genex-dev — GeneX developers

[GeneX-dev] Re: GeneX listservs and help

[Harry M. <hj...@nc...>]

Well Well well... Then it might mean that we could get GeneX running on
NT/2k as well?
Have to think about this, but it might be possible, tho still
aesthetically ugly... ;)

if apache and cygwin are ok, and you can implement the genex bits on the
cygwin side (so paths, permissions, system calls etcare handled the
linux way, then I guess there's no reason that it couldn't work.  R has
a win port (or probably better to run the Cygwin Linux port)

hjm




Todd Peterson wrote:
> 
> I have Oracle running on my PC (800 MHz AMD Thunderbird with 512M RAM) along
> with Apache. Works fine. Cygwin packages works great, too.
> 
> ----- Original Message -----
> From: "Harry Mangalam" <man...@ho...>
> To: <adl...@uc...>
> Cc: "genex developers" <gen...@nc...>
> Sent: Wednesday, May 16, 2001 3:51 PM
> Subject: GeneX listservs and help
> 
> > Hi ... sorry I didn't get your 1st name..
> >
> > Here are the email addresses to post to if you want help on the genex
> system:
> >
> > ge...@nc... [the NCGR geenx develpers list (~6 developers)]
> >
> >
> > gen...@li... [the larger Open Source group of genex
> developers ]
> >
> > to get the broadest response to your questions, you should post to the
> genex-dev list.
> >
> > We're pretty good at responding to queries right away.
> >
> > Hi All,
> >
> > Please welcome Thea Wilkins' group from UC Davis.  They have some
> interesting challenges, like porting to Oracle, running on NT.
> >
> > It occurs to me that one thing I DIDN'T consider was that it's
> theoretically possible (translatation: almost certainly impossible) to run
> the GeneX system on a Cygwin-supervised chunk of the system so
> > that GeneX doesn't know it;s running on NT...?
> >
> > Anyone ever tried to run something as complex as a RDBMS on Cygwin?
> > I know some of my codes which were written explicitly for Unixy OSs run
> amazingly well on Cygwin, but it IS another layer to consider and supporting
> the bizarrities of Microsoft's OSs is not something
> > I'm really interested in..
> >
> >
> >
> > --
> > Cheers,
> > Harry
> >
> > Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... ||
> man...@ho...
> >

-- 
Cheers,
Harry

Harry J Mangalam (949) 856 2847(v&f)  hj...@nc... || man...@ho...

[GeneX-dev] GeneX installation problemes

[Ron O. <Ls...@wi...>]

Dear developers,
I happened to get to the stage of running install-all.pl. There I had =
three surprisees: I didn't install xcluster, dtd2html and JAR. For all of =
these the install-all script (IAS) asked for paths which I couldn't pride. =
Following that senario I had the following questions and problems:
Assuming I will manage to install these components/applications how do I =
set their paths. Should I re-run IAS?
I couldn't find dtd2html at the URL specified in the script?
Where can I find  a rpm for JAR?
In addition IAS loged out the following error which I presume relate to =
the fact that dtd2html wasn't installed:

!! System Error: No such file or directory @ line: 2256  (  -tree  -all =
genexml-all.html  -dtd genexml  -home genexml-home.html  -out /usr/local/li=
b/dtd/genexml-html  -top genexml-    top.html  --treefile genexml-tree.html=
  /usr/local/lib/dtd/genexml.dtd)

!! System Error: No such file or directory @ line: 2267  (  -tree  -all =
ams-all.html  -dtd ams  -home ams-home.html  -out /usr/local/lib/dtd/genexm=
l-html  -top ams-top.html  --treefile ams-tree.html  /usr/local/lib/dtd/ams=
.dtd)

!! System Error: No such file or directory @ line: 2278  (  -tree  -all =
als-all.html  -dtd als  -home als-home.html  -out /usr/local/lib/dtd/genexm=
l-html  -top als-top.html  --treefile als-tree.html  /usr/local/lib/dtd/als=
.dtd)

!! System Error: No such file or directory @ line: 2289  (  -tree  -all =
usf-all.html  -dtd usf  -home usf-home.html  -out /usr/local/lib/dtd/genexm=
l-html  -top usf-top.html  --treefile usf-tree.html  /usr/local/lib/dtd/usf=
.dtd)

!! System Error: Illegal seek @ line: 2488  (make -s test)


insn't it?

