iChat archiving

On 10/19/2004 1:14 PM, "Sander Tekelenburg" <tekelenbeuronet.nl> wrote:
> The more I use iChat, the more I run into the problem that, unlike email,
> it's hard to quickly find something in my iChat archive.
>
> How do people deal with this (other than not using chat)? I do have iChat
> autosave chats, but haven't found a mechnism to quickly and effectively
> search through them - quite the contrary. If they were plain text it would be
> easy. Perhaps someone knows of a tool to convert these files to plain text?

Logorrhea

<http://spiny.com/logorrhea/>

I reviewed it for Macworld last year -- it's really a stellar app for doing
exactly what you want.

<http://www.macworld.com/2003/10/reviews/macgems/>

> The more I use iChat, the more I run into the problem that, unlike email,
> it's hard to quickly find something in my iChat archive.

I use 'iChat Transcript Manager'. It's a bit rough and could use some
improvements (such as double-clicking on a chat to open it in iChat),
but does the job of finding text in chats well. The only long-term
disadvantage is that it's not being actively developed, but it works
now and maybe that would change if enough interest was shown. There
are a couple of other similar programmes as well, but after trying
them I settled on iCTM. Do a search on MacUpdate for 'iChat' and see
what comes up to find the others.

<http://www.unsanity.org/archives/000149.php>
<http://www.macupdate.com/>

Cheers,
Jolin

_________________
=> Jolin Warren, Edinburgh, Scotland

At 13:36 -0700 UTC, on 2004/10/19, Dan Frakes wrote:

> Logorrhea

Thanks. Looks nice and it's definitely a start. But I find it a bit too
limited still. Can't change the text size/colour of chats for instance. I
also would like to be able to browse not just by buddy, but by date, or by
length of the chat.

I see Logorrhea has an export command though. At least that strips the bloody
HTML from the chats. The format of the output isn't yet clear to me, but it
looks like this might allow relatively easy parsing so you can send it to a
database, from where you could have much more control over it. I'll have to
look into that.

Such parsing and importing into a database in itself probably won't be too
hard, but of course then you'll want it in a database that offers just the
right searchability :) I suppose a nice interface would be something like
iTunes: a row of columns that you can use to sort, and a 'live filtering'
function (preferably including the ability to save search criteria, so a
search gets updated automatically as more chats are added.)

--
Sander Tekelenburg, <http://www.euronet.nl/~tekelenb/>

At 1:14 PM -0700 2004/10/19, Sander Tekelenburg wrote:
>The more I use iChat, the more I run into the problem that, unlike email,
>it's hard to quickly find something in my iChat archive.
>
>How do people deal with this (other than not using chat)? I do have iChat
>autosave chats, but haven't found a mechnism to quickly and effectively
>search through them - quite the contrary. If they were plain text it would be
>easy. Perhaps someone knows of a tool to convert these files to plain text?

As you probably know, the content of the .chat files includes
the plaintext, so strings, grep, and open work.

Agreed, it's nowhere near as nice as searching in an email
client, though.

--
Chris Pepper: <http://www.reppep.com/~pepper/>
Rockefeller University: <http://www.rockefeller.edu/>

> Logorrhea
>
> <http://spiny.com/logorrhea/>
>
> I reviewed it for Macworld last year -- it's really a stellar app for doing
> exactly what you want.

I like Logorrhea a lot too; it was Dan's article that first turned me on to
it!

Regarding those iChat logs... Will there ever be too many of them? Given the
number of different chats I do each day, I can see my iChats folder quickly
filling up with tens of thousands of iChat logs. I'm at 3,000 now.

It would be great if Logorrhea would allow you to delete chat logs too --
surely I don't need all of these, but I can never bring myself to dig into
that folder to manage its contents, nor should I really have to when it
makes perfect sense for a program like Logorrhea to handle that.

Khoi

work: www.behaviordesign.com
play: www.subtraction.com

At 22:08 -0400 UTC, on 2004/10/20, Chris Pepper wrote:

[...]

> As you probably know, the content of the .chat files includes
> the plaintext, so strings, grep, and open work.

Ah! I wasn't aware of the existence of "strings" :)

But how would you use strings to search an entire directory? It seems to only
accept files.

