Charcoal HQ

12:39 AM

in Tavern on the Meta on Meta Stack Exchange Chat, 18 hours ago, by Infinite Recursion

@Undo @bjb568, please collaborate and reduce redundant bot reports ^

Blaze

[Blaze] answer flagged by Undo: stackoverflow.com/a/27882709

[Blaze] answer flagged by Undo: stackoverflow.com/a/27882707

[Blaze] answer flagged by Undo: stackoverflow.com/a/27882637

[Blaze] answer flagged by Undo: stackoverflow.com/a/27882750

1 hour later…

1:54 AM

Hmm, got a few rude/not constructive comments.

Looks like I'm going to have to manually do this.

1 hour later…

3:12 AM

@hichris123 Classification?

@Andy Yeah. The usual naughty words aren't really helping too much.

@hichris123 Don't rush it. It took me a while for my stuff. Can you split them into broad categories or can you classify some of them using certain patterns?

@Andy I'm just thinking about broad categories currently. The problem is most seem to be good comments, and it's hard to find the bad ones.

[ SmokeDetector ] All-caps title: BAJOTZ ARTWORK [WATTPAD67] by Bajotz .X. on android.stackexchange.com

Most are good comments, from my experience. You have to go through a large number of them for find the bad ones

7 hours later…

10:23 AM

[ SmokeDetector ] All-caps title: DETERMINANT OF SKEW SYMMETRIC MATRIX OF ODD ORDER IS ZERO. HOW by MMS on math.stackexchange.com

3 hours later…

ProgramFOX

1:50 PM

!!/alive?

@ProgramFOX Of course

8 hours later…

9:43 PM

@hichris, what fields are you saving when retrieving comments?

@Andy Id, OwnerId, OwnerReputation, Score, CreationDate, IsEdited, PostId, Body (in Markdown)

@hichris123 Why isEdited?

@Undo shrugs Just cause?

@hichris123 link too by any chance?

@Andy you can derive the link from the Id

@hichris123 :D

9:45 PM

@Andy No but I'm manually constructing... ^

Undo is too quick for me.

I practice my chat skills daily.

@Undo claps politely

bows

If you are classifying rude comments, @hichris, would it be possible to get a dump those those that you marked rude? Preferable ones that were then flagged and removed too?

I'm guessing they are a very small percentage of what you've seen though, right?

It's /posts/comments, by the way.

@Andy Sure. It's a grand total of like 40 right now.

... or 20 ...

Do you want them now or when there's more of them?

9:54 PM

When there is more, please

Ok.

Going through Andy's dump

(thanks for that @Andy)

You're welcome

Just need a way to make it readable :P

Have you found some useful patterns?

9:57 PM

not yet

I can make it pretty next export

nah, I'm just going to feed it through Python

But, it's a giant list of dictionaries. You can make it pretty too :)

Yep :D

stackoverflow.com/questions/27892540/what-happend-to-this-site - mind giving me a screenshot, @Andy? I'm curious now.

10:00 PM

so. close.

I can't believe my Pentium 4 w/ 2 GB of RAM can handle all of this. :D

@Andy have a ballpark estimate on how many posts are in the dump? Want to make sure I'm parsing it right

Does 999 sound right?

Okay, ran Torch on that dump

67 x www.optimalstackfacts.org
62 x www.maletestosteronebooster.org
51 x www.x4facts.com
25 x www.tripleeffectseyeserum.org
24 x www.healthcaresup.com
17 x www.garciniacambogiaprofacts.com
17 x www.lxwpro-t.com
16 x www.thehealthyadvise.com
13 x www.filerepairforum.com
11 x dietandhealthguide.com
11 x www.supracleanse350facts.org
10 x hyperglycemiaabout.com
9 x sites.google.com
9 x www.idolizeadvancedeyeserumfacts.com
8 x diabacordoesitwork.com
8 x testeronexlhelp.com
7 x droz-garciniacambogia.org

Once we get post body blacklists working we can add those

10:36 PM

@hichris123 i.imgur.com/7Ns3WWE.jpg

lol thanks.

@Undo Probably. I have 1014 currently, so 999 isn't that far off considering it's slightly older

testing smokey in a minute

!!/alive?

@Undo You broke it!

yup :P

actually it's not started yet :P

10:51 PM

[ SmokeDetector ] SmokeDetector started in reverted mode at rev 7353961 (Undo1: Merge branch 'bodies' of github.com/Charcoal-SE/SmokeDetector into bodies) (hosted by Undo)

In theory this should drive us nuts

I added a... special regex :P

[ SmokeDetector ] body contains 'the': Is $\left(\sum {{x^n}\over{n!}} \right) \left(\sum {{y^n}\over{n!}} \right) = \left(\sum {{(x+y)^n}\over{n!}}\right)$ generalizable for series? by user507974 on math.stackexchange.com

... merge branch bodies into bodies?!

[ SmokeDetector ] body contains 'the': Energy source for a generation ship by celtschk on worldbuilding.stackexchange.com

[ SmokeDetector ] body contains 'the': For example: “I am looking for soaps” or “I am looking for soap”? which one is more correct? by Saeid on english.stackexchange.com