Моя диаграмма показывает совершенно неверные данные, и я хочу знать, где в моем коде я могу изменить то, что хочу получить, чтобы оно отображало правильные данные.
Отключено ли извлечение данных ? Или отображение данных отключено?
Если в коде нет ничего плохого, просветите меня. Я в отчаянии...
import os
import csv
import pandas as pd
from bs4 import BeautifulSoup
import tarfile
file_path = 'kungalv_slutpriser.tar.gz'
with tarfile.open(file_path, 'r:gz') as tar:
tar.extractall('extracted_html')
def extract_house_info(html_content):
house_info = {}
try:
soup = BeautifulSoup(html_content, 'html.parser')
# Extract data using BeautifulSoup (adjust selectors based on actual HTML structure)
house_info['date_of_sale'] = soup.find(string='Såld').find_next().text.strip() if soup.find(string='Såld') else None
house_info['address'] = soup.find(string='Adress').find_next().text.strip() if soup.find(string='Adress') else None
house_info['location'] = soup.find(string='Ort').find_next().text.strip() if soup.find(string='Ort') else None
area_text = soup.find(string='Boarea').find_next().text.strip() if soup.find(string='Boarea') else None
if area_text and area_text.split()[0].isdigit():
house_info['boarea'] = int(area_text.split()[0])
else:
house_info['boarea'] = None
biarea_tag = soup.find(string='Biarea')
if biarea_tag:
biarea_text = biarea_tag.find_next().text.strip()
if biarea_text.split()[0].isdigit():
house_info['biarea'] = int(biarea_text.split()[0])
house_info['totalarea'] = house_info['boarea'] + house_info['biarea']
else:
house_info['biarea'] = None
house_info['totalarea'] = house_info['boarea']
house_info['rooms'] = soup.find(string='Rum').find_next().text.strip() if soup.find(string='Rum') else None
house_info['plot_area'] = soup.find(string='Tomt').find_next().text.strip() if soup.find(string='Tomt') else None
closing_price_tag = soup.find(string='Pris')
if closing_price_tag:
house_info['closing_price'] = closing_price_tag.find_next().text.strip()
else:
house_info['closing_price'] = None
except Exception as e:
print(f"Error extracting data: {e}")
return house_info
# List to hold all extracted house data
all_houses = []
# Process each HTML file
for file_name in html_files:
file_path = os.path.join(html_folder, file_name)
try:
with open(file_path, 'r', encoding='utf-8') as file:
html_content = file.read()
house_info = extract_house_info(html_content)
all_houses.append(house_info)
except Exception as e:
print(f"Error processing file {file_name}: {e}")
# Create a DataFrame from the list of dictionaries
df = pd.DataFrame(all_houses)
# Save DataFrame to CSV file
df.to_csv('house_prices.csv', index=False)
print("Data has been written to 'house_prices.csv'.")
Подробнее здесь: https://stackoverflow.com/questions/790 ... o-my-chart
