From avasik at gmail.com Fri Jul 1 12:27:57 2005 From: avasik at gmail.com (Васик Алексей) Date: Fri, 1 Jul 2005 14:27:57 +0400 Subject: [spambayes-dev] Compilation & translation quest. Message-ID: <1749196073.20050701142757@gmail.com> Sorry for my bad English, it is not my native language. Yesterday I've just finished translate an Outlook addin to RUSSIA lang. But! There is a big problem: For the first, I try to compile a "clean dist" sources of 1.1a1 Installed tools are py2exe-0.5.4.win32-py2.4.exe python-2.4.1.msi pywin32-204.win32-py2.4.exe ResourcePackage-1.0.0.win32.exe InnoSetup (fresh) Being guided by the text of the readme-devel.txt -- o Run the setup.py file in the spambayes/Outlook2000/docs directory to generate the dynamic documentation. Ran it o Run sb_server and open the web interface. This gets resourcepackage to generate the needed files. Yes o Replace the __init__.py file in spambayes/spambayes/resources with a blank file to disable resourcepackage. Empty o Ensure that the version numbers in spambayes/spambayes/__init__.py and spambayes/spambayes/Version.py are up-to-date. Don' touch o Ensure that you don't have any other copies of spambayes in your PYTHONPATH, or py2exe will pick these up! If in doubt, run setup.py install. Yes o Run the "setup_all.py" script in the spambayes/windows/py2exe/ directory. This uses py2exe to create the files that Inno will install. The result is : [skipped] creating C:\SPAMBAYES\windows\py2exe\build\bdist.win32 creating C:\SPAMBAYES\windows\py2exe\build\bdist.win32\winexe creating C:\SPAMBAYES\windows\py2exe\build\bdist.win32\winexe\collect creating C:\SPAMBAYES\windows\py2exe\build\bdist.win32\winexe\temp creating C:\SPAMBAYES\windows\py2exe\dist creating C:\SPAMBAYES\windows\py2exe\dist\lib creating C:\SPAMBAYES\windows\py2exe\dist\bin *** generate typelib stubs *** creating C:\SPAMBAYES\windows\py2exe\build\bdist.win32\winexe\temp\win32com creating C:\SPAMBAYES\windows\py2exe\build\bdist.win32\winexe\temp\win32com\gen_py Traceback (most recent call last): File "C:\SPAMBAYES\windows\py2exe\setup_all.py", line 187, in ? zipfile = "lib/spambayes.modules", File "C:\Python24\lib\distutils\core.py", line 149, in setup dist.run_commands() File "C:\Python24\lib\distutils\dist.py", line 946, in run_commands self.run_command(cmd) File "C:\Python24\lib\distutils\dist.py", line 966, in run_command cmd_obj.run() File "C:\Python24\Lib\site-packages\py2exe\build_exe.py", line 183, in run self._run() File "C:\Python24\Lib\site-packages\py2exe\build_exe.py", line 214, in _run self.typelibs) File "C:\Python24\Lib\site-packages\py2exe\build_exe.py", line 1254, in collect_win32com_genpy mod = gencache.GetModuleForTypelib(*info) File "C:\Python24\Lib\site-packages\win32com\client\gencache.py", line 258, in GetModuleForTypelib mod = _GetModule(modName) File "C:\Python24\Lib\site-packages\win32com\client\gencache.py", line 629, in _GetModule mod = __import__(mod_name) ImportError: No module named 00062FFF-0000-0000-C000-000000000046x0x9x0 So I got only iss file without any needed files in py2exe\dist\ Whats the problem ? I think there's my mistakes, but where I mistake ? Thanx for the GREAT WORK for ALL TEAM MEMBERS! Russia stay with you :) Alexei From kenny.pitt at gmail.com Fri Jul 1 15:44:53 2005 From: kenny.pitt at gmail.com (Kenny Pitt) Date: Fri, 1 Jul 2005 09:44:53 -0400 Subject: [spambayes-dev] Compilation & translation quest. In-Reply-To: <1749196073.20050701142757@gmail.com> Message-ID: <42c548d8.5832928b.0471.78b7@mx.gmail.com> ????? ??????? wrote: > "C:\Python24\Lib\site-packages\win32com\client\gencache.py", line > 629, in _GetModule mod = __import__(mod_name) > ImportError: No module named > 00062FFF-0000-0000-C000-000000000046x0x9x0 > > So I got only iss file without any needed files in py2exe\dist\ > > Whats the problem ? I think there's my mistakes, but where I mistake ? You have to build on a system with Outlook 2000 installed. The error is coming when trying to import the COM type libraries for integrating into Outlook. The build will not work if you only have Outlook XP or Outlook 2003. In order for SpamBayes to work with Outlook 2000 and higher, we have to build to the lowest supported version. You can, however, get it working on your system by making a couple of local changes to your setup_all.py script. Starting on line 47 in the version I have there is a section that begins with "typelib = [". If you are using Outlook 2003 then make the following change to the 2 lines that follow the "typelib =": ('{00062FFF-0000-0000-C000-000000000046}', 0, 9, 0), ('{2DF8D04C-5BFA-101B-BDE5-00AA0044DE52}', 0, 2, 1), Change to: ('{00062FFF-0000-0000-C000-000000000046}', 0, 9, 2), ('{2DF8D04C-5BFA-101B-BDE5-00AA0044DE52}', 0, 2, 2), Notice that only the version numbers at the end change. If you are using