Thank you in advance for your support,

Ron Ophir, Ph.D.
DNA Microarray Informatics,
Biological Services
Weizmann Istitute of Science
POB 26
Rehovot 76100
Israel
e-mail: Ron...@we...
Phone: 972-8-9343456
Fax:972-8-9344113

[GeneX-dev] Re: genex-dump.gz

[Harry M. <man...@ho...>]

thanks very much for the update notice!  Glad to see that you got thru it OK!

Cheers
Harry


System Administrator wrote:
> 
> Many thanks for your assistance. I suppose the installation instructions
> on the website (http://genex.ncgr.org) are somewhat dated. I downloaded
> the file you specified and everything worked like a charm :) By the way, I
> am working on a document that steps through an installation process on a
> Solaris box. Even though, genex requires a good deal of development
> software, I was able to get the number of packages needed on a Solaris box
> to a bare minimum (31 as opposed to the standard 220). Once I have fully
> tested the installation, I'll send your support team a pointer to the URL.
> Thus far, everything is working beautifully. Again, thank you for your
> time and help.
> 
> On Fri, 11 May 2001, Harry Mangalam wrote:
> 
> > Hi there,
> >
> > It looks like you're using a very old version of the either the instructions or the installation package or both.
> >
> > We currently require Postgres 7 to handle some of the db constructs and you're using a postgres 6 data dump to try to initialize the database.
> >
> > (you're also trying to feed postgres a gzipped file which is sure to casue it to fail - you should at minimum decompress it before you try this:
> >
> > zcat  genex-dump-001202.pg6.gz | psql genex
> > ^
> > ('zcat' decompresses the file on the fly; 'cat' doesn't)
> >
> >
> >
> > However, PLEASE try the latest version which is at:
> >
> > http://genebox.ncgr.org/genex/download/genex-server/GeneX-Server-1.0.2.tar.gz
> >
> > and the appropriate DB instantiation scripts:
> >
> > http://genebox.ncgr.org/genex/download/DB/genex-initdb-latest.gz
> >
> > Also, we can be of MUCH MORE help if you send us the [Option.Reminders] file and the
> > [genex-errors.txt] file that are automatically generated by the installation script.
> >
> >
> > Damian Silvera wrote:
> > >
> > >    This is Damian Silvera from the University of Miami School of
> > > Medicine. We have dilligently attempted to bring your product into
> > > production but have hit a few snags. Currently, we are attempting to
> > > load the genex-dump.gz file into our genex database but have received
> > > the following error on three separate attempts on three separate
> > > servers:
> > >
> > > ERROR:  copy: line 1, CopyReadAttribute: end of record marker corrupted
> > > PQendcopy: resetting connection
> > >
> > > We downloaded genex-dump-001202.pg6.gz from
> > > http://www.ncgr.org/research/genex. The command we used was cat
> > > genex-dump-001202.pg6.gz | psql genex. If we went to the wrong place,
> > > used the wrong commands or downloaded the wrong file, could you let us
> > > know where to start.
> > >
> > > Regards
> >
> > --
> > Cheers,
> > Harry
> >
> > Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho...
> >

-- 
Cheers,
Harry

Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho...

[GeneX-dev] Re: Fw: GeneX installation

[Harry M. <man...@ho...>]

Hi Ron,

The address you want to use for support is either ge...@nc... or (better, wider)
gen...@li....

