Files
marketscrape-web/scraper.py
2022-12-21 15:14:32 -08:00

293 lines
11 KiB
Python

# Database
import database
# Regular Expressions
import re
# Web Scraping
import requests
from bs4 import BeautifulSoup
# Math
import statistics
import numpy as np
# Currency Conversion
from currency_converter import CurrencyConverter
# Sentiment Analysis
#nltk.download()
#nltk.download('stopwords')
from nltk.corpus import stopwords
from nltk.sentiment import SentimentIntensityAnalyzer
from nltk.tokenize import RegexpTokenizer
from nltk.stem import WordNetLemmatizer
from difflib import SequenceMatcher
# Pattern Matching
import re
def sentiment_analysis(text):
# Create a SentimentIntensityAnalyzer object
sia = SentimentIntensityAnalyzer()
sentiment = sia.polarity_scores(text)
# Get the sentiment scores
neg, neu, pos, compound = sentiment["neg"], sentiment["neu"], sentiment["pos"], sentiment["compound"]
# Assign a rating based on the compound score
if compound > 0.0:
rating = 5 * max(pos, compound)
elif compound < 0.0:
rating = 5 * min(neg, compound)
else:
rating = 5 * neu
return abs(rating)
def clean_text(text):
# Remove punctuation
tokenizer = RegexpTokenizer('\w+|\$[\d\.]+|http\S+')
tokenized = tokenizer.tokenize(text)
# Lowercase all words
tokenized = [word.lower() for word in tokenized]
# Remove stopwords
stop_words = stopwords.words('english')
# Filter out any tokens not containing letters
filtered = [word for word in tokenized if word not in stop_words and word.isalpha()]
# Lemmatize all words
lemmatizer = WordNetLemmatizer()
lemmatized = [lemmatizer.lemmatize(word) for word in filtered]
return " ".join(lemmatized)
def price_difference_rating(initial, final):
# If the listing price is less than or equal to the median price found online, set the rating to 5
if initial <= final:
rating = 5.0
else:
# If the listing price is greater than the median price found online, calculate the difference
difference = min(initial, final) / max(initial, final)
rating = (difference / 20) * 100
return rating
def get_listing_title(soup):
# Get the title of the listing
title = soup.find("meta", {"name": "DC.title"})
title_content = title["content"]
return title_content
def get_listing_description(soup):
# Get the description of the listing
description = soup.find("meta", {"name": "DC.description"})
description_content = description["content"]
return clean_text(description_content)
def get_listing_price(soup):
# Get the price of the listing
spans = soup.find_all("span")
# Check if the listing is free
free = [span.text for span in spans if "free" in span.text.lower()]
if (free):
return free
# Find the span that contains the price of the listing and extract the price
price = [str(span.text) for span in spans if "$" in span.text][0]
return price
def create_soup(url, headers):
# Create a request object
response = requests.get(url, headers=headers)
# Create a BeautifulSoup object
soup = BeautifulSoup(response.text, 'html.parser')
return soup
def convert_currency(price, base_currency, target_currency):
# Convert the price to the target currency
c = CurrencyConverter()
price = c.convert(price, base_currency, target_currency)
return price
def clean_listing_title(title):
# Certain symbols are not allowed in the search query for Google Shopping, so they must be removed
title = re.sub(r"#", "%2", title)
title = re.sub(r"&", "%26", title)
return title
def get_product_price(soup):
# Get the price of the product
prices = soup.find_all("span", {"class": "HRLxBb"})
# Extract the price from the span
values = []
for price in prices:
values.append(price.text)
# Remove the dollar sign from the price
normalized = [re.sub("\$", "", price) for price in values]
# Convert the price to a float
normalized = [re.search(r"[0-9,.]*", price).group(0) for price in normalized]
# Remove the commas from the price
normalized = [float(price.replace(",", "")) for price in normalized]
# Remove statistical outliers as to not skew the median price
outlierless = reject_outliers(np.array(normalized))
return outlierless
def get_product_description(soup):
# Get the description of the product
description = soup.find_all("div", {"class": "rgHvZc"})
return description
def clean_title_description(title):
# Remove punctuation
cleaned = re.sub(r"[^A-Za-z0-9\s]+", " ", title)
# Remove extra spaces
cleaned = re.sub(r"\s+", " ", cleaned)
return cleaned
def listing_product_similarity(soup, title, similarity_threshold):
# Get the median price of the product
normalized = get_product_price(soup)
# Get the product description
description = get_product_description(soup)
price_description = {}
# Iterate through the product descriptions
for key, value in zip(description, normalized):
google_shopping_title = clean_title_description(key.text.lower())
listing_title = clean_title_description(title.lower())
# Get the similarity between the listing title and the product description on Google Shopping
price_description[key.text] = [value, SequenceMatcher(None, google_shopping_title, listing_title).ratio()]
