Add ?censor and ?safe params

.gitignore

@@ -0,0 +1 @@
+/__pycache__/

CONTRIBUTING.md

@@ -2,7 +2,7 @@
 
 ## Sorting
 
-Don't bother sorting the file. Every now and then I'll sort and clean it up.
+Don't bother sorting the files. Every now and then I'll sort and clean up.
 
 ## License
 

README.markdown

@@ -15,6 +15,12 @@ Some interesting usage for that can be:
 git config --global alias.yolo '!git commit -m "$(curl -s https://whatthecommit.com/index.txt)"'
 ```
 
+## Make it safe for work
+
+    https://whatthecommit.com/?safe filters out any unsafe or swearing messages
+    https://whatthecommit.com/?censor censors them instead, and filters messages that wouldn't be funny if censored
+    https://whatthecommit.com/?censor=* censors using custom pattern
+
 Or use one of the following VSCode Extensions:
 
 - [WhatTheCommit](https://marketplace.visualstudio.com/items?itemName=Gaardsholt.vscode-whatthecommit) 

commit.py

@@ -1,89 +1,35 @@
 import os
-import sys
-import random
-import re
 import json
 import signal
-from typing import Dict, List
-
-from hashlib import md5
 
 import tornado.httpserver
 import tornado.ioloop
 import tornado.web
 from tornado.escape import xhtml_unescape
 from tornado.options import define, options
 
-define("port", default=5000, help="run on the given port", type=int)
-
-names = ['Nick', 'Steve', 'Andy', 'Qi', 'Fanny', 'Sarah', 'Cord', 'Todd',
-    'Chris', 'Pasha', 'Gabe', 'Tony', 'Jason', 'Randal', 'Ali', 'Kim',
-    'Rainer', 'Guillaume', 'Kelan', 'David', 'John', 'Stephen', 'Tom', 'Steven',
-    'Jen', 'Marcus', 'Edy', 'Rachel', 'Ethan', 'Dan', 'Darren', 'Greg']
-
-humans_file = os.path.join(os.path.dirname(__file__), 'static', 'humans.txt')
-messages_file = os.path.join(os.path.dirname(__file__), 'commit_messages.txt')
-messages: Dict[str, str] = {}
-
-# Create a hash table of all commit messages
-with open(messages_file, 'r', encoding='utf-8') as messages_input:
-    for line in messages_input.readlines():
-        messages[md5(line.encode('utf-8')).hexdigest()] = line
-
-names: List[str] = []
-
-with open(humans_file, 'r', encoding='utf-8') as humans_input:
-    humans_content = humans_input.read()
-    for line in humans_content.split("\n"):
-        if "Name:" in line:
-            data = line[6:].rstrip()
-            if data.find("github:") == 0:
-                names.append(data[7:])
-            else:
-                names.append(data.split(" ")[0])
-
-num_re = re.compile(r"XNUM([0-9,]*)X")
-
-def fill_line(message):
-    message = message.replace('XNAMEX', random.choice(names))
-    message = message.replace('XUPPERNAMEX', random.choice(names).upper())
-    message = message.replace('XLOWERNAMEX', random.choice(names).lower())
-
-    nums = num_re.findall(message)
-
-    while nums:
-        start = 1
-        end = 999
-        value = nums.pop(0) or str(end)
-        if "," in value:
-            position = value.index(",")
-            if position == 0: # XNUM,5X
-                end = int(value[1:])
-            elif position == len(value) - 1: # XNUM5,X
-                start = int(value[:position])
-            else: # XNUM1,5X
-                start = int(value[:position])
-                end = int(value[position+1:])
-        else:
-            end = int(value)
-        if start > end:
-            end = start * 2
-
-        randint = random.randint(start, end)
-        message = num_re.sub(str(randint), message, count=1)
+import messages
 
-    return message
+define("port", default=5000, help="run on the given port", type=int)
 
 class MainHandler(tornado.web.RequestHandler):
     def get(self, message_hash=None):
-        if not message_hash:
-            message_hash = random.choice(list(messages.keys()))
-        elif message_hash not in messages:
-            raise tornado.web.HTTPError(404)
+        safe_only = self.get_argument("safe", default=False) != False
+        censor = self.get_argument("censor", default=False)
