Upload 2 files

extract_news/2_extract_news.py

@@ -0,0 +1,99 @@
+# -*- coding: utf-8 -*-
+
+"""
+News Source Extractor:
+
+This script is designed to extract the content of news articles from various French media sources.
+The URLs of these articles are retrieved from the `base_news` table, where articles marked with 
+a `step` value of '0' are pending extraction.
+
+To install the necessary packages:
+pip install aiohttp mysql-connector-python
+
+Once extracted, the content of each article is saved locally for further processing. This separation 
+of content fetching and processing is intentional to optimize resource management.
+
+The script operates in batches, processing a defined number of entries (`NB_BY_STEP`) at a time.
+After extraction, the `step` value of the processed articles is updated to '1' to indicate completion.
+
+Author     : Guillaume Eckendoerffer
+Date       : 29-09-23
+Repository : https://github.com/Eckendoerffer/TorchTrainerFlow/
+             https://huggingface.co/datasets/eckendoerffer/news_fr
+"""
+
+import asyncio
+import aiohttp
+import time
+import mysql.connector
+import os
+
+# Database configuration
+db_config = {
+    "host": "[host]",
+    "user": "[user]",
+    "password": "[passwd]",
+    "database": "[database]"
+}
+
+NB_BY_STEP = 20
+path = os.getcwd()  
+
+def mysqli_return_number(conn, query):
+    cursor = conn.cursor()
+    cursor.execute(query)
+    result = cursor.fetchone()
+    cursor.close()
+    return result[0] if result else 0
+
+async def fetch_and_save(url, id_source):
+    try:
+        async with aiohttp.ClientSession() as session:
+            async with session.get(url) as response:
+                byte_content = await response.read()
+                
+                try:
+                    text_content = byte_content.decode('utf-8')
+                except UnicodeDecodeError:
+                    text_content = byte_content.decode('ISO-8859-1')
+
+                with open(f"{path}/sources/html_news/{id_source}.txt", "w", encoding="utf-8") as file:
+                    file.write(text_content)
+                    
+    except aiohttp.client_exceptions.TooManyRedirects:
+        print(f"Too many redirects for URL: {url}")
+
+
+async def main():
+    conn = mysql.connector.connect(**db_config)
+
+    while True:
+        time_start = time.time()
+
+        cursor = conn.cursor()
+        cursor.execute(f"SELECT `id`, `url` FROM `base_news` WHERE `step`='0' ORDER BY RAND() LIMIT {NB_BY_STEP}")
+        rows = cursor.fetchall()
+        cursor.close()
+
+        if not rows:
+            break
+
+        tasks = []
+        for row in rows:
+            id_source, url = row
+            cursor = conn.cursor()
+            cursor.execute(f"UPDATE `base_news` SET `step`='1' WHERE `id`='{id_source}' LIMIT 1")
+            print(f'{id_source}) {url}')
+            cursor.close()
+            tasks.append(fetch_and_save(url.strip(), id_source))
+
+        await asyncio.gather(*tasks)
+
+        nb_base = mysqli_return_number(conn, "SELECT COUNT(`id`) FROM `base_news` WHERE `step`='0'")
+        time_elapsed = time.time() - time_start
+        time_per_item = time_elapsed / NB_BY_STEP
+        print(f"Remaining: {nb_base} - Time: {time_per_item:.3f}s/item")
+
+    conn.close()
+
+asyncio.run(main())

