改进步骤
- 数据增强：使用GPT模型生成更多的训练数据。
- 使用更高级的模型：使用BERT或其他预训练的语言模型进行文本分类。
- 经验条和经验值显示：在网页端显示当前的经验值，并添加一个经验条。
数据增强和训练数据集
我们可以通过OpenAI的GPT模型来生成更多的训练数据。这里假设我们已经有一定数量的基础训练数据，并使用GPT模型生成更多样化的数据。
initial_data = {
'工作': [
'完成工作任务',
'解决问题',
'编写代码',
'参加会议'
],
'生活': [
'锻炼身体',
'学习新知识',
'看电影',
'休息'
],
'家务': [
'打扫卫生',
'洗衣服',
'做饭',
'整理房间'
]
}使用GPT生成更多数据
我们将使用OpenAI的API生成更多数据。请确保您已设置OpenAI API密钥：
import openai
openai.api_key = 'your_openai_api_key'
def generate_data(prompt, max_tokens=50):
response = openai.Completion.create(
engine="text-davinci-003",
prompt=prompt,
max_tokens=max_tokens,
n=5,
stop=None,
temperature=0.7
)
return [choice['text'].strip() for choice in response['choices']]
def augment_data(data):
augmented_data = {}
for category, examples in data.items():
augmented_data[category] = examples.copy()
for example in examples:
prompt = f"给出一些类似于'{example}'的{category}活动。"
new_examples = generate_data(prompt)
augmented_data[category].extend(new_examples)
return augmented_data
# 增强数据
augmented_data = augment_data(initial_data)使用BERT进行文本分类
我们将使用transformers库中的BERT模型进行分类。以下是完整的app.py代码：
import nltk
from nltk.tokenize import word_tokenize
from nltk.corpus import stopwords
from nltk.stem import WordNetLemmatizer
from transformers import BertTokenizer, BertForSequenceClassification
from transformers import Trainer, TrainingArguments
from sklearn.model_selection import train_test_split
from torch.utils.data import Dataset, DataLoader
from flask import Flask, render_template, request, jsonify
import torch
import os
import pandas as pd
from datetime import datetime
import openai
# 初始化nltk
nltk.download('punkt')
nltk.download('stopwords')
nltk.download('wordnet')
# 初始化分类器和评分模型
classifier = None
tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
# 初始化分类和评分标准
categories = ['工作', '生活', '家务']
score_criteria = {
'工作': {
'完成任务': 10,
'解决问题': 5,
'编程': 8,
'参加会议': 4,
'其他': 1
},
'生活': {
'锻炼': 5,
'学习': 3,
'休息': 2,
'看电影': 2,
'其他': 1
},
'家务': {
'打扫卫生': 5,
'洗衣服': 3,
'做饭': 4,
'整理房间': 3,
'其他': 1
}
}
# 初始化基础分数和等级
base_score = 0
level = '初级'
# 初始化日志文件和评分记录
daily_log = []
scores = []
# 初始化词形还原器
lemmatizer = WordNetLemmatizer()
# 设置OpenAI API密钥
openai.api_key = 'your_openai_api_key'
# 检查并创建示例数据文件
def create_example_files():
example_data = {
'工作': [
'完成工作任务\n',
'解决问题\n',
'编写代码\n',
'参加会议\n',
'其他工作相关活动\n'
],
'生活': [
'锻炼身体\n',
'学习新知识\n',
'看电影\n',
'休息\n',
'其他生活相关活动\n'
],
'家务': [
'打扫卫生\n',
'洗衣服\n',
'做饭\n',
'整理房间\n',
'其他家务相关活动\n'
]
}
for category, data in example_data.items():
filename = f'{category}_log.txt'
if not os.path.exists(filename):
with open(filename, 'w', encoding='utf-8') as file:
file.writelines(data)
# 使用GPT生成更多数据
def generate_data(prompt, max_tokens=50):
response = openai.Completion.create(
engine="text-davinci-003",
prompt=prompt,
max_tokens=max_tokens,
n=5,
stop=None,
temperature=0.7
)
return [choice['text'].strip() for choice in response['choices']]
# 增强数据
def augment_data(data):
augmented_data = {}
for category, examples in data.items():
augmented_data[category] = examples.copy()
for example in examples:
prompt = f"给出一些类似于'{example}'的{category}活动。"
new_examples = generate_data(prompt)
augmented_data[category].extend(new_examples)
return augmented_data
# 读取和增强数据
def load_and_augment_data():
create_example_files()
initial_data = {
'工作': [],
'生活': [],
'家务': []
}
for category in categories:
with open(f'{category}_log.txt', 'r', encoding='utf-8') as file:
initial_data[category] = file.read().splitlines()
return augment_data(initial_data)
# 自定义数据集类
class ActivityDataset(Dataset):
