Files
old-svevijesti/pyth/scrapingsingle.py
2024-01-06 08:17:05 +01:00

270 lines
9.7 KiB
Python

from bs4 import BeautifulSoup
import requests
from urllib.parse import urljoin
from openai import OpenAI , APIError
import os
from langchain.embeddings import OpenAIEmbeddings
from vectData import (insert_data ,is_similar_data ,get_similar, get_specific_data,get_all_links,cleansing ,modify_similar_data)
import json
from dotenv import load_dotenv
import tiktoken
load_dotenv()
cleansing()
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
client = OpenAI()
embeddings = OpenAIEmbeddings()
dlinks = ['https://klix.ba', 'https://srpskainfo.com', 'https://bljesak.info']
headers = {'User-Agent': 'Mozilla/5.0 (Linux; Android 5.1.1; SM-G928X Build/LMY47X) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/47.0.2526.83 Mobile Safari/537.36'}
def num_tokens_from_string(string: str, model="gpt-3.5-turbo") -> int:
encoding = tiktoken.encoding_for_model(model)
return len(encoding.encode(string))
def slice_text_at_2k_tokens(text):
encoding_name = "gpt-3.5-turbo"
max_tokens = 2000
encoding = tiktoken.encoding_for_model(encoding_name)
tokens = encoding.encode(text)
if len(tokens) <= max_tokens:
return [text]
sliced_tokens = tokens[:max_tokens]
sliced_text = encoding.decode(sliced_tokens)
return sliced_text
def replace_with_spaces(text):
allowed_chars = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyzČčĆćDždžĐ𩹮ž0123456789 "
cleaned_text = ''.join(char if char in allowed_chars else ' ' for char in text)
return cleaned_text
def fix_links(links_set):
modified_links = set()
for link in links_set:
if "www" in link:
modified_link = link.replace("www.", "")
modified_links.add(modified_link)
else:
modified_links.add(link)
return modified_links
total_links = set()
collected_news = set()
def get_article_links(url, already_checked):
response = requests.get(url,headers)
if response.status_code == 200:
soup = BeautifulSoup(response.text, 'html.parser')
articles = soup.find_all('article')
link_store = []
for article in articles:
links = article.find_all('a', href=True)
for link in links:
link_value = urljoin(url, link['href'])
if link_value not in already_checked:
link_store.append(link_value)
already_checked.add(link_value)
return link_store
already_checked = set()
for dlink in dlinks:
temp_links = get_article_links(dlink, already_checked)
if temp_links:
total_links.update(temp_links)
final_links = {item for item in total_links if item}
db_links = set(get_all_links())
new_links = final_links - db_links
final_links = new_links
final_links = set(final_links)
final_links = fix_links(final_links)
if __name__ == '__main__':
for link in final_links:
response = requests.get(link,headers)
soup = BeautifulSoup(response.text, 'html.parser')
titles = soup.find_all(['h2', 'h1','h3'])
title_text = ' '.join([title.get_text(strip=True) for title in titles])
texts = soup.find_all(['p'])
text_text = ' '.join([text.get_text(strip=True) for text in texts])
text_text = text_text
title_text = title_text
title_text = replace_with_spaces(title_text)
print(f"Tokens usage: {num_tokens_from_string(text_text, 'gpt-3.5-turbo')}")
text_text = slice_text_at_2k_tokens(text_text)
text_text = replace_with_spaces(str(text_text))
try:
completion = client.chat.completions.create(
model="gpt-3.5-turbo",
messages=[
{"role": "system", "content": "Data analytic, Journalist and News reporter"},
{"role": "user", "content": rf"Extract relevant information from the following input: Title: {title_text}, Text: {text_text}. Remove any non-news element related to the current text and title, and provide the cleaned data as a JSON object with 'title' and 'content' fields."}
]
)
generated_text = completion.choices[0].message.content
generated_text = generated_text
response_data = json.loads(generated_text)
