Split generator into separate python module

Author: odinhb

.gitignore

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

commit.py

@@ -1,84 +1,29 @@
 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)
-
-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)
+        found_message = messages.find_by_md5(message_hash)
 
-        message = fill_line(messages[message_hash])
+        if message_hash and not found_message:
+            raise tornado.web.HTTPError(404)
 
-        self.output_message(message, message_hash)
+        if found_message:
+            self.output_message(found_message, message_hash)
+        else:
+            message, generated_message_hash = messages.generate()
+            self.output_message(message, generated_message_hash)
 
     def output_message(self, message, message_hash):
         self.set_header('X-Message-Hash', message_hash)
@@ -129,6 +74,8 @@ 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()

messages.py

@@ -0,0 +1,76 @@
+import os
+import random
+import re
+from typing import Dict, List
+from hashlib import md5
+
+def generate():
+    digest = _pick_random_key(templates)
+    msg = templates[digest]
+    return (msg, digest)
+
+def find_by_md5(md5):
+    if md5 not in templates:
+        return None
+    else:
+        t = templates[md5]
+        return _fill_template(t)
+
+this_file = os.path.dirname(__file__)
+humans_file = os.path.join(this_file, 'static', 'humans.txt')
+all_messages_file = os.path.join(this_file, 'commit_messages.txt')
+tmp = os.path.join(this_file, 'tmp')
+os.makedirs(tmp, exist_ok=True)
+
+templates: Dict[str, str] = {}
+names: List[str] = []
+
+# Create a hash table of all commit message templates
+print("hashing messages...")
+with open(all_messages_file, 'r', encoding='utf-8') as f:
+    for line in f:
+        templates[md5(line.encode('utf-8')).hexdigest()] = line
+
+with open(humans_file, 'r', encoding='utf-8') as f:
+    for line in f:
+        if "Name:" in line:
+            data = line[6:].rstrip()
+            if data.find("github:") == 0:
+                names.append(data[7:])
+            else:
+                names.append(data.split(" ")[0])
+
+def _pick_random_key(templates):
+    return random.choice(list(templates.keys()))
+
+num_re = re.compile(r"XNUM([0-9,]*)X")
+
+def _fill_template(txt):
+    txt = txt.replace('XNAMEX', random.choice(names))
+    txt = txt.replace('XUPPERNAMEX', random.choice(names).upper())
+    txt = txt.replace('XLOWERNAMEX', random.choice(names).lower())
+
+    nums = num_re.findall(txt)
+
+    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)
+        txt = num_re.sub(str(randint), txt, count=1)
+
+    return txt