Outlook XP, the "9, 2" in the first line should be "9, 1" instead. I can't remember if the second line is "2, 1" or "2, 2" for XP so just try both. -- Kenny Pitt From TronixSoft at comcast.net Sat Jul 2 09:48:37 2005 From: TronixSoft at comcast.net (TronixSoft) Date: Sat, 2 Jul 2005 00:48:37 -0700 Subject: [spambayes-dev] web service Message-ID: <20050702074846.A24D11E4002@bag.python.org> Hi, I'm writing spam filtering service for the "free" email services bundled with Windows server 2003. (Which consist of just SMTP and pop3 services.) I would like to leverage the client side spambayes database to provide server side filtering. The approach I would like to take is to leave spambayes intact on the client and leverage its (superb) efforts, and have the (email) server updated with the results. Here is what the server will do: Receive email Check for spam based on the recipient's individual (spambayes) database If spam, take preconfigured action (move, delete, tag, etc.) If not spam, cache the results for a preconfigured amount of time (say, a day) and wait for a ham/spam report from the client If a ham/spam report arrives, update the database accordingly If no report arrives within timeout period, take a preconfigured course of action (update db based on current results, assume good/bad, ignore, etc.) Forward email to the next hop as appropriate Listen for updates from spambayes clients via web service Here is what the client would do, without an update to spambayes Receive email Allow spambayes to process Forward results (clues?) for the current email to the server; or send copy of db to server and let server digest the entire db; each via additional Outlook plug-in -or- Forward db to sever as a manual/calendar update (i.e. not on arrival of each email) Here is what the client could do with an update to spambayes (presumptuous, I know and apologize) Receive email Process Connect to server via web services (xml/soap) and upload results for the current email Yes, this would mean a separate db per user on the server which I think is a good thing, but may cause scalability issues. However, I would like to get started 'somewhere'. I also realize that I could forgo the web services and just have the xml package emailed to a service account inbox on the server, but I'm not too keen on the idea of automating the sending of email from Outlook or otherwise. Also, by using a web service, this would leave the door open to many options and features. One would be a single interface by which a user can update preferences on the server, regardless of the client (web page, binary, from within spambayes itself). My question is: Can someone help me with a utility to export or convert the spambayes db to text? The reason I'm asking for help is due to a desire to code this in VB.NET. I took some classes in C many years ago, but ended up spending all my time "growing up" with VB. I could learn python and figure this out, but I'm hoping to focus on the areas I know best. Not to step on anyone's pickle, but how about an option in spambayes to choose whether to use a db, or xml for storage. Drive space is cheap, I wouldn't mind a larger database if it meant that I could parse it from another application (OS, etc.) XML would make sharing the db a lot easier. Just a thought. By the way, I may even develop a script version whereby the xml package can be copied to a share or emailed to a special inbox on the server that a script will monitor, digest and update config/db files for each user accordingly. Currently I'm using a simple VB script on the server, implementing some WMI calls (cpu friendly way to monitor folder for new files), and a white/black list. Of course I understand this to be very rudimentary. But it is a surprisingly effective (and free) way to handle some of the more obvious spam. The white list is based on email addresses, and the black list is comprised of keywords (such as viagra, etc.) Any thoughts, comments, encouragement, and/or assistance would be greatly appreciated. Thanks, Robert -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mail.python.org/pipermail/spambayes-dev/attachments/20050702/5e27a42b/attachment.htm From tameyer at ihug.co.nz Mon Jul 4 05:03:44 2005 From: tameyer at ihug.co.nz (Tony Meyer) Date: Mon, 4 Jul 2005 15:03:44 +1200 Subject: [spambayes-dev] web service In-Reply-To: Message-ID: > Can someone help me with a utility to export or convert > the spambayes db to text? See the sb_dbexpimp.py script in the source distribution. It will convert a SpamBayes token database in any of the supported forms to/from CSV. > Not to step on anyone's pickle, but how about an option in > spambayes to choose whether to use a db, or xml for storage. We support pickle, various dbms, mysql, postgresql, and ZODB. If you want something else, then