Download all posts from a Facebook page, in txt or csv

scrapefb.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-

#Written by Luca Tringali
# The code is described here:
# https://www.codice-sorgente.it/2018/08/facebook-scraping-scaricare-tutti-i-post-delle-pagine-facebook/
#Released under GNU GPL3

#USAGE: python3 ./scrapefb.py https://facebookpageurl/ ./ CSV
#The second argument is the output folder, the third one (in caps) is the format TXT or CSV

import urllib.request
import urllib.parse
import re
import html
import sys
import os
import json
import datetime
import time
from socket import timeout

#useragent = 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_3) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/35.0.1916.47 Safari/537.36'
useragent = 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/86.0.4240.111 Safari/537.36'

def geturl(thisurl):
        global useragent
        if thisurl == '':
            return ''
        req = urllib.request.Request(
            thisurl,
            data=None,
            headers={
                'User-Agent': useragent
            }
        )

        thishtml = ""
        try:
            f = urllib.request.urlopen(req,timeout=300)
            ft = f.read() #we should stop if this is taking too long
        except:
            ft = ""
        try:
            encoding = f.info().get_content_charset()
            if encoding == None:
                encoding = 'windows-1252'
            thishtml = ft.decode(encoding)
        except:
            try:
               thishtml = ft.decode('utf-8', 'backslashreplace')
            except:
               thishtml = str(ft)
        try:
            thishtml = html.unescape(thishtml)
        except:
            thishtml = ""
        return thishtml


