def train_emoji_vectors():
with open("data/emojis/emoji_df.csv",
mode="r",
encoding="utf-8",
newline="") as file:
lines = reader(file)
codes = []
for emoji in lines:
if ":" not in emoji[1] and "," not in emoji[1]:
clean = re.sub(r"[^a-zA-Z0-9\s-]+", "", emoji[1])
lower = clean.lower()
emoji = lower.split()
emoji = '-'.join(emoji)
codes.append(emoji)
getemojis.get_emojis(codes)
emoji_vectors = {}
for code in codes:
vector = get_vector(code)
if isinstance(vector, np.ndarray):
emoji_vectors[code] = get_vector(code)
data = [f"{len(emoji_vectors)} 300"]
for emoji in emoji_vectors:
numbers = " ".join(
str(number) for number in np.nditer(emoji_vectors[emoji]))
line = f"{emoji} {numbers}"
data.append(line)
text = "\n".join(data)
with open("data/emojis/emoji_vector_model.txt", "w") as file:
file.write(text)
def post(self):
settings = request.get_json()
emoji = settings['emoji']
filename = os.path.dirname(f"data/emojis/{emoji}.png")
if os.path.exists(filename):
return send_file(f"data/emojis/{emoji}.png",
mimetype="image/png")
else:
getemojis.get_emojis(emoji)
return send_file(f"data/emojis/{emoji}.png",
mimetype="image/png")
def analyze_users(num_of_tweets=2000, force_new_tweets=False, emojis=False):
users = []
with open("data/handles.json", "r") as file:
people = json.load(file)
for name in people:
person = ChiUser(name)
tweets = None
# Check for successful load of user data
if not person.load_user_data():
force_new_tweets = True
elif len(person.tweets) <= 0:
force_new_tweets = True
# Check if we are forcing the system to get new tweets
if force_new_tweets:
if num_of_tweets > 0:
print(f"Grabbing tweets for {person.username}")
tweets = twitterhandler.get_tweets(person.username,
num_of_tweets)
# Do the analysis
print(f"Analyzing tweets for {person.username}")
person.find_words(tweets)
# user.print_words()
if isinstance(person, ChiUser):
print(f"Working statistical analysis for {person.username}")
chisquaredmodel.add_user(person)
chisquaredmodel.calculate_user(person)
elif isinstance(person, BayesUser):
print(f"Working statistical analysis for {person.username}")
bayesmodel.add_user(person)
bayesmodel.calculate_user(person)
else:
person.add_interests(person.words)
# user.print_interests()
if emojis:
print(f"Getting emojis for {person.username}")
get_emojis(person)
users.append(person)
Metrics.analyze_users(users)
from getemojis import get_emojis
import re
import json
import codecs
with open("./.header.lua", "r") as file:
header = file.read()
with open("./.source.lua", "r") as file:
source = file.read()
encoded = "{" + ",".join(
['["{}"]="{}"'.format(key, val)
for key, val in get_emojis().items()]) + "}"
middle = "local codes = " + encoded
with codecs.open("emoji.lua", "w+", "utf-8", errors="surrogatepass") as file:
file.write("\n\n".join([header, middle, source]))