the code is pretty simple to write. (See storage.py). > Drive space is cheap, I wouldn't mind a larger database if > it meant that I could parse it from another application > (OS, etc.) You can use any of the existing formats from other applications. If you use pickle, then that probably restricts you to programs written in Python (although it's possible there's a C library to read Python's pickle format, and you could do it via the Python source anyway). For dbm, look at (e.g) libdbm. SQL is SQL. You're more likely to use ZODB if you're using Python, but again, you can use it from elsewhere. > XML would make sharing the db a lot easier. Just a thought. XML would be hideous overkill for this sort of thing. "word1034" is excessively verbose. =Tony.Meyer From mhammond at skippinet.com.au Tue Jul 5 05:09:10 2005 From: mhammond at skippinet.com.au (Mark Hammond) Date: Tue, 5 Jul 2005 13:09:10 +1000 Subject: [spambayes-dev] Compilation & translation quest. In-Reply-To: <1749196073.20050701142757@gmail.com> Message-ID: <0cdf01c5810e$ea749000$090a0a0a@enfoldsystems.local> > The result is : > [skipped] > creating C:\SPAMBAYES\windows\py2exe\build\bdist.win32 > creating C:\SPAMBAYES\windows\py2exe\build\bdist.win32\winexe > creating C:\SPAMBAYES\windows\py2exe\build\bdist.win32\winexe\collect > creating C:\SPAMBAYES\windows\py2exe\build\bdist.win32\winexe\temp > creating C:\SPAMBAYES\windows\py2exe\dist > creating C:\SPAMBAYES\windows\py2exe\dist\lib > creating C:\SPAMBAYES\windows\py2exe\dist\bin > *** generate typelib stubs *** > creating > C:\SPAMBAYES\windows\py2exe\build\bdist.win32\winexe\temp\win32com > creating > C:\SPAMBAYES\windows\py2exe\build\bdist.win32\winexe\temp\win3 > 2com\gen_py > Traceback (most recent call last): > File "C:\SPAMBAYES\windows\py2exe\setup_all.py", line 187, in ? > zipfile = "lib/spambayes.modules", > File "C:\Python24\lib\distutils\core.py", line 149, in setup > dist.run_commands() > File "C:\Python24\lib\distutils\dist.py", line 946, in run_commands > self.run_command(cmd) > File "C:\Python24\lib\distutils\dist.py", line 966, in run_command > cmd_obj.run() > File "C:\Python24\Lib\site-packages\py2exe\build_exe.py", > line 183, in run > self._run() > File "C:\Python24\Lib\site-packages\py2exe\build_exe.py", > line 214, in _run > self.typelibs) > File "C:\Python24\Lib\site-packages\py2exe\build_exe.py", > line 1254, in collect_win32com_genpy > mod = gencache.GetModuleForTypelib(*info) > File > "C:\Python24\Lib\site-packages\win32com\client\gencache.py", > line 258, in GetModuleForTypelib > mod = _GetModule(modName) > File > "C:\Python24\Lib\site-packages\win32com\client\gencache.py", > line 629, in _GetModule > mod = __import__(mod_name) > ImportError: No module named > 00062FFF-0000-0000-C000-000000000046x0x9x0 Currently you need Office 2000 installed to build spambayes. This is because we reference the Office 2000 type-libraries, which allows us to work in all office versions 2000 or later. If we referenced the Office XP type-libraries, SpamBayes would no longer work on systems with only office 2000. For the purposes of testing, you could change setup_all.py to reference the typelibs for your version of Office. To do this you will need to run "\Python23\lib\site-packages\win32com\client\makepy.py -i", and select the relevant typelibs from the list. makepy will print the new values you can use for the selected typelibrary. In general, most will be very similar to the originals, but with a different version number. You will also need to update Outlook2000\setup.py with these new values. Again, once you have done that, your version will not work with Office 2000 - but that may be OK if you don't intend distributing the built version. Mark From faber at faberbox.com Tue Jul 5 17:29:01 2005 From: faber at faberbox.com (Faber) Date: Tue, 5 Jul 2005 17:29:01 +0200 Subject: [spambayes-dev] Unicode version of PyMeldLite Message-ID: <200507051729.01333.faber@faberbox.com> Hello guys, I've used your PyMeldLite module for a web application, and I've modified it in order to support Unicode strings (well, let's say utf-8 decoded ones ;) ). Probably it's not the greatest patch I've wrote, but it works for me and could be used as start-point for implement full Unicode support in PyMeldLite. Let me know if you're interested in my patched version, and where can I send it. Regards -- Faber http://www.faberbox.com/ Like music and mathematical equations, computer language is just that, language, and it communicates information either to a computer or to those who can read it...For the purposes of First Amendment analysis, this court finds that source code is speech. -- Judge Marilyn Hall Patel From richie at entrian.com Tue Jul 5 22:20:49 2005 From: richie at entrian.com (Richie Hindle) Date: Tue, 05 Jul 2005 21:20:49 +0100 Subject: [spambayes-dev] Unicode version of PyMeldLite In-Reply-To: <200507051729.01333.faber@faberbox.com> References: <200507051729.01333.faber@faberbox.com> Message-ID: <9nqlc190ucosmu3vvbp9fk6gls3s7dc5kp@4ax.com> Faber, > I've used your PyMeldLite module for a web application, and I've modified it > in order to support Unicode strings (well, let's say utf-8 decoded > ones ;) ). > [...] > Let me know if you're interested in my patched version, and where can I send > it. Yes please! Ideally you'd submit it to the SourceForge patch manager for SpamBayes, at http://sourceforge.net/tracker/?group_id=61702&atid=498105 (follow the "Login and Submit New") link. But you'll need to register with SourceForge to do that - if you don't want to do that, you can email the patch to me. Thanks! -- Richie Hindle richie at entrian.com From BHARGENS at nordco.com Wed Jul 6 18:47:35 2005 From: BHARGENS at nordco.com (Hargens, Becky) Date: Wed, 6 Jul 2005 11:47:35 -0500 Subject: [spambayes-dev] Good Morning Message-ID: <539358A519F9AE429F7243E6F32E789101B72CEA@nordcoex2k.nordco.com> I was on your website FAQ page and couldn't find an answer to several questions: 1. How do you get Spambayes to accept all those in your contact lists as "good", not as "suspect" spam without clicking on one suspect spam at a time? After installing the software, I discovered email I received from people at work was going into the suspect folder. I would have thought the program, during it's initial training phase, would have gone through all the address/contact lists and set those as "good". It didn't, and it's a pain to train each one. 2. I have been trying over and over to train the software to accept an email address as good (ie., not suspect spam). I was clicking on "recover from spam", assuming this was what was telling the program that the email addy was ok. Apparently it isn't. Any suggestions? From Spalding at ntconnections.com.au Sat Jul 9 09:38:01 2005 From: Spalding at ntconnections.com.au (Natalia) Date: Sat, 9 Jul 2005 02:38:01 -0500 Subject: [spambayes-dev] Need software? Click here. Message-ID: <4334104283.8493557587@12-221-119-7.client.insightBB.com> Software distribution. http://nitxw.c1g9ftu59mu1rvu.lhotacg.com Golf is the most fun I've ever had, with my clothes on. Intellectual passion dries out sensuality. From EganBranis at jahitchcock.com Sat Jul 9 23:16:10 2005 From: EganBranis at jahitchcock.com (Branislav Egan) Date: Sat, 9 Jul 2005 16:16:10 -0500 Subject: [spambayes-dev] Works GGood Message-ID: <20050709211619.DB1281E4005@bag.python.org> Hello, with these ladies, one of whom, indeed, had been for a little whilesuch heaps upon the beds of the six Spaniards that by the time sheis on the gun-deck. You'll return to it at once, and take your crewAmazed as he was, his amazement would have been deeper had he beenword for it that no harm should come to you so that no harm came toLord Jeffreys, whose terrible fame had come ahead of him fromthat I must, I'll take none that I needn't. But.... He broke offYou speak excellent Spanish.If the Spaniards had reached it, there would be lights. He knocked,have an angel for his niece? said he recklessly, for he was recklessEnglish seaman's story, disregarding any evidence that might belieye've served me for a hostage. Ye'll be well advised to avoid amust belong to either Cuba or Hispaniola. Now knowing Cuba to lieMr. Nuttall looked wildly this way and that a moment, then boltedHis lordship's smile brought lines like gashes into his leatheryknow ye wouldn't like them. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mail.python.org/pipermail/spambayes-dev/attachments/20050709/4428fdd8/attachment.htm From emil.pedersen at its.uu.se Wed Jul 13 14:49:36 2005 From: emil.pedersen at its.uu.se (Emil Pedersen) Date: Wed, 13 Jul 2005 14:49:36 +0200 Subject: [spambayes-dev] sb_imapfilter or imaplib - another crash on bad date Message-ID: <9E5377B3F440026FDC575E98@emil.its.uu.se> Hi, I've had some imapfilter crashes related to bad dates. From what I can found it resembles of [890645], but appears when saving (appending) the mail. I'm not sure whatever this should be considered a bug or not, or where it's best fixed (in imaplib or spambayes)... Anyway, by catching the execption raised from the append call it works fine (given my setup, spambayes-1.0.3 & python-2.3.5). So here's my "fix": emilp at emil:proj/SPAMBAYES> diff sb_imapfilter.py.ORG sb_imapfilter.py 522,525c522,529 < response = imap.append(self.folder.name, flgs, tme, < self.as_string()) < if response[0] == "OK": < break --- > # Catch exceptions that may occure if msg_time is realy > # bad (I got one like "19-Jan-2038 04:10:47 +0100"). > # Just pass the exceptions, and the loop will try again, > # eventualy with an desent time. > try: > response = imap.append(self.folder.name, flgs, tme, > self.as_string()) > except