In order for this to work, you have to have a recent (> 7.0) version of the Postgres source on your system (or a least the postgres include and lib directories); usually they're installed in ;

/usr/local/pgsql/include
/usr/local/pgsql/lib

ie:
/usr/local
|
+-- pgsql/
    |-- bin
    |-- include            < !
    |   |-- access
    |   |-- commands
    |   |-- executor
    |   |-- lib
    |   |-- libpq
    |   |-- libpq++
    |   |-- port
    |   |   `-- linux
    |   `-- utils
    `-- lib                < !


so your environment variables would be set like this:

setenv POSTGRES_INCLUDE /usr/local/pgsql/include
setenv POSTGRES_LIB /usr/local/pgsql/lib


If you had installed your postgres lib and include dir's in your /usr/lib and /usr/include, your described approach would have worked, but it's usually better to keep them separate.

Harry


> I am tying to install GeneX as the database server for Weizmann Institute
> expression data. I am currently in the final stage of the relevant component
> of CPAN. I have a problem with the environment variable POSTGRES_INCLUDE and
> POSTGRES_LIB. These values were not set by postgres installation. Therefore
> I set the POSTGRES_INCLUDE to /usr/include and POSTGRES_LIB to /usr/lib.
> However, didn't help. Maybe you could spare me a good advise based on your
> experience.
> Thank you in advance,
> 
> Ron Ophir, Ph.D.
> DNA Microarray Informatics,
> Biological Services
> Weizmann Istitute of Science
> POB 26
> Rehovot 76100
> Israel
> e-mail: Ron...@we...
> Phone: 972-8-9343456
> Fax:972-8-9344113
> 
>   --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>                Name: TEXT.htm
>    TEXT.htm    Type: Hypertext Markup Language (text/html)
>            Encoding: quoted-printable

-- 
Cheers,
Harry

Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho...

Re: [GeneX-dev] VNC server/Java client working - some questions..

[Harry M. <man...@ho...>]

Hi Michael,  Thanks for the comments!

Michael Pear wrote:
> 
> Harry,
> 
> My reaction is that handling vnc in this way goes beyond the threshold for
> several reasons.
> 
> 1) dealing with things they may not understand or have access to such as ssh
> from a Mac or PC...I don't want my servers to have telnet access!

Yes, I agree, but letting an anonymous user in WITHOUT any authentication is worse.  There are ssh clients for all platforms and I was thinking of including URLs to all of them in the help file. 
But.....

> 
> 2) Requires a login for each user

But eventually, we will be using a user login for data-loading and this is something that can be handled by the sysadmin/madman.

> 3) For the goal of displaying something in a browser, it seems
> counterintuitive to have to prepare
> something manually on the server

....This I strongly agree with - that's why I spent so much time trying to get it to run anonymously, but in the end, I backed off as I was uncomfortable with the idea of allowing anonymous access to
a server.  I config'ed the user (nobody or apache or whoever) as having a minimal desktop with no obvious way of starting a telnet session, (and the user doesn't have a default shell, but I'm sure
with a little seaching I could do it.

If it's anonymous, it also requries a lot of server-side monitoring to watch Xvnc sessions and clean up after they've gone idle, otherwise you have a lot of idle Xvnc's hanging about.  It's certainly
possible to do, but it's sort of fidgety.
 
> I had set up vnc to work on one of my installations of GeneX, and it required
> some machinations due to the password checking, particular as everything is
> run
> by the web server uid (apache in my case). In the end, it required a
> combination of
> manually creating the passwd file for a user, copying it to an area readable
> by the web
> server, and modifying the vncserver script to skip a specific mode check on
> the passwd file.
> 
> I'm not sure how the procedure below solves the problem. It gives the user
> access to
> an X window display in their browser, but how do you connect xgobi or other
> processes
> which are initiated by the  web server?

xgobi is told to send its display to the screen selected (which the user connects to).  It works.


