etherpump/etherdump/commands/rss.py

#!/usr/bin/env python
from __future__ import print_function
from argparse import ArgumentParser
import sys, json, re, os
from datetime import datetime
from urllib import urlencode
from urllib2 import HTTPError
from jinja2 import FileSystemLoader, Environment
from common import *
from time import sleep
import dateutil.parser

"""
rss:
    Generate an RSS feed from an etherdump.


TODO NEXT
add back limit and ordering parameters to create filters to make a latest changes feed!

"""

def group (items, key=lambda x: x):
    ret = []
    keys = {}
    for item in items:
        k = key(item)
        if k not in keys:
            keys[k] = []
        keys[k].append(item)
    for k in sorted(keys):
        keys[k].sort()
        ret.append(keys[k])
    return ret

def base (x):
    return re.sub(r"(\.raw\.html)|(\.diff\.html)|(\.meta\.json)|(\.raw\.txt)$", "", x)

def main (args):
    p = ArgumentParser("Check for pads that have changed since last sync (according to .meta.json)")

    p.add_argument("input", nargs="+", help="filenames")
    p.add_argument("--templates", default=None, help="templates path")

    p.add_argument("--padinfo", default=".etherdump/settings.json", help="settings, default: .etherdump/settings.json")
    p.add_argument("--zerorevs", default=False, action="store_true", help="include pads with zero revisions, default: False (i.e. pads with no revisions are skipped)")

    p.add_argument("--skip", default=None, type=int, help="skip this many items, default: None")
    p.add_argument("--type", default="lastchanges", help="type of feed, default: lastchanges")

    p.add_argument("--title", default="etherpad", help="rss feed channel title, default: etherpad")
    p.add_argument("--description", default="", help="channel description, default: empty")
    p.add_argument("--language", default="en-US", help="feed language, default: en-US")
    p.add_argument("--updatePeriod", default="daily", help="updatePeriod, possible values: hourly, daily, weekly, monthly, yearly; default: daily")
    p.add_argument("--updateFrequency", default=1, type=int, help="update frequency within the update period (where 2 would mean twice per period); default: 1")
    p.add_argument("--padurl", default=None, help="to use as channel link, default: padurl")
    p.add_argument("--feedurl", default="feed.xml", help="to use as feeds own (self) link, default: feed.xml")
    p.add_argument("--generator", default="https://gitlab.com/activearchives/etherdump", help="generator, default: https://gitlab.com/activearchives/etherdump")
    p.add_argument("--itemlink", default="pad", help="item to link to in feed, possible values: pad, text, html, dhtml; default: pad")

    p.add_argument("--chronological", default=False, action="store_true", help="order chronologically, default: False")
    args = p.parse_args(args)

    tmpath = args.templates
    if tmpath == None:
        tmpath = os.path.split(os.path.abspath(__file__))[0]
        tmpath = os.path.split(tmpath)[0]
        tmpath = os.path.join(tmpath, "data", "templates")

    env = Environment(loader=FileSystemLoader(tmpath))
    template = env.get_template("rss.xml")

    info = loadpadinfo(args.padinfo)

    inputs = args.input
    inputs.sort()
    inputs = group(inputs, base)

    def loadmeta(paths):
        for p in paths:
            if p.endswith(".meta.json"):
                with open(p) as f:
                    return json.load(f)

    def fixdates (padmeta):
        d = dateutil.parser.parse(padmeta["lastedited_iso"])
        padmeta["lastedited"] = d
        padmeta["lastedited_822"] = d.strftime("%a, %d %b %Y %H:%M:%S +0000")
        return padmeta

    pads = map(loadmeta, inputs)
    pads = map(fixdates, pads)
    args.pads = pads

    # args.timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")

    padurlbase = re.sub(r"api/1.2.9/$", "p/", info["apiurl"])
    if type(padurlbase) == unicode:
        padurlbase = padurlbase.encode("utf-8")
    args.padurl = args.padurl or padurlbase
    args.utcnow = datetime.utcnow().strftime("%a, %d %b %Y %H:%M:%S +0000")

    # from pprint import pprint
    # pprint(inputs)

    print (template.render(vars(args)).encode("utf-8"))