+        if censor == "":
+            censor = True
 
-        message = fill_line(messages[message_hash])
+        found_message = messages.find_by_md5(message_hash, censor=censor)
 
-        self.output_message(message, message_hash)
+        if message_hash and not found_message:
+            raise tornado.web.HTTPError(404)
+
+        if found_message:
+            self.output_message(found_message, message_hash)
+        else:
+            message, generated_message_hash = (
+                messages.generate(safe_only=safe_only, censor=censor))
+            self.output_message(message, generated_message_hash)
 
     def output_message(self, message, message_hash):
         self.set_header('X-Message-Hash', message_hash)
@@ -134,6 +80,8 @@ def try_exit(self):
     tornado.options.parse_command_line()
     signal.signal(signal.SIGINT, application.signal_handler)
     http_server = tornado.httpserver.HTTPServer(application)
-    http_server.listen(os.environ.get("PORT", 5000))
+    port = os.environ.get("PORT", 5000)
+    print("ready for requests (on port %s)" % (port))
+    http_server.listen(port)
     tornado.ioloop.PeriodicCallback(application.try_exit, 100).start()
     tornado.ioloop.IOLoop.instance().start()

commit_messages/README.md

@@ -0,0 +1,58 @@
+these are the source text files from which all the commit messages
+are generated
+
+- put commit messages with swear words in censorable.txt, this makes
+them funny by using [grawlixes](https://en.wikipedia.org/wiki/Grawlix)
+
+- put other nsfw in unsafe, because they cannot be made funny and safe
+by using grawlixes
+
+- put safe for work messages in safe.txt.
+
+## Script for filtering of swear words
+
+This was used to create the initial list of unsafe words.
+
+```py
+BAD_WORDS = [
+    "shit",
+    "piss",
+    "fuck",
+    "cunt",
+    "cocksucker",
+    "motherfucker",
+    "tits",
+
+    "cock",
+    "fucker",
+
+    "fart",
+    "turd",
+    "twat",
+
+    "dicksucker",
+    "fucking",
+
+    "sex",
+    "sexy",
+]
+
+safe_f = open("commit_messages/safe.txt", "w")
+unsafe_f = open("commit_messages/unsafe.txt", "w")
+
+with open("./commit_messages.txt") as original_f:
+    for line in original_f:
+        bad = False
+
+        for bad_word in BAD_WORDS:
+            if bad_word in line.lower():
+                bad = True
+
+        if bad:
+            unsafe_f.write(line)
+        else:
+            safe_f.write(line)
+
+safe_f.close()
+unsafe_f.close()
+```

commit_messages/censorable.txt

@@ -0,0 +1,44 @@
+"Get that shit outta my master."
+Code was clean until manager requested to fuck it up
+Derp search/replace fuckup
+Either Hot Shit or Total Bollocks
+FUCKING XUPPERNAMEX
+Fixed some shit
+Fixed the fuck out of #XNUMX!
+Fuck it, YOLO!
+Fucking egotistical bastard. adds expandtab to vimrc
+Fucking submodule bull shit
+Fucking templates.
+Herping the fucking derp right here and now.
+I don't get paid enough for this shit.
+I hate this fucking language.
+I'm too old for this shit!
+It fucking compiles \:D/
+Merge pull request #67 from Lazersmoke/fix-andys-shit Fix andys shit
+My boss forced me to build this feature... Pure shit.
+REALLY FUCKING FIXED
+Revert "fuckup".
+SHIT ===> GOLD
+SOAP is a piece of shit
+Shit code!
+Some shit.
+WHO THE FUCK CAME UP WITH MAKE?
+Why The Fuck?
+a lot of shit
+clarify further the brokenness of C++. why the fuck are we using C++?
+fix some fucking errors
+fixed shit that havent been fixed in last commit
+fixing project shit
+fuckup.
+holy shit it's functional
+if you're not using et, fuck off
+include shit
+refuckulated the carbonator
+someday I gonna kill someone for this shit...
+this is how we generate our shit.
+I don't give a damn 'bout my reputation
+arrgghh... damn this thing for not working.
+download half the damn internet to parse a pdf
+fix that damn sign!!!
+hopefully going to get a successful build got damn it
+still trying to render a damn cube