def prefer_envar(configs: dict) -> dict:
for config in list(configs):
config_envar = f"{ENVAR_PREFIX}{config}".lower()
if os.environ.get(config_envar):
configs[config]=os.environ.get(config_envar)
log.info(f"loading {config_envar} from envar. Value: {configs.get(config)}")
else:
log.debug(f"no environment config for: {config_envar}")
return configs
def init(self):
log.info("using cobe to generate comments")
main_db = self.config.get("cobe_main_db")
# make sure db was initialized correctly
if os.path.isfile(main_db):
# set the initial size
self.size = os.path.getsize(main_db)
else:
log.info(f"cobe db failed to initialize. exiting")
sys.exit()
log.debug('filling cobe database for commenting')
# loop through learning comments until we reach the min db size
while self.size <= tobytes(self.config.get("cobe_min_db_size")):
log.info(
f"cobe db size is: {str(bytesto(self.size, 'm'))}mb, need {self.config.get('cobe_min_db_size')} - learning..."
)
# just learn from random subreddits for now
subreddit = get_subreddit(getsubclass=True)
log.info(f"learning from /r/{subreddit}")
# get the comment generator function from pushshift
comments = self.psapi.get_comments(subreddit)
# go through 500 comments per subreddit
for x in range(500):
# get the comment from the generator function
try:
comment = next(comments)
except StopIteration as e:
log.info(f"end of comments")
# bot responses are better when it learns from short comments
if len(comment.body) < 240:
log.debug(
f"learning comment: {comment.body.encode('utf8')}")
# only learn comments that don't contain an avoid word
if not any(word in comment.body for word in AVOID_WORDS):
self.brain.learn(comment.body.encode("utf8"))
# update the class size variable so the while loop
# knows when to break
self.size = os.path.getsize(main_db)
log.info(
f"database min size ({self.config.get('cobe_min_db_size')}) reached"
)
self.ready = True
import random
from apis import reddit_api
from logs.logger import log
from config.reddit.reddit_sub_lists import REDDIT_APPROVED_SUBS
from config.common_config import CONFIG_ROOT
with open(f"{CONFIG_ROOT}/reddit/reddit_avoid_subs.txt", "r") as subfile:
AVOID_SUBS = subfile.read().splitlines()
subfile.close()
with open(f"{CONFIG_ROOT}/reddit/reddit_avoid_words.txt", "r") as wordfile:
AVOID_WORDS = wordfile.read().splitlines()
wordfile.close()
log.debug(f"avoiding subs: {AVOID_SUBS}")
def get_subreddit(nsfw=False, getsubclass=False):
# if the subreddit list is being used jut return one from there
if REDDIT_APPROVED_SUBS:
log.info(f"picking subreddit from approved list")
subreddit = reddit_api.subreddit(
random.choice(REDDIT_APPROVED_SUBS).strip())
log.info(f"using subreddit: {subreddit.display_name}")
else:
log.info(f"picking a random subreddit")
# otherwise we'll do some logic to get a random subreddit
subreddit_ok = False
while not subreddit_ok:
subreddit = reddit_api.random_subreddit(nsfw=nsfw)