alexis2/parser.py

#!/usr/bin/python3 
import threading
from queue import Queue 

from db import db
from source import Source
from linklist import LinkList
from news import News

def parseSource(s):
    sel=None
    rss=False
    if "selector" in s:
        sel=s["selector"]
    if "rss" in s:
        rss=bool(["rss"])
    l=LinkList(s["link"],sel,rss=rss)

    kw=d.get_keywords(s["category"])
    if l is not None:
        print("[Source] %s: %d total articles" % ( s["name"], len(l.links)))
        for a in l.links:
            if not d.article_exists(a):
                q.put(a)
        print("[Source] %s: %d new articles" % ( s["name"], q.qsize()))


    if not q.empty():
        threads=[]
        for i in range(num_worker_threads):
            t = threading.Thread(target=news_worker,args=[s["_id"],kw])
            t.daemon=True
            t.start()
            threads.append(t)
    
        print("Waiting for %d queue elements" % q.qsize())
        q.join() #block until all tasks are done

        #stop workers
        for i in range(num_worker_threads):
            q.put(None)
        for t in threads:
            t.join()
    print("[Source] %s => Finished parsing" %s["name"])

def news_worker(source_id,kw):
    while True:
        item = q.get()
        if item is None:
            break
        n=News(item,source_id,kw)
        print("[Thread %s] %s" % (threading.current_thread().name,n.URL))
        d.insert_article(n.get())
        q.task_done()

q=Queue()
d=db()
d.purge_error()
num_worker_threads=100
for s in d.sources():
    parseSource(s)