grande refactor -> Delojza now is a class

This commit is contained in:
Tomáš Mládek 2019-04-17 18:05:17 +02:00 committed by Tomáš Mládek
parent 4ff94b2af6
commit c249de6e64
5 changed files with 238 additions and 230 deletions

2
.gitignore vendored
View file

@ -1,4 +1,4 @@
out
delojza.ini
downloaded.lst
delojza.log
initial.txt

8
delojza.ini Normal file
View file

@ -0,0 +1,8 @@
[delojza]
tg_api_key = ***REMOVED***
[tumblr]
consumer_key = ***REMOVED***
consumer_secret = ***REMOVED***
oauth_key = ***REMOVED***
oauth_secret = ***REMOVED***

View file

@ -1,10 +1,12 @@
#!/usr/bin/env python3
import errno
import logging
import os
import re
import shutil
import sys
from configparser import ConfigParser
from glob import glob
import filetype
@ -15,249 +17,260 @@ import youtube_dl
from telegram import MessageEntity
from telegram.ext import Updater, CommandHandler, MessageHandler, Filters
DIR = os.path.dirname(os.path.realpath(__file__))
TMP_DIR = '/var/tmp'
OUT_DIR = DIR + '/out'
logging.basicConfig(level=logging.INFO,
format='%(asctime)s - %(name)s - %(levelname)s - %(message)s')
logger = logging.getLogger("kunsax")
client = pytumblr.TumblrRestClient(
'***REMOVED***',
'***REMOVED***',
'***REMOVED***',
'***REMOVED***'
)
markov = None
def add_to_corpus(text):
global markov
text = text.lower()
new_sentence = markovify.NewlineText(text)
markov = markovify.combine([markov, new_sentence])
with open("initial.txt", 'a') as f:
f.write(text + '\n')
def datestr(date):
return date.strftime("%Y-%m-%d@%H%M")
def mkdir_p(path):
try:
os.makedirs(path)
except OSError as exc: # Python >2.5
except OSError as exc:
if exc.errno == errno.EEXIST and os.path.isdir(path):
pass
else:
raise
def ytdl_has(url):
ies = youtube_dl.extractor.gen_extractors()
for ie in ies:
if ie.suitable(url) and ie.IE_NAME != 'generic' \
and '/channel/' not in url:
# Site has dedicated extractor
return True
return False
def datestr(date):
return date.strftime("%Y-%m-%d@%H%M")
def download_ydl(urls, subdir, date, extract=False, filename=None):
ydl_opts = {
'noplaylist': True,
'restrictfilenames': True,
'outtmpl': TMP_DIR + '/' + datestr(date) + '__%(title)s__%(id)s.%(ext)s'
}
if extract:
ydl_opts['format'] = 'bestaudio'
# ydl_opts['postprocessors'] = [{
# 'key': 'FFmpegExtractAudio',
# 'preferredcodec': 'wav'
# }]
with youtube_dl.YoutubeDL(ydl_opts) as ydl:
ydl.download(urls)
out_dir = OUT_DIR + '/' + subdir + '/'
for filename in map(ydl.prepare_filename, map(ydl.extract_info, urls)):
globbeds = glob(os.path.splitext(filename)[0] + '.*')
for globbed in globbeds:
logger.info("Moving %s to %s..." % (globbed, out_dir))
shutil.move(globbed, out_dir)
return []
class DelojzaBot:
def __init__(self, tg_api_key, out_dir, tmp_dir='/var/tmp', tumblr_keys=None, markov=None):
self.logger = logging.getLogger("kunsax")
self.out_dir = out_dir
self.logger.debug('OUT_DIR: ' + out_dir)
self.tmp_dir = tmp_dir
self.logger.debug('TMP_DIR: ' + tmp_dir)
self.markov = markov
def download_raw(urls, subdir, date, extract=False, filename=None):
filenames = []
for url in urls:
local_filename = OUT_DIR + '/' + subdir + '/' + "%s__%s" % (datestr(date), filename or url.split('/')[-1])
r = requests.get(url, stream=True)
with open(local_filename, 'wb') as f:
for chunk in r.iter_content(chunk_size=1024):
if chunk:
f.write(chunk)
if not re.match(r'.*\..{3,5}$', os.path.split(local_filename)[-1]):
kind = filetype.guess(local_filename)
if kind is None:
logger.error("File has no extension and could not be determined!")
else:
logger.info('Moving file without extension... %s?' % kind.extension)
shutil.move(local_filename, local_filename + '.' + kind.extension)
filenames.append(local_filename)
return filenames
self.updater = Updater(tg_api_key)
dp = self.updater.dispatcher
dp.add_handler(CommandHandler("start", self.tg_start))
dp.add_error_handler(self.tg_error)
last_hashtag = None
dp.add_handler(MessageHandler(Filters.entity(MessageEntity.URL), self.handle_url))