I could use a repeat loop in AS to go through all the files, but it would be
very slow to use a multitude of "do shell script" calls... If strings can't
be directed to traverse an entire directory, a shell script would b eneeded
to do this fast enough. And I have no experience with shell scripting... :-/

--
Sander Tekelenburg, <http://www.euronet.nl/~tekelenb/>

At 5:35 PM +0200 2004/10/21, Sander Tekelenburg wrote:
>At 22:08 -0400 UTC, on 2004/10/20, Chris Pepper wrote:
>
>[...]
>
>> As you probably know, the content of the .chat files includes
>> the plaintext, so strings, grep, and open work.
>
>Ah! I wasn't aware of the existence of "strings" :)
>
>But how would you use strings to search an entire directory? It seems to only
>accept files.
>
>I could use a repeat loop in AS to go through all the files, but it would be
>very slow to use a multitude of "do shell script" calls... If strings can't
>be directed to traverse an entire directory, a shell script would b eneeded
>to do this fast enough. And I have no experience with shell scripting... :-/

Sander,

        You can use "grep -L" to find matching files, then use
strings or the open command on the returned files. If you do this
frequently, it's probably worth writing a one-line shell function
(like an alias, but it can embed arguments into the middle of the
command) involving grep, xargs, and either open or strings.

                                                Chris
--
Chris Pepper: <http://www.reppep.com/~pepper/>
Rockefeller University: <http://www.rockefeller.edu/>

>At 1:14 PM -0700 2004/10/19, Sander Tekelenburg wrote: The more I use
>iChat, the more I run into the problem that, unlike email, it's hard
>to quickly find something in my iChat archive.
>
>How do people deal with this (other than not using chat)? I do have
>iChat autosave chats, but haven't found a mechnism to quickly and
>effectively search through them - quite the contrary. If they were
>plain text it would be easy. Perhaps someone knows of a tool to
>convert these files to plain text?

I switched yesterday to Adium exactly for this reason. I wanted to open
my iChat archives in BBEdit to do a search. With Adium, you can choose
to keep your iChat logs (as Adium call them) in plain text.
<http://www.adiumx.com/index.php>

You can then retrieve these logs in ~/Library/Applications Support/Adium
2.0/Users/Default/Logs. What is nice is that Adium creates a folder for
each person you chat with.

Also Adium comes with it own log viewer (Windows->Log Viewer or cmd-L)
that I prefer over Logorrhea.

Although, I find that Adium has too much bell and whistles for my taste,
it looks like a nice app and I'm happy with it after one day of usage.

HTH

Cheers
-Emmanuel
--
Emmanuel Décarie / Programmation pour le Web - Programming for the Web
<http://scriptdigital.com/> - Blog: <http://blog.scriptdigital.com> - AIM: scriptdigital

Posted by tekelenb on 06:55am Oct 21, 2004

At 13:36 -0700 UTC, on 2004/10/19, Dan Frakes wrote:

> Logorrhea

Thanks. Looks nice and it's definitely a start. But I find it a bit too limited still. Can't change the text size/colour of chats for instance. I also would like to be able to browse not just by buddy, but by date, or by length of the chat.

Hi -- the co-author of Logorrhea here. As it happens, Dan recently sent me a list of feature requests, including yours. I will definitely try to get to them fairly soon.

I see Logorrhea has an export command though. At least that strips the bloody HTML from the chats. The format of the output isn't yet clear to me, but it looks like this might allow relatively easy parsing so you can send it to a database, from where you could have much more control over it. I'll have to look into that.

The format is kind of plain text. The source for the previous dot release is available if you're interested in figuring out how it works or rolling your own export. Once I stop being lazy, I'll also release the latest source.

Such parsing and importing into a database in itself probably won't be too hard, but of course then you'll want it in a database that offers just the right searchability :) I suppose a nice interface would be something like iTunes: a row of columns that you can use to sort, and a 'live filtering' function (preferably including the ability to save search criteria, so a search gets updated automatically as more chats are added.)

Live filtering is a good idea -- I may look into this.

Let me know if you have any other thoughts -- definitely interested in improving Logorrhea, although it's hard to find the time to invest in a free product. ;)

- Ladd

At 11:54 -0400 UTC, on 2004/10/21, Chris Pepper wrote:

> At 5:35 PM +0200 2004/10/21, Sander Tekelenburg wrote:

[...]

>>But how would you use strings to search an entire directory? It seems to only
>>accept files.

[...]

> You can use "grep -L" to find matching files, then use
> strings or the open command on the returned files. If you do this
> frequently, it's probably worth writing a one-line shell function
> (like an alias, but it can embed arguments into the middle of the
> command) involving grep, xargs, and either open or strings.

Sorry. I suppose I don't understand how to use grep -L. If you could you give
an example that would be great :)

Doing it for a single file is clear:

$ strings [file] | grep -i [query]

But as far as I can see you'd need to do that in a loop, feeding it each file
inside ~/Documents/iChats. Perhaps I should learn how to write such a loop in
a (bash) shell script. I know how to do it in AS of course, but then I'd need
to use a do shell script call for each file, which would make it terribly
slow.