title = response_data["title"]
text = response_data["content"]
#print("*********************************")
#print(f"Title: {title}")
#print("---------------------------------")
#print(f"Content : {text}")
#print("*********************************")
vector = embeddings.embed_query(generated_text)
if not is_similar_data(title, text, link, vector, threshold=0.98):
similar_d = "NO"
insert_data(title, text, link, vector,similar_d)
except Exception as e:
print(f"Error in completion: {e}")
continue
def comb_similar():
print("Checking similar")
similar_article = get_similar()
grouped_data = {}
for sa in similar_article:
if similar_article:
first_t = get_specific_data(sa[0])
second_t = get_specific_data(sa[1])
link_f = first_t[0][2]
link_s = second_t[0][2]
f_text = first_t[0][1]
s_text = second_t[0][1]
f_title = first_t[0][0]
s_title = second_t[0][0]
if f_title in grouped_data:
grouped_data[f_title].append((f_text, link_f))
else:
grouped_data[f_title] = [(f_text, link_f)]
if s_title in grouped_data:
grouped_data[s_title].append((s_text, link_s))
else:
grouped_data[s_title] = [(s_text, link_s)]
for title, tuples in grouped_data.items():
if len(tuples) == 3:
text1, link1 = tuples[0]
text2, link2 = tuples[1]
text3, link3 = tuples[2]
t1check = num_tokens_from_string(text1)
t2check = num_tokens_from_string(text2)
t3check = num_tokens_from_string(text3)
slice_if_more = t1check,t2check,t3check
if slice_if_more < 2000:
combined_text = f"{text1}{text2}{text3}"
combined_text = slice_text_at_2k_tokens(combined_text)
user_message = rf"Here is text {combined_text}, combined from 3 sources, filter text, and make news content, return as JSON only with 'content' field"
if link1 != link2 and link1 != link3 and link2 != link3:
link = f"{link1} {link2} {link3}"
else:
link = link1
else:
user_message = rf"Here are 3 texts {text1} {text2} and {text3}, combine the following texts into a cohesive news remove any non-news related to both texts and provide the cleaned data as a JSON only with 'content' field."
if link1 != link2 and link1 != link3 and link2 != link3:
link = f"{link1} {link2} {link3}"
else:
link = link1
else:
ftcheck = num_tokens_from_string(f_text)
stcheck = num_tokens_from_string(s_text)
fscomb = ftcheck + stcheck
if fscomb <2000:
combined_text = f"{f_text}{s_text}"
user_message = rf"Here is text {combined_text}, combined from 2 sources, filter text, and make news content, return as JSON only with 'content' field"
if link_f != link_s:
link = f"{link_f} {link_s}"
else:
link = link_f
else:
user_message = rf"Here are 2 texts {f_text} and {s_text}, combine the following texts into a cohesive news remove any non-news related to both texts and provide the cleaned data as a JSON only with 'content' field."
if link_f != link_s:
link = f"{link_f} {link_s}"
else:
link = link_f
try:
completion = client.chat.completions.create(
model="gpt-3.5-turbo",
messages=[
{"role": "system", "content": "Data analytic, Journalist and News reporter"},
{"role": "user", "content": user_message}
]
)
generated_text = completion.choices[0].message.content
if similar_article:
if f_title == s_title:
print(f_title)
modify_similar_data(first_t,"SOURCE")
similar_article.remove(sa)
print("Modified")
else:
print(f"First: {f_title}")
print(f"Second: {s_title}")
modify_similar_data(first_t,"SOURCE")
modify_similar_data(second_t,"SOURCE")
similar_article.remove(sa)
print("Modified")
else:
print("Similar list is empty")
response_data = json.loads(generated_text)
title = f_title
text = response_data["content"]
vector = embeddings.embed_query(generated_text)
if not is_similar_data(title, text, link, vector, threshold=0.98):
similar_d = "NO"
insert_data(title, text, link, vector, similar_d)
except Exception as e:
print(f"Error in completion: {e}")
continue