dp.add_handler(
MessageHandler(
Filters.photo | Filters.video | Filters.video_note | Filters.audio | Filters.voice | Filters.document,
self.handle_rest))
dp.add_handler(MessageHandler(Filters.entity(MessageEntity.HASHTAG), self.handle_hashtag))
dp.add_handler(MessageHandler(Filters.text, self.handle_text))
if tumblr_keys:
self.client = pytumblr.TumblrRestClient(*tumblr_keys)
def get_first_hashtag(message):
global last_hashtag
hashtags = list(map(message.parse_entity,
list(filter(lambda e: e.type == 'hashtag', message.entities))))
hashtags += list(map(message.parse_caption_entity,
list(filter(lambda e: e.type == 'hashtag', message.caption_entities))))
if len(hashtags) == 0:
if last_hashtag is not None and last_hashtag[0] == message.from_user:
prehashtag = last_hashtag[1]
last_hashtag = None
else:
return None
else:
prehashtag = hashtags[0]
hashtag = prehashtag[1:].upper()
if "PRAS" in hashtag:
hashtag = "PRAS"
return hashtag
self.last_hashtag = None
@staticmethod
def ytdl_can(url):
ies = youtube_dl.extractor.gen_extractors()
for ie in ies:
if ie.suitable(url) and ie.IE_NAME != 'generic' \
and '/channel/' not in url:
# Site has dedicated extractor
return True
return False
def handle_hashtag(bot, update):
global last_hashtag
hashtags = list(map(update.message.parse_entity,
list(filter(lambda e: e.type == 'hashtag', update.message.entities))))
if len(hashtags) > 0:
last_hashtag = (update.message.from_user, hashtags[0])
def download_ytdl(self, urls, subdir, date, extract=False, filename=None):
ydl_opts = {
'noplaylist': True,
'restrictfilenames': True,
'outtmpl': os.path.join(self.tmp_dir, datestr(date), '__%(title)s__%(id)s.%(ext)s') # HOW?
}
if extract:
ydl_opts['format'] = 'bestaudio'
# ydl_opts['postprocessors'] = [{
# 'key': 'FFmpegExtractAudio',
# 'preferredcodec': 'wav'
# }]
with youtube_dl.YoutubeDL(ydl_opts) as ydl:
ydl.download(urls)
out_dir = os.path.join(self.out_dir, subdir)
for filename in map(ydl.prepare_filename, map(ydl.extract_info, urls)):
globbeds = glob(os.path.splitext(filename)[0] + '.*')
for globbed in globbeds:
self.logger.info("Moving %s to %s..." % (globbed, out_dir))
shutil.move(globbed, out_dir)
return []
# noinspection PyBroadException
def handle(urls, message, download, tumblr=False, filename=None):
try:
hashtag = get_first_hashtag(message)
if hashtag is None:
logger.info("Ignoring %s due to no hashtag present..." % urls)
return
logger.info("Downloading %s" % urls)
reply = 'Downloading'
if hashtag:
mkdir_p(OUT_DIR + '/' + hashtag)
reply += ' to "' + hashtag + '"'
reply += '...'
extract = False
if hashtag in ('AUDIO', 'RADIO') and download != download_raw:
extract = True
reply += ' (And also guessing you want to extract the audio)'
message.reply_text(reply)
filenames = download(urls,
hashtag or '.', message.date,
extract=extract, filename=filename)
if hashtag == 'TUMBLR':
message.reply_text('(btw, queueing to tumblr)')
for filename in filenames:
client.create_photo('kunsaxan', state="queue", data=filename)
def download_raw(self, urls, subdir, date, extract=False, filename=None):
filenames = []
for url in urls:
local_filename = os.path.join(self.out_dir, subdir,
"%s__%s" % (datestr(date), filename or url.split('/')[-1]))
r = requests.get(url, stream=True)
with open(local_filename, 'wb') as f:
for chunk in r.iter_content(chunk_size=1024):
if chunk:
f.write(chunk)
if not re.match(r'.*\..{3,5}$', os.path.split(local_filename)[-1]):
kind = filetype.guess(local_filename)
if kind is None:
self.logger.error("File has no extension and could not be determined!")
else:
self.logger.info('Moving file without extension... %s?' % kind.extension)
shutil.move(local_filename, local_filename + '.' + kind.extension)
filenames.append(local_filename)
return filenames
except:
_, exc_value, __ = sys.exc_info()
if "Timed out" not in str(exc_value):
message.reply_text("Something is FUCKED: %s" % exc_value)