def __init__(self, texts, labels, tokenizer, max_len):
self.texts = texts
self.labels = labels
self.tokenizer = tokenizer
self.max_len = max_len
def __len__(self):
return len(self.texts)
def __getitem__(self, item):
text = self.texts[item]
label = self.labels[item]
encoding = self.tokenizer.encode_plus(
text,
add_special_tokens=True,
max_length=self.max_len,
return_token_type_ids=False,
padding='max_length',
return_attention_mask=True,
return_tensors='pt',
truncation=True
)
return {
'text': text,
'input_ids': encoding['input_ids'].flatten(),
'attention_mask': encoding['attention_mask'].flatten(),
'labels': torch.tensor(label, dtype=torch.long)
}
# 训练分类器
def train_classifier():
global classifier
data = load_and_augment_data()
texts = []
labels = []
for category, examples in data.items():
texts.extend(examples)
labels.extend([categories.index(category)] * len(examples))
train_texts, val_texts, train_labels, val_labels = train_test_split(
texts, labels, test_size=0.1, random_state=42
)
train_dataset = ActivityDataset(train_texts, train_labels, tokenizer, max_len=128)
val_dataset = ActivityDataset(val_texts, val_labels, tokenizer, max_len=128)
training_args = TrainingArguments(
output_dir='./results',
num_train_epochs=3,
per_device_train_batch_size=16,
per_device_eval_batch_size=16,
warmup_steps=500,
weight_decay=0.01,
logging_dir='./logs',
logging_steps=10,
evaluation_strategy='epoch'
)
model = BertForSequenceClassification.from_pretrained('bert-base-uncased', num_labels=len(categories))
trainer = Trainer(
model=model,
args=training_args,
train_dataset=train_dataset,
eval_dataset=val_dataset,
)
trainer.train()
classifier = model
# 预处理文本数据
def preprocess_text(data):
processed_data = []
for text in data:
# 分词
tokens = word_tokenize(text.lower())
# 去除停用词和标点符号
stop_words = set(stopwords.words('english'))
filtered_tokens = [token for token in tokens if token.isalnum() and token not in stop_words]
# 词形还原
lemmatized_tokens = [lemmatizer.lemmatize(token) for token in filtered_tokens]
# 重新组合为文本
processed_text = ' '.join(lemmatized_tokens)
processed_data.append(processed_text)
return processed_data
# 智能分类工作日常
def classify_activity(activity):
inputs = tokenizer(activity, return_tensors='pt', truncation=True, padding=True, max_length=128)
outputs = classifier(**inputs)
predictions = torch.argmax(outputs.logits, dim=1)
return categories[predictions.item()]
# 根据分类和评分标准计算活动得分
def score_activity(activity, category):
score = 0
for word, points in score_criteria[category].items():
if word in activity:
score += points
return score
# 记录每天的工作日常到Excel
def log_daily_activity(activity):
global base_score, level
# 智能分类
category = classify_activity(activity)
# 计算得分
score = score_activity(activity, category)
# 更新基础分数和等级
base_score += score
if base_score >= 100:
level = '高级'
elif base_score >= 50:
level = '中级'
# 记录日志
daily_log.append((activity, category, score))
# 更新评分记录
scores.append(score)
# 获取当前时间
timestamp = datetime.now().strftime('%Y-%m-%d %H:%M:%S')
# 记录到Excel
log_to_excel(timestamp, activity, category, score)
# 记录数据到Excel文件
def log_to_excel(timestamp, activity, category, score):
filename = 'activity_log.xlsx'
if os.path.exists(filename):
df = pd.read_excel(filename)
else:
df = pd.DataFrame(columns=['Timestamp', 'Activity', 'Category', 'Score'])
new_entry = pd.DataFrame([[timestamp, activity, category, score]], columns=['Timestamp', 'Activity', 'Category', 'Score'])
df = pd.concat([df, new_entry], ignore_index=True)
df.to_excel(filename, index=False)
# 创建Flask应用
app = Flask(__name__)
@app.route('/')