IMAP4.error: > pass > else: > if response[0] == "OK": > break Should I report this as a bug (including the suggested fix), or can I just post it here and some "real" developer will catch it? I'm not familiar with the normal bug tracking procedure, so please help me out or point me in the right direction. Sincerely yours, Emil From T.A.Meyer at massey.ac.nz Thu Jul 14 03:07:11 2005 From: T.A.Meyer at massey.ac.nz (Meyer, Tony) Date: Thu, 14 Jul 2005 13:07:11 +1200 Subject: [spambayes-dev] sb_imapfilter or imaplib - another crash on bad date Message-ID: (Please excuse the HTML mail; working from a crappy webmail system) > I've had some imapfilter crashes related to bad dates. From what I can > found it resembles of [890645], but appears when saving (appending) the > mail. > > I'm not sure whatever this should be considered a bug or not, or where > it's best fixed (in imaplib or spambayes)... If it really does seem like a bug with imaplib, then that should definitely be reported (open a bug tracker on >) - I recall that we did come across one back when first writing imapfilter, although I don't recall what it was offhand. However, even in that case, we still really want to patch SpamBayes to handle it as well, for two reasons. The first is that (hopefully <0.5 wink>) the Python release cycle is slower that the SpamBayes release cycle. It's likely that a new SpamBayes release (e.g. 1.1a2) will come out before the next Python release (I think 2.3 is done, so that would be 2.4.2, or 2.5). The other reason is that we aim to have SpamBayes work for any versions of Python greater than or equal to 2.2; it's nice not to have to force people to have the most recent versions of something. [...] > Should I report this as a bug (including the suggested fix), or can I > just post it here and some "real" developer will catch it? The first thing to do, if possible, would be to check whether this has already been fixed in CVS. I know that Sjoerd has checked in some imapfilter fixes to do with dates recently, but I don't recall what they were offhand, and don't have easy access to check right now. The easiest way to check is to use SF's ViewCVS (the "CVS" and then "View Reposity" links on >) to look at sb_imapfilter.py. If the problem is already fixed in CVS, then you don't need to do anything. If it's not then ideally either a bug report or patch tracker (link above) for SpamBayes. If it's a bug and you set it to "imapfilter", then it will automatically assign it to me (anadelonbrin) and I'll get to it at some point. Having a patch, rather than just a bug report drastically reduces the amount of time it'll take for me to deal with it. However, reporting it here also stands a reasonable chance of it being noticed and fixed (it probably increases the time, since I'm more likely to review what's waiting for me on the tracker than review spambayes-dev emails that I haven't got to yet). Thanks for the email and the patch (even if the fix is already in CVS) - greatly appreciated! (People to work on imapfilter are particularly thin on the ground). =Tony.Meyer From sesoinc at juno.com Fri Jul 15 09:57:14 2005 From: sesoinc at juno.com (SE SO INC) Date: Fri, 15 Jul 2005 00:57:14 -0700 Subject: [spambayes-dev] Your Front Door Opening Message-ID: An HTML attachment was scrubbed... URL: http://mail.python.org/pipermail/spambayes-dev/attachments/20050715/af5b59ce/attachment.htm From robert.mezzone at pjsolomon.com Tue Jul 19 23:15:56 2005 From: robert.mezzone at pjsolomon.com (Robert Mezzone) Date: Tue, 19 Jul 2005 17:15:56 -0400 Subject: [spambayes-dev] Outlook Plug-In Bug? Message-ID: Hi, We noticed the following problem with version 1.0.3 and 1.0.4 of the Outlook Plug-in All machines are WindowsXP SP2 all latest patches All machines are Office 2003 SP1 all latest patches Scenario User logs into windows User opens Word or any application that supports the Send-To feature User sends an email using the Send-To As Attachment feature. Message is delivered Try to send a second attachment Error "Operation Failed" Look in Task Manager - there are two OUTLOOK.EXE process' running. The first process did not end gracefully when the first message was sent. For some reason the two processes are causing a problem that prevents messages from being delivered from an external application. Disable the Spambayes COM Add-In within outlook and the OUTLOOK.EXE process ends gracefully every time. Scenario 2 User logs into Windows Opens Outlook Opens Word Uses the Send-To As Attachment feature User can send documents all day long provided they don't close Outlook. Every indication points to the Spambayes program causing the Operation Failed error. I haven't been able to reproduce this with 1.1a1 Is this a known bug in the older versions? Thanks. Robert Mezzone -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mail.python.org/pipermail/spambayes-dev/attachments/20050719/0f95ef31/attachment.htm From T.A.Meyer at massey.ac.nz Wed Jul 20 02:58:23 2005 From: T.A.Meyer at massey.ac.nz (Meyer, Tony) Date: Wed, 20 Jul 2005 12:58:23 +1200 Subject: [spambayes-dev] Outlook Plug-In Bug? Message-ID: > User opens Word or any application that supports the Send-To feature [...] I suspect that this is this bug: [ 1103976 ] Outlook.exe will not terminate when launched via mailto My money is on this being an Outlook bug. It's possible we can work around it somehow, of course. Since this only appears to effect Outlook 2003, and I'm using Outlook 2002, it's pretty difficult for me to look into ways around this (not that I have the time right at the moment, anyway). Kenny is assigned the above bug report, and is going to look into it when he has time, I believe (which won't be for a while as well, IIRC). Whether or not this was a problem in earlier versions of the plug-in is an interesting question, to which I don't know the answer. It's possible that it was, especially since the problem doesn't effect old versions of Outlook. If you (or anyone else) wants to try old plug-in versions, I have copies of them all (or could unhide them on sourceforge temporarily). If the problem doesn't occur with old versions, then that would certainly be valuable information for the bug report, since that gives an idea of where to find a solution. =Tony.Meyer From robert.mezzone at pjsolomon.com Wed Jul 20 03:19:04 2005 From: robert.mezzone at pjsolomon.com (Robert Mezzone) Date: Tue, 19 Jul 2005 21:19:04 -0400 Subject: [spambayes-dev] Outlook Plug-In Bug? Message-ID: I'd be more than happy to try older versions. I should have saved my copies. I have to do some further testing but I can't recall having the problem with 1.1a1 and I've been using it for months. I want to try upgrading one of the 1.0.x machines tomorrow to see what happens. I downloaded Process Explorer from sysinterinals. Perhaps I can figure out what is causing the process to not end gracefully. Robert -----Original Message----- From: Meyer, Tony [mailto:T.A.Meyer at massey.ac.nz] Sent: Tue 7/19/2005 8:58 PM To: Robert Mezzone; spambayes-dev at python.org Subject: RE: [spambayes-dev] Outlook Plug-In Bug? > User opens Word or any application that supports the Send-To feature [...] I suspect that this is this bug: [ 1103976 ] Outlook.exe will not terminate when launched via mailto My money is on this being an Outlook bug. It's possible we can work around it somehow, of course. Since this only appears to effect Outlook 2003, and I'm using Outlook 2002, it's pretty difficult for me to look into ways around this (not that I have the time right at the moment, anyway). Kenny is assigned the above bug report, and is going to look into it when he has time, I believe (which won't be for a while as well, IIRC). Whether or not this was a problem in earlier versions of the plug-in is an interesting question, to which I don't know the answer. It's possible that it was, especially since the problem doesn't effect old versions of Outlook. If you (or anyone else) wants to try old plug-in versions, I have copies of them all (or could unhide them on sourceforge temporarily). If the problem doesn't occur with old versions, then that would certainly be valuable information for the bug report, since that gives an idea of where to find a solution. =Tony.Meyer -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mail.python.org/pipermail/spambayes-dev/attachments/20050719/366ded58/attachment.htm From robert.mezzone at pjsolomon.com Wed Jul 20 17:36:51 2005 From: robert.mezzone at pjsolomon.com (Robert Mezzone) Date: Wed, 20 Jul 2005 11:36:51 -0400 Subject: [spambayes-dev] Outlook Plug-In Bug? Message-ID: The following were tested on a PC running WINXP SP2 all patches and Office 2003 SP1 all patches. I installed a version and tested. Uninstalled, rebooted, deleted c:\program files\spambayes as well as the User Profile under XP prior to installing a different version. 1.0 Outlook.exe process does not end gracefully (I found this on a sourceforge site) 1.0.3 Outlook.exe process does not end gracefully 1.0.4 Outlook.exe process does not end gracefully 1.1a.1 Outlook.exe process does not end gracefully FWIW: I tried Inboxer 2.1 since it's based on the Spambayes engine and it doesn't hang. If I have time later I will build a second machine with OfficeXP and run the above tests. hth Robert -----Original Message----- From: Meyer, Tony [mailto:T.A.Meyer at massey.ac.nz] Sent: Tuesday, July 19, 2005 8:58 PM To: Robert Mezzone; spambayes-dev at python.org Subject: RE: [spambayes-dev] Outlook Plug-In Bug? > User opens Word or any application that supports the Send-To feature [...] I suspect that this