def get_first_hashtag(self, message):
hashtags = list(map(message.parse_entity,
list(filter(lambda e: e.type == 'hashtag', message.entities))))
hashtags += list(map(message.parse_caption_entity,
list(filter(lambda e: e.type == 'hashtag', message.caption_entities))))
if len(hashtags) == 0:
if self.last_hashtag is not None and self.last_hashtag[0] == message.from_user:
prehashtag = self.last_hashtag[1]
self.last_hashtag = None
else:
return None
else:
prehashtag = hashtags[0]
hashtag = prehashtag[1:].upper()
if "PRAS" in hashtag:
hashtag = "PRAS"
return hashtag
def handle_hashtag(self, bot, update):
hashtags = list(map(update.message.parse_entity,
list(filter(lambda e: e.type == 'hashtag', update.message.entities))))
if len(hashtags) > 0:
self.last_hashtag = (update.message.from_user, hashtags[0])
# noinspection PyBroadException
def handle(self, urls, message, download, tumblr=False, filename=None):
try:
hashtag = self.get_first_hashtag(message)
if hashtag is None:
self.logger.info("Ignoring %s due to no hashtag present..." % urls)
return
self.logger.info("Downloading %s under '%s'" % (urls, hashtag))
reply = 'Downloading'
if hashtag:
mkdir_p(os.path.join(self.out_dir, hashtag))
reply += ' to "' + hashtag + '"'
reply += '...'
extract = False
if hashtag in ('AUDIO', 'RADIO') and download != self.download_raw:
extract = True
reply += ' (And also guessing you want to extract the audio)'
message.reply_text(reply)
filenames = download(urls,
hashtag or '.', message.date,
extract=extract, filename=filename)
if hashtag == 'TUMBLR':
message.reply_text('(btw, queueing to tumblr)')
for filename in filenames:
self.client.create_photo('kunsaxan', state="queue", data=filename)
return filenames
except:
_, exc_value, __ = sys.exc_info()
if "Timed out" not in str(exc_value):
message.reply_text("Something is FUCKED: %s" % exc_value)
def handle_url(self, bot, update):
ytdl_urls = list(filter(self.ytdl_can,
map(lambda e: update.message.parse_entity(e),
filter(lambda e: e.type == 'url',
update.message.entities))))
if len(ytdl_urls) > 0:
self.handle(ytdl_urls, update.message, self.download_ytdl)
# noinspection PyBroadException
def handle_rest(self, bot, update):
file, filename, tumblr = None, None, False
if len(update.message.photo) > 0:
photo = max(update.message.photo, key=lambda p: p.width)
file = photo.file_id
tumblr = True
elif update.message.document is not None:
filename = update.message.document.file_name
file = update.message.document.file_id
elif update.message.audio is not None:
filename = update.message.audio.title
file = update.message.audio.file_id
elif update.message.video is not None:
file = update.message.video.file_id
elif update.message.video_note is not None:
file = update.message.video_note.file_id
elif update.message.voice is not None:
file = update.message.voice.file_id
if file is not None:
url = bot.getFile(file).file_path
self.handle([url], update.message, self.download_raw, tumblr=tumblr, filename=filename)
def handle_text(self, bot, update):
self.markov.add_to_corpus(update.message.text)
def tg_start(self, bot, update):
update.message.reply_text(self.markov.make_sentence())
def tg_error(self, bot, update, error):
self.logger.error(error)
if "Timed out" in str(error):
if update is not None:
update.message.reply_text(self.markov.make_sentence(tries=100) or "Mmmm, I like it...")
self.handle_rest(bot, update)
else:
if update is not None:
update.message.reply_text("Something is fucked: %s" % error)
def run_idle(self):
self.updater.start_polling()
self.logger.info("Started Telegram bot...")
self.updater.idle()
def handle_url(bot, update):
ytdl_urls = list(filter(ytdl_has,
map(lambda e: update.message.parse_entity(e),
filter(lambda e: e.type == 'url',
update.message.entities))))
if len(ytdl_urls) > 0:
handle(ytdl_urls, update.message, download_ydl)
class MarkovBlabberer:
def __init__(self, filepath):
self.logger = logging.getLogger('markov')