def index():
return render_template('index.html', base_score=base_score, level=level)
@app.route('/log_activity', methods=['POST'])
def log_activity():
activity = request.form['activity']
log_daily_activity(activity)
return jsonify({'base_score': base_score, 'level': level})
@app.route('/simulate')
def simulate():
global base_score, level
# 模拟每天的活动记录
activities = [
'完成工作任务',
'锻炼身体',
'解决问题',
'学习新知识',
'编写代码',
'看电影',
'休息',
'参加会议',
'打扫卫生',
'洗衣服',
'做饭',
'整理房间',
'其他活动'
]
for activity in activities:
log_daily_activity(activity)
# 返回当前分数和等级
return jsonify({'base_score': base_score, 'level': level})
if __name__ == '__main__':
# 训练分类器和评分模型
train_classifier()
# 启动Flask应用
app.run(debug=True)HTML代码（templates/index.html）
我们将创建一个表单，让用户输入他们的活动，并通过Ajax发送到服务器，同时添加经验值和经验条的显示：
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>升级动画</title>
<style>
body {
text-align: center;
font-family: Arial, sans-serif;
}
.character {
width: 200px;
height: 200px;
margin: 50px auto;
border: 2px solid #000;
border-radius: 10px;
background-color: #f0f0f0;
position: relative;
}
.character img {
width: 100%;
height: 100%;
display: none;
}
.character img.active {
display: block;
}
.level {
font-size: 1.5em;
margin-top: 20px;
}
.upgrade {
font-size: 1.2em;
color: green;
}
.form-container {
margin: 20px auto;
}
.progress-bar {
width: 80%;
margin: 20px auto;
background-color: #e0e0e0;
border-radius: 10px;
overflow: hidden;
}
.progress-bar-inner {
height: 20px;
background-color: #76c7c0;
width: 0;
}
</style>
</head>
<body>
<div class="character">
<img src="/static/character_level1.png" id="level1" class="active">
<img src="/static/character_level2.png" id="level2">
<img src="/static/character_level3.png" id="level3">
</div>
<div class="level" id="level">当前等级: 初级</div>
<div class="upgrade" id="upgrade"></div>
<div class="form-container">
<form id="activity-form">
<input type="text" id="activity" name="activity" placeholder="输入今天的活动" required>
<button type="submit">记录活动</button>
</form>
</div>
<div class="progress-bar">
<div class="progress-bar-inner" id="progress-bar-inner"></div>
</div>
<div id="experience">当前经验值: 0</div>
<script>
let baseScore = {{ base_score }};
let level = "{{ level }}";
function updateLevel(newLevel) {
const levelImages = {
'初级': document.getElementById('level1'),
'中级': document.getElementById('level2'),
'高级': document.getElementById('level3')
};
for (let level in levelImages) {
if (level === newLevel) {
levelImages[level].classList.add('active');
document.getElementById('level').textContent = `当前等级: ${level}`;
document.getElementById('upgrade').textContent = `升级到${level}!`;
} else {
levelImages[level].classList.remove('active');
}
}
}
function updateExperience(newScore) {
baseScore = newScore;
document.getElementById('experience').textContent = `当前经验值: ${baseScore}`;
let progressBar = document.getElementById('progress-bar-inner');
progressBar.style.width = `${Math.min((baseScore % 50) * 2, 100)}%`;
}
function logActivity(event) {
event.preventDefault();
const activity = document.getElementById('activity').value;
fetch('/log_activity', {
method: 'POST',
headers: {
'Content-Type': 'application/x-www-form-urlencoded'
},
body: `activity=${encodeURIComponent(activity)}`
})
.then(response => response.json())
.then(data => {
updateLevel(data.level);
updateExperience(data.base_score);
document.getElementById('activity').value = '';
});
}
document.getElementById('activity-form').addEventListener('submit', logActivity);
</script>
</body>
</html>
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
