scrape/scrape_large.py

import os
import re
import requests
from tqdm import tqdm
from datasets import Dataset
from bs4 import BeautifulSoup

# try to find info 
def from_wiki_script(author_soup):
    try:
        text = str(author_soup.find_all("script")[0])
        try:
            birth = re.findall(r"Født i \d{4}",text)[0]
            birth = re.findall(r"\d{4}",birth)[0]
        except IndexError:
            birth = None

        try:
            death = re.findall(r"Døde i \d{4}",text)[0]
            death = re.findall(r"\d{4}",death)[0]
        except IndexError:
            death = None
    except KeyError:
        birth = None
        death = None
    return birth, death

def from_infobox(author_soup):
    #right infobox
    try:
        boxes = author_soup.find_all("table")
        try:
            boxes = [i for i in boxes if "biography" in i["class"]][0]
            try:
                #vals first 4 digit nr after string
                death = re.findall(r"(?<=Død).*?[^\d]*(\d{4})",str(boxes))[0]
            except IndexError:
                death = None
            try:
                birth = re.findall(r"(?<=Født).*?[^\d]*(\d{4})",str(boxes))[0]
            except IndexError:
                birth = None
        except IndexError:
            birth = None
            death = None
    except KeyError:
        birth = None
        death = None

    return birth, death

#last resort, find first two 4 digit nums in first textbox
def from_wiki_text(author_soup):
    try:
        text = list(author_soup.find_all("p"))[0].get_text()
        try:
            birth = re.findall(r"\d{4}",text)[0]
        except IndexError:
            birth = None
        try:
            death = re.findall(r"\d{4}",text)[1]
        except IndexError:
            death = None
    except KeyError:
        birth = None
        death = None
    
    return birth, death

def none_to_q(val:str) -> str: 
    """ If value is None replaces it with ?"""
    if val is None:
        val = "?"
    else:
        pass
    return val

def find_wiki_birth_death(author_soup):
    birth, death = from_wiki_script(author_soup)
    if birth is None and death is None:
        birth, death = from_infobox(author_soup)
    else:
        pass
    if birth is None and death is None:
        birth, death = from_wiki_text(author_soup)
    else:
        pass

    birth = none_to_q(birth)
    death = none_to_q(death)

    return birth, death

#filter for people
def is_a_person(tag):
    return( (tag.has_attr('href')) and 
            (tag.has_attr('title')) and 
            (len(tag.attrs) == 2) and 
            ("index" not in tag.get("href")) and
            (":") not in  tag.get("href"))
#filter for wikimedia commons
def is_a_person_commons(tag):
    return( (tag.has_attr('href')) and 
            (tag.has_attr('title')) and 
            (len(tag.attrs) == 2) and 
            ("index" not in tag.get("href")) and
            (("Writers" not in tag.get("title")) and 
             ("ategories" not in tag.get("title")) and 
             ("Denmark"  not in tag.get("title"))) and
            ("Category" in tag.get("title"))
        )
#filter for author subcategories
def is_a_subcategory(tag):
    return( (tag.has_attr('href')) and 
            (tag.has_attr('title')) and
            ("Dansksprogede" in tag.get("title")) 
    )

def flatten(twodlist :list[list,list]) -> list:
    """ flatten a list by 1 dimension"""
    onedlist = [x for xs in twodlist for x in xs]
    return onedlist

def extract_authors(people,
                    authors:list[dict[str,str]],
                    name_list:list[str]
                    ) -> list[list[dict[str,str]], list[str]]:    
    for i in people:
        author_name = i.get("title")
        author_link = i.get("href")
        if author_name not in name_list:
            #find their death
            author_page = requests.get(f"https://da.wikipedia.org{author_link}")
            author_soup = BeautifulSoup(author_page.content, 'html.parser')
                
            birth, death = find_wiki_birth_death(author_soup)
                
            author_row={
                "link": f"https://da.wikipedia.org{author_link}",
                "name":author_name,
                "born":birth,
                "died":death,
                "name_yr":f"{author_name} ({birth}-{death})"
                        }
            authors.append(author_row)
            name_list.append(author_name)
        else:
            pass
    return authors, name_list

def extract_authors_commons(people,
                    authors:list[dict[str,str]],
                    name_list:list[str]
                    ) -> list[list[dict[str,str]], list[str]]:
    for i in people:
        author_name = i.get_text()
        author_link = i.get("href")
        if author_name not in name_list:
            #find their death
            author_page = requests.get(f"https://commons.wikimedia.org{author_link}")
            author_soup = BeautifulSoup(author_page.content, 'html.parser')
                
            boxes = author_soup.find_all("table")
            try:
                box = [i for i in boxes if "Date of birth" in str(i)][0]
                try:
                    #vals first 4 digit nr after string
                    death = re.findall(r"(?<=Date of death).*?[^\d]*(\d{4})",str(box))[0]
                except IndexError:
                    death = None
                try:
                    birth = re.findall(r"(?<=Date of birth).*?[^\d]*(\d{4})",str(box))[0]
                except IndexError:
                    birth = None
            except IndexError:
                birth = None
                death = None

