnglishWords)             # before we've heard anything, all the words we know are possible
    for letter in word:                        # then start listening to the word letter by letter
        soFar += letter                        # add the newly heard letter to the portion of the word heard so far
        for word in set(candidates):           # now look through the candidate words
            if not word.startswith(soFar):     # if the word seen so far is NOT consistent with a word we know
                candidates.remove(word)        # remove the word from the candidates
        print("These are the possible words when we've heard {} so far:\n{}".format(str(soFar),str(candidates)))
    return candidates

drussellmrichie

"""Usage:
  gradeEmailer.py [--name=NAME] [--password=PASSWORD] [--dry-run] <sender-email> <grades-path>

Options:
  --name=NAME           The name of the grader.
                        Used in the signature and From field of the email.
                        If not given, <sender-email> will be used.
  --password=PASSWORD   The password to the sender's email account.
                        If not given, it will be prompted for.
  --dry-run             To test, send emails to yourself instead of to the students.

drussellmrichie

import spacy
import numpy as np

nlp = spacy.load('en', disable=['ner'])


def tree_height(root):
    """
    Find the maximum depth (height) of the dependency parse of a spacy sentence by starting with its root
    Code adapted from https://stackoverflow.com/questions/35920826/how-to-find-height-for-non-binary-tree

drussellmrichie

words = [{1: {'text': 'elephant', 'position': (320, 180)}},
        {2: {'text': 'wug', 'position': (220, 140)}},
        {3: {'text': 'dog', 'position': (320, 180)}}]

import pyactr as actr

environment = actr.Environment(focus_position=(0,0))
lex_decision = actr.ACTRModel(
    environment=environment,
    automatic_visual_search=False,

drussellmrichie

from transformers import pipeline
from numpy import argmax
import shap
from transformers import RobertaConfig, RobertaModel
from transformers import RobertaTokenizer
import torch
import numpy as np
import scipy as sp
# import datasets
import pandas as pd