extract_news/4_extract_news_url.py

@@ -0,0 +1,131 @@
+# -*- coding: utf-8 -*-
+
+"""
+Random Line Fetcher for Large Datasets
+
+Extracts and stores relevant links from local French online news articles.
+
+pip install beautifulsoup4 mysql-connector-python colorama
+
+Author     : Guillaume Eckendoerffer
+Date       : 28-09-23
+Repository : https://github.com/Eckendoerffer/TorchTrainerFlow/
+"""
+
+import os
+from bs4 import BeautifulSoup
+import mysql.connector
+import hashlib
+from colorama import Fore, init
+
+# Database configuration
+db_config = {
+    "host": "[host]",
+    "user": "[user]",
+    "password": "[passwd]",
+    "database": "[database]"
+}
+
+conn = mysql.connector.connect(**db_config)
+cursor = conn.cursor()
+query = "SELECT `key_media` FROM `base_news` WHERE `key_media` != ''"
+cursor.execute(query)
+keys = cursor.fetchall()
+formatted_keys = "|".join([key[0] for key in keys]) + "|"
+
+init(autoreset=True)
+
+def get_dom_path(url):
+    from urllib.parse import urlparse
+    parsed_url = urlparse(url)
+    return f"{parsed_url.scheme}://{parsed_url.netloc}"
+
+def get_html_content(file_path):
+    with open(file_path, 'r', encoding='utf8', errors='ignore') as file:
+        return file.read()
+
+def mysqli_return_number(conn, query, params=None):
+    cursor = conn.cursor()
+    cursor.execute(query)
+    result = cursor.fetchone()
+    cursor.close()
+    return result[0] if result else 0
+
+def mysqli_return_count(conn, query):
+    cursor = conn.cursor()
+    cursor.execute(query)
+    result = cursor.fetchone()
+    cursor.close()
+    return result[0] if result else 0
+
+def process_news_source():
+    global formatted_keys
+
+    cursor = conn.cursor()
+    query = ("SELECT `id`, `url`, `media` FROM `base_news` WHERE `link`='0' AND `step` > 0 AND `id` > 215000 AND `url` NOT LIKE 'https://avis-vin.%' AND `url` NOT LIKE 'https://www.elle.fr/%' AND `url` NOT LIKE 'www.lamontagne.fr/%' AND `url` NOT LIKE 'https://www.rtbf.be/%' AND `url` NOT LIKE 'https://www.tf1info.fr/%' AND `url` NOT LIKE 'https://www.futura-sciences.com/%' AND `url` NOT LIKE 'https://cdn-elle.ladmedia.fr/%' ORDER BY Rand() LIMIT 1")
+    cursor.execute(query)
+    row = cursor.fetchone()
+
+    if not row:
+        return 'No unprocessed news source found.'
+        
+    id_source, url_source, id_media = row
+    dom = get_dom_path(url_source)
+    cursor.execute(f"UPDATE `base_news` SET `link`='1' WHERE `id`='{id_source}' LIMIT 1")
+    conn.commit()
+
+    querys = "SELECT COUNT(`id`) FROM `base_news` WHERE `step`='0'"
+    nb_link = mysqli_return_count(conn, querys)
+
+    file_path = f"sources/html_news/{id_source}.txt"
+    if os.path.exists(file_path):
+        html_content = get_html_content(file_path)
+    else:
+        return  
+
+    print(f"{nb_link} {url_source} {id_media} ({len(html_content)})")
+
+    soup = BeautifulSoup(html_content, 'html.parser')
+    nb_add = 0
+    for link in soup.find_all('a'):
+        url = link.get('href')
+        if url is None:
+            continue
+        url = url.split("#")[0]
+        url = url.split("?")[0]
+
+        if not url:
+            continue
+        if not "//" in url:
+            url = f"{dom}/{url}" if url[0] != '/' else f"{dom}{url}"
+        elif "http" not in url:
+            url = 'https:' + url
+        if not url.startswith(("http://", "https://")) or url.count(' ') or url.count('%') or url.count('\''):
+            continue 
+
+        key = hashlib.md5(url.encode()).hexdigest()
+        nb_base_news = formatted_keys.count(f'{key}|')
+
+        if url.startswith(dom):
+            if nb_base_news:
+                #print(Fore.YELLOW + url)
+                continue
+            elif (
+                url.count("-") > 6 and
+                not any(substring in url for substring in ['replay', 'video', 'login', '/inloggen', '?redirect', '.jpg', '.png', 'mailto'])
+            ):
+                print(Fore.GREEN + url)
+                insert_query = f"INSERT INTO `base_news` (`id`, `key_media`, `media`, `url`, `step`) VALUES (NULL, '{key}', '{id_media}', '{url}', '0');"
+                cursor.execute(insert_query)
+                conn.commit()
+                formatted_keys = f'{formatted_keys}{key}|'
+                nb_add += 1
+        else:
+            #print(Fore.RED + url)
+            continue
+
+while True:
+    process_news_source()
+
+conn.close()
+