is this bug: [ 1103976 ] Outlook.exe will not terminate when launched via mailto My money is on this being an Outlook bug. It's possible we can work around it somehow, of course. Since this only appears to effect Outlook 2003, and I'm using Outlook 2002, it's pretty difficult for me to look into ways around this (not that I have the time right at the moment, anyway). Kenny is assigned the above bug report, and is going to look into it when he has time, I believe (which won't be for a while as well, IIRC). Whether or not this was a problem in earlier versions of the plug-in is an interesting question, to which I don't know the answer. It's possible that it was, especially since the problem doesn't effect old versions of Outlook. If you (or anyone else) wants to try old plug-in versions, I have copies of them all (or could unhide them on sourceforge temporarily). If the problem doesn't occur with old versions, then that would certainly be valuable information for the bug report, since that gives an idea of where to find a solution. =Tony.Meyer From T.A.Meyer at massey.ac.nz Fri Jul 22 02:55:06 2005 From: T.A.Meyer at massey.ac.nz (Meyer, Tony) Date: Fri, 22 Jul 2005 12:55:06 +1200 Subject: [spambayes-dev] Outlook Plug-In Bug? Message-ID: > I'd be more than happy to try older versions. Old versions are (for the moment at least) here: (After 1.0a9 was just 1.0b1 and 1.0rc1 IIRC, which were all essentially the same as 1.0, which you have already tried). I would suggest trying the oldest version first. If it has the same problem, then it's highly likely that they all do (so every version has had this problem), since it's unlikely that we accidentally fixed the problem in one version, then accidentally broke it again later. =Tony.Meyer From tameyer at ihug.co.nz Fri Jul 22 02:56:32 2005 From: tameyer at ihug.co.nz (Tony Meyer) Date: Fri, 22 Jul 2005 12:56:32 +1200 Subject: [spambayes-dev] Outlook Plug-In Bug? In-Reply-To: Message-ID: > The following were tested on a PC running WINXP SP2 all patches and > Office 2003 SP1 all patches. Thanks for that. > FWIW: I tried Inboxer 2.1 since it's based on the Spambayes engine and > it doesn't hang. Interesting. That indicates that even if this is an Outlook bug as I suspect, there must be a way to work around it. Perhaps if we can't figure it out Sean (of Inboxer) might be willing to contribute the fix. =Tony.Meyer From kenny.pitt at gmail.com Fri Jul 22 17:29:36 2005 From: kenny.pitt at gmail.com (Kenny Pitt) Date: Fri, 22 Jul 2005 11:29:36 -0400 Subject: [spambayes-dev] Outlook Plug-In Bug? In-Reply-To: Message-ID: <42e110b2.48c9ffa6.6956.111c@mx.gmail.com> Tony Meyer wrote: >> The following were tested on a PC running WINXP SP2 all patches and >> Office 2003 SP1 all patches. > > Thanks for that. > >> FWIW: I tried Inboxer 2.1 since it's based on the Spambayes engine >> and it doesn't hang. > > Interesting. That indicates that even if this is an Outlook bug as I > suspect, there must be a way to work around it. Perhaps if we can't > figure it out Sean (of Inboxer) might be willing to contribute the > fix. Knowing if the problem goes away if we go back far enough will be useful information. I wonder if something may have changed in recent versions of pywin32 or py2exe? -- Kenny Pitt From polynomials at remypublishing.com Fri Jul 22 21:47:52 2005 From: polynomials at remypublishing.com (Sara) Date: Fri, 22 Jul 2005 21:47:52 +0200 Subject: [spambayes-dev] Shopping for software? Now in your language & currency! Message-ID: <3042927164.314688789@cc523849-a.hnglo1.ov.home.nl> Software for System Builders, Resellers, and Hardware Purchasers Only. http://ptick.6da39n6h3y6d37o.ekisthmoidgj.info For what is liberty but the unhampered translation of will into act? If they give you ruled paper, write the other way. From michaelmurdock at gmail.com Fri Jul 22 20:33:26 2005 From: michaelmurdock at gmail.com (Michael Murdock) Date: Fri, 22 Jul 2005 12:33:26 -0600 Subject: [spambayes-dev] SpamBayes for Document Categorization? Message-ID: Hello, I am interested in using SpamBayes as the core classifier for a system I want to write that classifies document instances into categories. Instances might be formatted in Word, PDF, text, or html. Of course I don't expect SpamBayes to know how to read all these different formats. So for the sake of discussion, let's just say it could process the text of any document I throw at it. For the sake of discussion, let's say I have five categories with many document instances (training examples) from each of these five categories: Doc Category #1 - streaming media protocols Doc Category #2 - media format conversion tools Doc Category #3 - DirectShow Doc Category #4 - media content management systems Doc Category #5 - none of the above In my proposed system I drag a document instance into a watch folder, which causes a text classifier to open it, analyze it and "tag" it somehow to indicate to which of the five categories it belongs (say by moving it into one of five directories). Here are my five concerns. *1. Embedding the SpamBase code into my app. * My first concern is whether or not the SpamBayes training and classifier code is structured such that it can be embedded into this kind of tool. I'm pretty comfortable with Python. But rewriting major pieces of SpamBayes to do this app would not be fun, nor feasible. *2. SpamBayes for Non-Email-Types of Classification.