def scrapefacebook(mypage, output = "./", ascsv = False):
    TOSELECT_FB = 'pages_reaction_units'
    startposts = "{\"__html\":"
    endposts = "]]"
    maxresults = 300
    towait = 10
    lstart = '/pages_reaction_units/more/?page_id='
    lending = '&cursor={"card_id":"videos","has_next_page":true}&surface=www_pages_home&unit_count='+str(maxresults)+'&referrer&dpr=1&__user=0&__a=1'
    allhtml = geturl(mypage)
    try:
        start = mypage.index("https://")
        end = mypage.index('/',start+8)
        fbdomain = mypage[start:end]
        indexes = [(m.start(0), m.end(0)) for m in re.finditer(TOSELECT_FB, allhtml[start+1:])]
        start = indexes[0][0]
        end = allhtml.index('"',start+1)
        thislink = allhtml[start:end]
        #estrapola page_id
        #getting page ID from page source HTML
        #https://it-it.facebook.com/pages_reaction_units/more/?page_id=286796408016028&cursor={"card_id":"videos","has_next_page":true}&surface=www_pages_home&unit_count=300&referrer&dpr=1&__user=0&__a=1
        start = thislink.index("page_id=")
        end = thislink.index('&',start+9)
        pageid = thislink[start+8:end]
        start = mypage.index("facebook.com")
        pagename = mypage[start+12:]
        pagename = re.sub(r'[^A-Za-z0-9]',"",pagename)
    except:
        fbdomain = ""
        pageid = ""
    fname = output + "fb_" + pagename + ".txt"
    if ascsv:
        fname = output + "fb_" + pagename + ".csv"
    alllinks = []
    linksfile = output + "fb_" + pagename + ".tmp"
    if os.path.isfile(linksfile):
        alllinks = [line.rstrip('\n') for line in open(linksfile, encoding='utf-8')]
    timelineiter = 0
    ripristino = False
    active = True
    while active:
        link = fbdomain + lstart + pageid + lending
        if timelineiter == 0 and len(alllinks)>0:
            link = alllinks[len(alllinks)-1]
            ripristino = True
        with open(linksfile, "a", encoding='utf-8') as lfile:
            lfile.write(link+'\n')
        print(link)
        newhtml = geturl(link)
        try:
            start = newhtml.index(startposts)
            end = newhtml.index(endposts)
            postshtml = newhtml[start:end]
            #eliminazione caratteri unicode surrogati
            #deleting unicode surrogates
            postshtml = postshtml.encode("utf-8").decode('unicode-escape')
            postshtml = re.sub(r'[\uD800-\uDFFF]',"",postshtml)
            #dividi per data-utime e tieni solo quello che sta tra <p> e </p>
            #splitting posts by their Unitx Time (data-utime) and keeping only text inside <p> </p> tags
            postsarray = re.split('data-utime', postshtml)
            timearray = []
            for i in range(len(postsarray)):
                try:
                    start = postsarray[i].index('"')
                    end = postsarray[i].index('"',start+2)
                    utime = postsarray[i][start:end]
                    utime = re.sub(r'[^0-9]',"",utime)
                    utimei = int(utime)
                except:
                    utimei = 0
                thistime = datetime.datetime.utcfromtimestamp(utimei).strftime('%Y-%m-%d %H:%M:%S')
                timearray.append(thistime)
                indexes = [(m.start(0), m.end(0)) for m in re.finditer('<p>(.*?)<\\\\/p>', postsarray[i])]
                thispost = ""
                for n in range(len(indexes)):
                    start = indexes[n][0]
                    end = indexes[n][1]
                    thispost = thispost + postsarray[i][start:end]
                #pulisco i tag non necessari
                #cleaning unnecessary tags
                postsarray[i] = re.sub(r'<.*?>',"",thispost)
                #tolgo gli slash non necessari
                #cleaning unnecessary slashes
                postsarray[i] = re.sub('\\\\/',"/",postsarray[i])
            print(postsarray)
            try:
                maxresults = 8
                start = newhtml.index('&cursor=')
                end = newhtml.index("&unit_count=", start+1)
                lending = newhtml[start:end]
                #eliminazione caratteri unicode surrogati
                #again, cleaning unicode surrogates
                lending = lending.encode("utf-8").decode('unicode-escape')
                lending = re.sub(r'[\uD800-\uDFFF]',"",lending)
                lending = urllib.parse.unquote(lending)
                lending = lending + '&unit_count='+str(maxresults)+'&dpr=1&__user=0&__a=1'
                #https://it-it.facebook.com/pages_reaction_units/more/?page_id=286796408016028&cursor={"timeline_cursor":"timeline_unit:1:00000000001528624041:04611686018427387904:09223372036854775793:04611686018427387904","timeline_section_cursor":{},"has_next_page":true}&surface=www_pages_home&unit_count=8&dpr=1&__user=0&__a=1
            except:
                active = False
        except:
            postsarray = []
            timearray = []
        #salvo il risultato in un file: se è il primo ciclo creo il file, altrimenti aggiungo
        #saving results in a file, creating a new one or appending to an existing one if needed
        if fname != "":
            postsfile = ""
            for i in range(len(postsarray)):
                if postsarray[i] != "":
                    if ascsv:
                        postsfile = postsfile + timearray[i] + "\t"
                    postsfile = postsfile + postsarray[i] + "\n"
            if timelineiter == 0 and ripristino==False:
                text_file = open(fname, "w", encoding='utf-8')
                text_file.write(postsfile)
                text_file.close()
            else:
                with open(fname, "a", encoding='utf-8') as myfile:
                    myfile.write(postsfile)
        timelineiter = timelineiter +1
        time.sleep(towait)


if __name__ == '__main__':
    START_PAGE = "https://it-it.facebook.com/chiesapastafarianaitaliana/"
    if len(sys.argv)>1:
        START_PAGE = sys.argv[1]
    if "facebook.com" in START_PAGE:
        output = "./"
        if len(sys.argv)>2:
            output = sys.argv[2]
        ascsv = False
        if len(sys.argv)>3:
            if sys.argv[3] == "CSV":
                ascsv = True
        scrapefacebook(START_PAGE, output, ascsv)

Ciao Luca, questo script è usabile anche in caso io voglia scaricare i post di un profilo facebook o funziona solo su una pagina? Grazie

Funzionava anche su profili pubblici, ma da un po' di tempo Facebook ha cambiato le chiamate, per cui bisognerebbe rifare da capo lo scraping