--
Sander Tekelenburg, <http://www.euronet.nl/~tekelenb/>

At 5:22 PM +0200 2004/10/23, Sander Tekelenburg wrote:
>At 11:54 -0400 UTC, on 2004/10/21, Chris Pepper wrote:
>
>> At 5:35 PM +0200 2004/10/21, Sander Tekelenburg wrote:
>
>[...]
>
>>>But how would you use strings to search an entire directory? It
>>>seems to only
>>>accept files.
>
>[...]
>
>> You can use "grep -L" to find matching files, then use
>> strings or the open command on the returned files. If you do this
>> frequently, it's probably worth writing a one-line shell function
>> (like an alias, but it can embed arguments into the middle of the
>> command) involving grep, xargs, and either open or strings.
>
>Sorry. I suppose I don't understand how to use grep -L. If you could you give
>an example that would be great :)
>
>Doing it for a single file is clear:
>
>$ strings [file] | grep -i [query]
>
>But as far as I can see you'd need to do that in a loop, feeding it each file
>inside ~/Documents/iChats. Perhaps I should learn how to write such a loop in
>a (bash) shell script. I know how to do it in AS of course, but then I'd need
>to use a do shell script call for each file, which would make it terribly
>slow.

        No, you can start directly with grep to find the matching
files (it doesn't need strings), so I'd say:

open `grep -iL sander ~/Documents/iChats/*chat`

        Except that iChat filenames contain spaces, so you need to
use xargs (which I've never used) to quote them before open or
strings can handle the arguments properly.

        If you're not going to open them directly, you could use
strings or grep to find the interesting lines in the relevant files...

                                                Chris
--
Chris Pepper: <http://www.reppep.com/~pepper/>
Rockefeller University: <http://www.rockefeller.edu/>

On Tue, 19 Oct 2004 13:14:56 -0700, Sander Tekelenburg
<tekelenbeuronet.nl> wrote:
> How do people deal with this (other than not using chat)? I do have iChat
> autosave chats, but haven't found a mechnism to quickly and effectively
> search through them - quite the contrary. If they were plain text it would be
> easy. Perhaps someone knows of a tool to convert these files to plain text?

I don't have a GOOD method, but I do have a method that works somewhat:

% cd Documents/iChats/
% strings * | grep "what I am looking for" | more
% strings * | grep -l "what I am looking for" | more

the first strings line searches for the text, the second tells me the
name of the file(s) that contain the text. Efficeint? No, but it is
fast. I wish the ichat logs were simple XML. Maybe in Tiger?

I did this earlier today to find a conversation I had in February
about rounded IDE cables.

of course, if I know the conversation was with "Fred Mertz" I can do

% strings Mertz* | grep -l "what I am looking for" | more

--
 ::::== <http://2blog.kreme.com> ==::::

On Mon, 25 Oct 2004 14:47:10 -0700, Khoi Vinh <publicsubtraction.com> wrote:
> Regarding those iChat logs... Will there ever be too many of them? Given the
> number of different chats I do each day, I can see my iChats folder quickly
> filling up with tens of thousands of iChat logs. I'm at 3,000 now.

I would rather have the option to combine chats. I have several
hundred for each of about 6 people, and then a few hundred other
"misc" logs. I would like to be able to take 300 logs and combine
them into a single file, even if it were just a txt file (XML would be
better though).

I haven't looked at the iChat log format to closely.

--
 ::::== <http://2blog.kreme.com> ==::::
 :: Don't get saucy with me, Bernaise ::

On Mon, 25 Oct 2004 14:52:21 -0700, Chris Pepper <pepperreppep.com> wrote:
> At 5:22 PM +0200 2004/10/23, Sander Tekelenburg wrote:
> >At 11:54 -0400 UTC, on 2004/10/21, Chris Pepper wrote:
> >> At 5:35 PM +0200 2004/10/21, Sander Tekelenburg wrote:
> No, you can start directly with grep to find the matching
> files (it doesn't need strings), so I'd say:
>
> open `grep -iL sander ~/Documents/iChats/*chat`

I prefer to see the grep results first, so I can pick out the right
log file and not spawn 100 chats onto my screen.

strings ~/Documents/iChats/* | grep -i sander | more

THEN I do

strings ~/Documents/iChats/* | grep -iL "the sander phrase I was looking for"

to get the name of the file

This is very quick with the command line utilizing up arrows and
control-w and command-C/V

--
 ::::== <http://2blog.kreme.com> ==::::
 :: Don't get saucy with me, Bernaise ::

I use and find quite enough for my needs, a perl script posted on
osxhints. Find it by search www.osxhints.com for "A simple Perl script
to search iChat logs". It using a shell command called 'strings' that
attempts to find text strings in binary files. Its free and I've found
it works quite well.

HTH