Implement a new "/help me" command that suggests something to try

Thanks to Acme for the original "Have you tried" list as obtained from:

    http://web.mit.edu/puzzle/www/resources.html

as:

    http://web.mit.edu/puzzle/www/resources/haveyoutried.pdf

diff --git a/turbot/have_you_tried.py b/turbot/have_you_tried.py
new file mode 100644 (file)
index 0000000..4f0619b
--- /dev/null
+++ b/turbot/have_you_tried.py
@@ -0,0 +1,105 @@
+import random
+
+things_to_try = [
+
+    # For starters
+    "reading the title and the blurb",
+    "typing it into a search engine",
+    ("pursuing even a train of thought that doesn’t seem " +
+     "to give a cogent answer at first"),
+    "looking at the list of helpful websites",
+    "looking at the binder of useful stuff",
+
+    # Letters and words
+    "alphabetizing",
+    "using the leftover letters to spell something",
+    "rearranging the letters (aka “anagramming” or “transposing”)",
+    "looking for unusual letter frequencies",
+    "determining if it is a Rot 13/Caesar shift in general",
+    "shifting from letters to numbers",
+    "treating letters as protein sequences",
+    "treating letters as call letters",
+    "turning it into a cryptic crossword clue",
+    "seeing if these words relate to song lyrics",
+    "seeing if there are any acronyms of phrases",
+    ("diagonalizing (taking the first letter of the first answer, " +
+     "the second letter of the second. . .)"),
+    "looking at a computer keyboard",
+    "a pun",
+
+    # Numbers
+    "graphing it functionally",
+    "graphing it parametrically",
+    "checking to see if your numbers are in the right base (including hex)",
+    "shifting from numbers to letters",
+    "using it as a phone number",
+    "using it as an IP address",
+    "matching numbers to MIT buildings",
+    "matching numbers to MIT courses",
+    "treating numbers as dates",
+    "treating numbers as atomic weights, numbers, etc.",
+    "treating numbers as latitude/longitude or GPS coordinates",
+    "treating numbers as radio stations",
+    "treating numbers as ASCII numbers",
+    "treating numbers as PLU numbers",
+    "treating numbers as ISBN numbers",
+    "seeing if there are any strange sequences",
+    "seeing if prime numbers are involved",
+    "seeing if fundamental constants are involved",
+    "asking what other numbers are close to the one you have",
+
+    # Thirteen ways of looking at a puzzle
+    "looking at it in a mirror",
+    "squinting at it from far away",
+    "tilting it",
+    "looking at it upside down",
+    "looking through it",
+    "rewriting it neatly",
+    "rewriting it on graph paper",
+    "saying it out loud to someone else",
+    "putting yourself in the constructor’s shoes",
+    "making it 3D",
+    "cutting it up",
+    "folding it",
+    "connecting the dots",
+
+    # What is it",
+    "looking at a song/poem/book/movie/TV show",
+    "Braille",
+    "Morse code",
+    "overlaying it on a map of MIT",
+    "getting on the T",
+    "checking it for pop culture references with Tanis, Jenn, Denis or James",
+    "using the Library of Congress or Dewey decimal system",
+    ("asking whether it has anything to do personally with the " +
+     "constructing team"),
+    "Palm graffiti",
+    "running it by someone who goes to MIT",
+
+    # Zen
+    "trusting your instincts",
+    "trusting someone else’s instincts",
+    "asking “what’s weird about this?”",
+    "taking a step back",
+    "asking for fresh brains",
+    "checking your work",
+    "asking someone else to check your work",
+    "explaining your work to someone else",
+    "doing what you’ve already done again to the output",
+    "consulting the list of expertise",
+    "leaving the room",
+    "asking people if it looks like anything they recognize",
+    "thinking about what’s missing",
+    "asking yourself whether you’ve used all the information",
+    "keeping even a strange-looking result",
+    "brute force",
+    "not thinking about it",
+    "splitting up tasks to solve in a group",
+    "talking to a hunt veteran",
+    "talking to a hunt newbie",
+    "asking “what’s the pattern”",
+    "rereading the instructions",
+]
+
+def have_you_tried():
+    return "Have you tried {}?".format(random.choice(things_to_try))

diff --git a/turbot/help.py b/turbot/help.py
index dbecfe36be3a5251c165fa5c9d8b8a7063a8df72..9a4d9fdbbf51494ede1a8cf99e2d7cc4b25790b6 100644 (file)
--- a/turbot/help.py
+++ b/turbot/help.py
@@ -25,6 +25,7 @@ Type `/help <command-name>` for more details on any command below.
     `/puzzle`: Display state of current puzzle (puzzle channel)
 
 *Solving support* (from any channel)
+    `/help me`: Get a suggestion of something to try
     `/rot [*|COUNT] <text_to_rotate>`: Perform a Caesar shift on text
 """
 

diff --git a/turbot/interaction.py b/turbot/interaction.py
index 53f45a701a9fb9bae68d41316af01fe29f518a14..f784a0266203862711d4c20f16656195f712a09f 100644 (file)
--- a/turbot/interaction.py
+++ b/turbot/interaction.py
@@ -18,6 +18,7 @@ from turbot.puzzle import (
 )
 from turbot.round import round_quoted_puzzles_titles_answers
 from turbot.help import turbot_help
+from turbot.have_you_tried import have_you_tried
 import turbot.rot
 import turbot.sheets
 import turbot.slack
@@ -1334,21 +1335,34 @@ def help_command(turb, body, args):
     Displays help on how to use Turbot.
     """
 
+    channel_name = body['channel_name'][0]
     channel_id = body['channel_id'][0]
     response_url = body['response_url'][0]
 
+    # Process "/help me" first. It calls out to have_you_tried rather
+    # than going through our help system.
+    #
+    # Also, it reports in the current channel, (where all other help
+    # output is reported privately to the invoking user).
+    if args == "me":
+        to_try = have_you_tried()
+
+        # If this is a direct message then there's not a usable channel_id
+        # and we have to use the response_url instead
+        if channel_name == "directmessage":
+            requests.post(response_url,
+                          json = {"text": to_try},
+                          headers = {"Content-type": "application/json"})
+        else:
+            turb.slack_client.chat_postMessage(
+                channel=channel_id, text=to_try)
+        return lambda_ok
+
     help_string = turbot_help(args)
 
-    # The "/help me" command is special in that it reports in the
-    # current channel, (where all other commands report privately to
-    # the invoking user).
-    if args == "me":
-        turb.slack_client.chat_postMessage(
-            channel=channel_id, text=help_string)
-    else:
-        requests.post(response_url,
-                      json = {"text": help_string},
-                      headers = {"Content-type": "application/json"})
+    requests.post(response_url,
+                  json = {"text": help_string},
+                  headers = {"Content-type": "application/json"})
 
     return lambda_ok