self.filepath = filepath
with open(filepath) as f:
text = f.read()
self.markov = markovify.NewlineText(text.lower())
self.logger.info("Sentence of the day: " + self.make_sentence())
# noinspection PyBroadException
def handle_rest(bot, update):
file, filename, tumblr = None, None, False
if len(update.message.photo) > 0:
photo = max(update.message.photo, key=lambda p: p.width)
file = photo.file_id
tumblr = True
elif update.message.document is not None:
filename = update.message.document.file_name
file = update.message.document.file_id
elif update.message.audio is not None:
filename = update.message.audio.title
file = update.message.audio.file_id
elif update.message.video is not None:
file = update.message.video.file_id
elif update.message.video_note is not None:
file = update.message.video_note.file_id
elif update.message.voice is not None:
file = update.message.voice.file_id
def make_sentence(self, tries=100):
return self.markov.make_sentence(tries=tries)
if file is not None:
url = bot.getFile(file).file_path
handle([url], update.message, download_raw, tumblr=tumblr, filename=filename)
def handle_text(bot, update):
add_to_corpus(update.message.text)
def start(bot, update):
update.message.reply_text(markov.make_sentence())
def error(bot, update, error):
logger.error(error)
if "Timed out" in str(error):
if update is not None:
update.message.reply_text(markov.make_sentence(tries=100) or "Mmmm, I like it...")
handle_rest(bot, update)
else:
if update is not None:
update.message.reply_text("Something is fucked: %s" % error)
def main():
global markov
with open("initial.txt") as f:
text = f.read()
markov = markovify.NewlineText(text.lower())
logger.info("Sentence of the day: " + markov.make_sentence())
updater = Updater("***REMOVED***")
dp = updater.dispatcher
dp.add_handler(CommandHandler("start", start))
dp.add_error_handler(error)
dp.add_handler(MessageHandler(Filters.entity(MessageEntity.URL), handle_url))
dp.add_handler(
MessageHandler(
Filters.photo | Filters.video | Filters.video_note | Filters.audio | Filters.voice | Filters.document,
handle_rest))
dp.add_handler(MessageHandler(Filters.entity(MessageEntity.HASHTAG), handle_hashtag))
dp.add_handler(MessageHandler(Filters.text, handle_text))
updater.start_polling()
logger.info("Started Telegram bot...")
updater.idle()
def add_to_corpus(self, text):
text = text.lower()
new_sentence = markovify.NewlineText(text)
self.markov = markovify.combine([self.markov, new_sentence])
with open(self.filepath, 'a') as f:
f.write(text + '\n')
if __name__ == '__main__':
main()
logging.basicConfig(level=logging.INFO,
format='%(asctime)s - %(name)s - %(levelname)s - %(message)s')
_DIR_ = os.path.dirname(os.path.realpath(__file__))
CONFIG_PATHS = ['/etc/delojza/delojza.ini',
os.path.join(os.getenv("HOME"), ".config/delojza/delojza.ini"),
os.path.join(_DIR_, "delojza.ini")]
config = ConfigParser()
try:
CONF_FILE = next(conf_path for conf_path in CONFIG_PATHS if os.path.isfile(conf_path))
config.read(CONF_FILE)
except StopIteration:
logging.error("No config file found, stopping.")
sys.exit(-1)
markov = MarkovBlabberer("initial.txt")
delojza = DelojzaBot(config.get('delojza', 'tg_api_key'),
config.get('delojza', 'OUT_DIR', fallback=os.path.join(_DIR_, "out")),
tmp_dir=config.get('delojza', 'tmp_dir', fallback="/var/tmp"),
tumblr_keys=(config.get('tumblr', 'consumer_key'),
config.get('tumblr', 'consumer_secret'),
config.get('tumblr', 'oauth_key'),
config.get('tumblr', 'oauth_secret')),
markov=None)
delojza.run_idle()

View file

@ -1,6 +0,0 @@
#!/bin/bash
DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"
cd ${DIR}
./update.sh &
source ./.venv/bin/activate
python3 delojza.py 2>&1 |tee -a delojza.log

View file

@ -1,7 +0,0 @@
#!/bin/bash
while :;do
NUM=$( grep 'INFO - Downloading' delojza.log|wc -l)
echo $NUM
curl -s 'https://kunsaxan.sdbs.cz/counter.php?key=delojza7953713b19ef2ea055156c8dc175bf80&count='$NUM
sleep 300;
done