home
·
contact
·
privacy
projects
/
plomlombot-irc.git
/ blobdiff
commit
grep
author
committer
pickaxe
?
search:
re
summary
|
shortlog
|
log
|
commit
|
commitdiff
|
tree
raw
|
inline
| side by side
Minor improvements to markov generator.
[plomlombot-irc.git]
/
plomlombot.py
diff --git
a/plomlombot.py
b/plomlombot.py
index 786b072bf483d1c69cfd19e18c18895214bbbf6b..7bf2208ab94a71774b4f1b7c6a49cf0e58661ad0 100755
(executable)
--- a/
plomlombot.py
+++ b/
plomlombot.py
@@
-190,16
+190,23
@@
def handle_command(command, argument, notice, target, session):
if not os.access(markovfeed_name, os.F_OK):
notice("NOT ENOUGH TEXT TO MARKOV.")
return
if not os.access(markovfeed_name, os.F_OK):
notice("NOT ENOUGH TEXT TO MARKOV.")
return
+
+ # Lowercase incoming lines, ensure they end in a sentence end mark.
file = open(markovfeed_name, "r")
lines = file.readlines()
file.close()
tokens = []
for line in lines:
file = open(markovfeed_name, "r")
lines = file.readlines()
file.close()
tokens = []
for line in lines:
- line = line.replace("\n", "").lower()
+ line = line.lower().replace("\n", "")
+ if line[-1] not in ".!?":
+ line += "."
tokens += line.split()
if len(tokens) <= select_length:
notice("NOT ENOUGH TEXT TO MARKOV.")
return
tokens += line.split()
if len(tokens) <= select_length:
notice("NOT ENOUGH TEXT TO MARKOV.")
return
+
+ # Replace URLs with escape string for now, so that the Markov selector
+ # won't see them as different strings. Stash replaced URLs in urls.
urls = []
url_escape = "\nURL"
url_starts = ["http://", "https://", "<http://", "<https://"]
urls = []
url_escape = "\nURL"
url_starts = ["http://", "https://", "<http://", "<https://"]
@@
-215,6
+222,9
@@
def handle_command(command, argument, notice, target, session):
urls += [tokens[i][:length]]
tokens[i] = url_escape + tokens[i][length:]
break
urls += [tokens[i][:length]]
tokens[i] = url_escape + tokens[i][length:]
break
+
+ # For each snippet of select_length, use markov() to find continuation
+ # token from selections. Replace present users' names with malkovich.
for i in range(len(tokens) - select_length):
token_list = []
for j in range(select_length + 1):
for i in range(len(tokens) - select_length):
token_list = []
for j in range(select_length + 1):
@@
-224,11
+234,12
@@
def handle_command(command, argument, notice, target, session):
for i in range(select_length):
snippet += [""]
msg = ""
for i in range(select_length):
snippet += [""]
msg = ""
+ malkovich = "malkovich"
while 1:
new_end = markov(snippet)
while 1:
new_end = markov(snippet)
- for name in session.uses_in_chan:
+ for name in session.use
r
s_in_chan:
if new_end[:len(name)] == name.lower():
if new_end[:len(name)] == name.lower():
- new_end =
"malkovich"
+ new_end[len(name):]
+ new_end =
malkovich
+ new_end[len(name):]
break
if len(msg) + len(new_end) > 200:
break
break
if len(msg) + len(new_end) > 200:
break
@@
-236,17
+247,21
@@
def handle_command(command, argument, notice, target, session):
for i in range(select_length - 1):
snippet[i] = snippet[i + 1]
snippet[select_length - 1] = new_end
for i in range(select_length - 1):
snippet[i] = snippet[i + 1]
snippet[select_length - 1] = new_end
+
+ # Replace occurences of url escape string with random choice from urls.
while True:
index = msg.find(url_escape)
if index < 0:
break
msg = msg.replace(url_escape, choice(urls), 1)
while True:
index = msg.find(url_escape)
if index < 0:
break
msg = msg.replace(url_escape, choice(urls), 1)
- notice(msg + "malkovich.")
+
+ # More meaningful ways to randomly end sentences.
+ notice(msg + malkovich + ".")
def twt():
def try_open(mode):
try:
def twt():
def try_open(mode):
try:
- twtfile = open(session.twtfile,
"w"
)
+ twtfile = open(session.twtfile,
mode
)
except (PermissionError, FileNotFoundError) as err:
notice("CAN'T ACCESS OR CREATE TWT FILE: " + str(err))
return None
except (PermissionError, FileNotFoundError) as err:
notice("CAN'T ACCESS OR CREATE TWT FILE: " + str(err))
return None
@@
-293,6
+308,7
@@
def handle_url(url, notice, show_url=False):
except (requests.exceptions.TooManyRedirects,
requests.exceptions.ConnectionError,
requests.exceptions.InvalidURL,
except (requests.exceptions.TooManyRedirects,
requests.exceptions.ConnectionError,
requests.exceptions.InvalidURL,
+ UnicodeError,
requests.exceptions.InvalidSchema) as error:
notice("TROUBLE FOLLOWING URL: " + str(error))
return
requests.exceptions.InvalidSchema) as error:
notice("TROUBLE FOLLOWING URL: " + str(error))
return
@@
-314,7
+330,7
@@
class Session:
self.io = io
self.nickname = nickname
self.channel = channel
self.io = io
self.nickname = nickname
self.channel = channel
- self.uses_in_chan = []
+ self.use
r
s_in_chan = []
self.twtfile = twtfile
self.io.send_line("NICK " + self.nickname)
self.io.send_line("USER " + username + " 0 * : ")
self.twtfile = twtfile
self.io.send_line("NICK " + self.nickname)
self.io.send_line("USER " + username + " 0 * : ")
@@
-382,14
+398,14
@@
class Session:
elif tokens[1] == "353":
names = tokens[5:]
names[0] = names[0][1:]
elif tokens[1] == "353":
names = tokens[5:]
names[0] = names[0][1:]
- self.uses_in_chan += names
+ self.use
r
s_in_chan += names
elif tokens[1] == "JOIN":
name = name_from_join_or_part(tokens)
if name != self.nickname:
elif tokens[1] == "JOIN":
name = name_from_join_or_part(tokens)
if name != self.nickname:
- self.uses_in_chan += [name]
+ self.use
r
s_in_chan += [name]
elif tokens[1] == "PART":
name = name_from_join_or_part(tokens)
elif tokens[1] == "PART":
name = name_from_join_or_part(tokens)
- del(self.use
s_in_chan[self.use
s_in_chan.index(name)])
+ del(self.use
rs_in_chan[self.user
s_in_chan.index(name)])
def parse_command_line_arguments():
parser = argparse.ArgumentParser()
def parse_command_line_arguments():
parser = argparse.ArgumentParser()