            birth = none_to_q(birth)
            death = none_to_q(death)
                
            author_row={
                "link": f"https://commons.wikimedia.org{author_link}",
                "name":author_name,
                "born":birth,
                "died":death,
                "name_yr":f"{author_name} ({birth}-{death})"
                        }
            authors.append(author_row)
            name_list.append(author_name)
        else:
            pass
        
    return authors, name_list

def is_next_page(tag):
    return (tag.get_text() == "næste side")

def main():
    authors = []
    name_list = []
    #people from main page
    print(f"https://da.wikipedia.org/wiki/Kategori:Dansksprogede_forfattere")
    page = requests.get(f"https://da.wikipedia.org/wiki/Kategori:Dansksprogede_forfattere")
    soup = BeautifulSoup(page.content, 'html.parser')
    #1 get all peeps from page
    people = list(soup.find_all("ul" and "li" and "a" and is_a_person))
    authors, name_list = extract_authors(
            people,
            authors,
            name_list
        )
    ##### go into subcategories
    sub_c = soup.find_all("ul" and "li" and "a" and is_a_subcategory)
    for i in sub_c:
        
        if "Danmark" not in i.get("title"):
            #if author not from denmark (more pages, less people)
            new_link = f"https://da.wikipedia.org/{i.get("href")}"
            page = requests.get(new_link)
            soup = BeautifulSoup(page.content, 'html.parser')
            people = list(soup.find_all("ul" and "li" and "a" and is_a_person))
            authors, name_list = extract_authors(
                                                people,
                                                authors,
                                                name_list
                                                )
            print(f"DONE: {i.get("title")}")

        elif "Danmark" in i.get("title"):
            #if author from denmark (less pages, more people)
            print("Processing Authors from Denmark (alphabetic order)...\n")
            #alphabet_list = "A B".split()
            alphabet_list = "A B C D E F G H I J K L M N O P Q R S T U V W X Y Z Æ Ø Å".split()
            for letter in tqdm(alphabet_list):
                page = requests.get(f"https://da.wikipedia.org/w/index.php?title=Kategori:Dansksprogede_forfattere_fra_Danmark&from={letter}")
                soup = BeautifulSoup(page.content, 'html.parser')
                people = list(soup.find_all("ul" and "li" and "a" and is_a_person))
                authors, name_list = extract_authors(
                    people,
                    authors,
                    name_list
                )
    ###other webpages
    print(f"Processing https://commons.wikimedia.org/wiki/Category:Writers_from_Denmark_by_name")
    #abc_list = "A B".split()
    abc_list = "A B C D E F G H I J K L M N O P Q R S T U V W X Y Z".split()
    for abc in tqdm(abc_list):
        page = requests.get(f"https://commons.wikimedia.org/w/index.php?title=Category:Writers_from_Denmark_by_name&from={abc}")
        soup = BeautifulSoup(page.content, 'html.parser')
        people = list(soup.find_all("ul" and "li" and "a" and is_a_person_commons))
        authors, name_list = extract_authors_commons(
                        people,
                        authors,
                        name_list
                    )
    
    print(f"Processing https://da.wikipedia.org/wiki/Kategori:Personer_i_Dansk_Biografisk_Leksikon")
    
    #get names from page, next page, repeat
    for abc in tqdm(abc_list):
        page = requests.get(f"https://commons.wikimedia.org/w/index.php?title=Category:Writers_from_Denmark_by_name&from={abc}")
        soup = BeautifulSoup(page.content, 'html.parser')
        people = list(soup.find_all("ul" and "li" and "a" and is_a_person_commons))
        authors, name_list = extract_authors_commons(
                        people,
                        authors,
                        name_list
                    )
    #another webpage
    p_counter = 0
    print(f"Processing https://da.wikipedia.org/wiki/Kategori:Personer_i_Dansk_Biografisk_Leksikon")

    page = requests.get(f"https://da.wikipedia.org/wiki/Kategori:Personer_i_Dansk_Biografisk_Leksikon")
    soup = BeautifulSoup(page.content, 'html.parser')
    #get names from page, next page, repeat until no more next page
    while len(soup.find_all("a" and is_next_page)) > 0:

        people = list(soup.find_all("ul" and "li" and "a" and is_a_person))
        authors, name_list = extract_authors(
                        people,
                        authors,
                        name_list
                    )
        p_counter += 1        
        new_page = soup.find_all("a" and is_next_page)[0]["href"]
        new_link = f"https://da.wikipedia.org/{new_page}"
        page = requests.get(new_link)
        soup = BeautifulSoup(page.content, 'html.parser')
        print(f"Scraped page {p_counter}/~30...")
    else:
        #last page
        print("Scraping last page...")
        people = list(soup.find_all("ul" and "li" and "a" and is_a_person))
        authors, name_list = extract_authors(
                        people,
                        authors,
                        name_list
                    )
        
    ds = Dataset.from_list(authors)
    ds.to_parquet(os.path.join(".","da_people_large.parquet"))


if __name__ == "__main__":
    main()