* Does it even make sense to start with SpamBayes since my problem domain doesn't have anything like email headers or the presence of an attachment, etc. that SpamBayes probably uses in its core feature extraction? *3. Discriminatory Training* My next concern relates to the lack of discriminatory training between categories. I think the way SpamBayes works is my training on a particular class, say class 1, is building a model with which to make the discrimination: Is this document instance a member of class 1 or not? When I train the model for class 1 do I only include positive instances (the ham) of Category 1? Or do I also include negative instances from the other categories (spam)? If the model for Category-1 is only trained on positive instance from that category, then this trained model is independent of the trained models for categories 2 through 5. And when it comes time to make a classification the model that responds "loudest" is the one selected. But, and here's my concern, there has never been a proabability model created that *discriminates between* the categories. Does this make sense what I am describing? I guess I'm thinking about Maximum Likelihood training of acoustic models in a speech recognition system, which has this lack of discriminatory training and I'm wondering if multi-class naive-Bayes classifiers have this same kind of shortcoming. *4. Adding a New Document Category.* Let's say I have trained the models on my five classes (as described above) and everything is working fine and I decide to add a new document category. Do the first five models need to be trained from scratch (to include the negative instances in this new sixth category)? Or can SpamBayes models be "incrementally" trained by just training on these new class-6 negative examples? *5. Size of Training Sample* My final concern relates to the number of training documents I would need. I'm guessing that each of my documents, no matter how long or short, reduces to a single feature vector for training and classification. Is this correct? If so, it would seem that I would need *at least* hundreds of examples from each category and probably thousands. Yes? No? Thanks for any thoughts you might have on my concerns and questions. ~Michael. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mail.python.org/pipermail/spambayes-dev/attachments/20050722/95a8332b/attachment.htm From benhdj at users.sourceforge.net Sun Jul 31 09:37:51 2005 From: benhdj at users.sourceforge.net (Benjamin Han) Date: Sun, 31 Jul 2005 00:37:51 -0700 Subject: [spambayes-dev] Integrating SpamBayes with a Mail.app plugin Message-ID: <435F4C0F-5B1E-430B-A413-E02B3F1F7E14@users.sourceforge.net> Hi list, I just started working on integrating SpamBayes 1.0.4 into JunkMatcher, a Mail.app plugin I wrote on OS X. I've heard a lot of good things about SB and I'd love to have it work for me. :-) In JunkMatcher messages are filtered based on conducting a list of tests, and these include both "property tests" and "pattern tests". The latter are regex matching, and all other kinds of tests are property tests (including blacklists querying, etc.). (more info about JunkMatcher: http://junkmatcher.sf.net ) My plan is to integrate SB as a property test. I've downloaded the code and took a brief look. It seems that I only need to replicate the sb_filter.py so SpamBayes would kick in when checking. I'd appreciate if someone can do a bit hand-holding for me - here are a few newbie questions: 1. Is Classifier.*_spamprob() thread-safe? I will be calling them from multiple threads. Anything I should be aware of? 2. I've looked around and tried to figure out what would be saved into SB's database. Am I correct in saying that it only saves the classification results and IDs, not the raw message source? This is because JunkMatcher already saved raw email sources and I'd like to avoid duplicating data. 3. At least as the first cut, I'll provide a very dumb-down UI for SB (i.e., no user configuration UI will be provided). I already know I need to give some default settings as to where to deposit hammie.db. Any other default settings I should be aware of? Or using the defaults in Options.py is all I need to do? Thanks in advance, Ben