prices = []
# Iterate through the product descriptions and their similarity scores
for key, value in price_description.items():
# If the similarity score is greater than the similarity threshold, add the price to the list of prices
if value[1] >= similarity_threshold:
prices.append(value[0])
return prices
def find_viable_product(title, ramp_down):
title = clean_listing_title(title)
headers = {
"User-Agent":
"Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/70.0.3538.102 Safari/537.36 Edge/18.19582"
}
url = "https://www.google.com/search?q=" + title + "&sa=X&biw=1920&bih=927&tbm=shop&sxsrf=ALiCzsbtwkWiDOQEcm_9X1UBlEG1iaqXtg%3A1663739640147&ei=-KYqY6CsCLez0PEP0Ias2AI&ved=0ahUKEwigiP-RmaX6AhW3GTQIHVADCysQ4dUDCAU&uact=5&oq=REPLACE&gs_lcp=Cgtwcm9kdWN0cy1jYxADMgUIABCABDIFCAAQgAQyBQgAEIAEMgsIABCABBCxAxCDATIECAAQAzIFCAAQgAQyBQgAEIAEMgUIABCABDIFCAAQgAQyBQgAEIAEOgsIABAeEA8QsAMQGDoNCAAQHhAPELADEAUQGDoGCAAQChADSgQIQRgBUM4MWO4TYJoVaAFwAHgAgAFDiAGNA5IBATeYAQCgAQHIAQPAAQE&sclient=products-cc"
soup = create_soup(url, headers)
# Set the similarity threshold to a initial value, and decrease it when no products are found
similarity_threshold = 0.45
try:
prices = listing_product_similarity(soup, title, similarity_threshold)
# The length of the list of prices should be greater than 0 if there are viable products
assert len(prices) > 0
except AssertionError:
print("Error: no viable products found, now searching for more general products...")
while len(prices) == 0:
# If no viable products are found, the search is further generalized by 5%, until a reasonable number of products are found
ramp_down += 0.05
prices = listing_product_similarity(soup, title, similarity_threshold - ramp_down)
# Get the median price of the viable products
median = statistics.median_grouped(prices)
return min(prices), max(prices), median
def valid_url(url):
if re.search(r"^https://www.facebook.com/", url):
return True
else:
return False
# The larger the value of m is, the less outliers are removed
# Source: https://stackoverflow.com/questions/62802061/python-find-outliers-inside-a-list
def reject_outliers(data, m=1.5):
distribution = np.abs(data - np.median(data))
m_deviation = np.median(distribution)
standard = distribution / (m_deviation if m_deviation else 1.)
return data[standard < m].tolist()
def print_results(title, initial_price, sentiment_rating, price_rating, average_rating, median, lower_bound, upper_bound):
print("\n● Listing:")
print(" ○ Product: {}".format(title))
print(" ○ Price: ${:,.2f}".format(initial_price))
print("● Similar products:")
print(" ○ Range: ${:,.2f} - ${:,.2f}".format(lower_bound, upper_bound))
print(" ○ Median: ${:,.2f}".format(median))
print("● Ratings:")
print(" ○ Description: {:,.2f}/5.00".format(sentiment_rating))
print(" ○ Price: {:,.2f}/5.00".format(price_rating))
print(" ○ Overall: {:,.2f}/5.00".format(average_rating))
def main():
# Initialize the database
database.initialize()
# Get the URL of the Facebook Marketplace listing
url = input("Enter URL: ")
# Check if the URL is valid
if valid_url(url):
pass
else:
print("Error: URL is not from Facebook Marketplace.")
exit(1)
# Shorten the URL listing to the title of the listing
shortened_url = re.search(r".*[0-9]", url).group(0)
# Use the shortened URL and convert it to mobile, to get the price of the listing
mobile_url = shortened_url.replace("www", "m")
# Find the ID of the product
market_id = (re.search(r"\/item\/([0-9]*)", url)).group(1)
records = database.retrieve(market_id)
if records:
title = records[1]
initial_price = records[2]
sentiment_rating = records[3]
price_rating = records[4]
average_rating = records[5]
median = records[6]
lower_bound = records[7]
upper_bound = records[8]
elif not records:
# Get the sentiment rating of the listing
sentiment_rating = sentiment_analysis(get_listing_description(create_soup(url, headers=None)))
# Get the title of the listing
title = get_listing_title(create_soup(url, headers=None))
# Get the minimum, maximum, and median prices of the viable products found on Google Shopping
list_price = get_listing_price(create_soup(mobile_url, headers=None))
if list_price[0] == "FREE":
print("This product is free!")
return
initial_price = int(re.sub("[\$,]", "", list_price))
lower_bound, upper_bound, median = find_viable_product(title, ramp_down=0.0)
# Calculate the price difference between the listing and the median price of the viable products, and generate ratings
price_rating = price_difference_rating(initial_price, median)
average_rating = statistics.mean([sentiment_rating, price_rating])
# Add the listing to the database
database.insert(market_id, title, initial_price, sentiment_rating, price_rating, average_rating, median, lower_bound, upper_bound)
print_results(title, initial_price, sentiment_rating, price_rating, average_rating, median, lower_bound, upper_bound)
if __name__ == "__main__":
main()