> >From what I can tell, all of the security checking for VNC is done in the
> "vncserver" script then
> it starts "Xvnc". Perhaps another approach is to replace the vncserver
> script with a variant that
> handles the check on the user appropriately for the GeneX situation.
> I gather that the problem is that the user ends up with shell access to the
> server, so you want
> them to be appropriately constrained.
> 
> Perhaps we should step back and identify the security concerns, and ask the
> question whether they can be properly addressed with a locally modified
> "vncserver" that could be incorporated into GeneX.

I'm definitely willing to try this (and I agree it is the way to go, but then we end up maintaining a branch of the VNC server distro), but unless you feel very strongly against it (I'll wait to hear
from you) I'm going to commit my changes to CVS to let you see how they work and try them on your side.  We can alway it back down.

Or I can just allow you access to the version I have running on my machine so you can see how it works..?

Let me know....

Harry


> Regards,
> 
> Michael
> ----- Original Message -----
> From: "Harry Mangalam" <man...@ho...>
> To: "genexdev at SF" <gen...@li...>
> Sent: Tuesday, May 08, 2001 7:37 PM
> Subject: [GeneX-dev] VNC server/Java client working - some questions..
> 
> > After spending an egregious amount of time trying to get the whole thing
> to run automatically & anonymously as 'nobody' and finally getting it to
> more or less work (with a lot of security breaches and
> > dubious work-arounds), I've backed off and am hoping that the following is
> still acceptable.
> >
> > it's considerably less fidgety (for me), more secure (for everyone), but
> more fidgety for users.  Id this too intrusive or is it still below the
> horizon of fatal annoyance for a biologist?
> >
> > The user has to:
> >
> > 1) Use telnet or ssh to log into the GeneX server with real
> >     username / password.
> >
> > 2) start vncserver with no parameters:
> >     'vncserver'
> >
> >     if it's the 1st time that you're using it, it will ask you to generate
> a
> >     password - do a good one (it's stored encrypted in ~/.vnc/passwd)
> using
> >     the vncpasswd util - just type 'vncpasswd' and enter a good password 2
> X
> >     in response to the prompts.
> >
> >     then run vncserver again as before.
> >
> > 3) Note the display that vncserver will write to:
> >
> >     New 'X' desktop is [hostname]:5
> >                                   ^
> >
> > When you fill out the CyberT form, fill out the xgobi parameter as normal
> > "View top [    ] results by 'p'."  and then select that display number
> above
> > to write to at the bottom.  "VNC Display"
> >
> > Once the analysis is done, a URL will be generated that, once clicked,
> will start the Java applet and connect to the VNC server, once you supply
> the password that you set
> > with the vncpasswd application.
> >
> > Is that an acceptable procedure?  If no horrible objections, I'll move the
> changes into CVS and let people play with them.
> >
> > --
> > Cheers,
> > Harry
> >
> > Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... ||
> man...@ho...
> >
> > _______________________________________________
> > Genex-dev mailing list
> > Gen...@li...
> > http://lists.sourceforge.net/lists/listinfo/genex-dev
> 
> _______________________________________________
> Genex-dev mailing list
> Gen...@li...
> http://lists.sourceforge.net/lists/listinfo/genex-dev

-- 
Cheers,
Harry

Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho...

Re: [GeneX-dev] VNC server/Java client working - some questions..

[Michael P. <mic...@ho...>]

Harry,

My reaction is that handling vnc in this way goes beyond the threshold for
several reasons.

1) dealing with things they may not understand or have access to such as ssh
from a Mac or PC...I don't want my servers to have telnet access!

2) Requires a login for each user

3) For the goal of displaying something in a browser, it seems
counterintuitive to have to prepare
something manually on the server

I had set up vnc to work on one of my installationsof GeneX, and it required
some machinations due to the password checking, particular as everything is
run
by the web server uid (apache in my case). In the end, it required a
combination of
manually creating the passwd file for a user, copying it to an area readable
by the web
server, and modifying the vncserver script to skip a specific mode check on
the passwd file.

I'm not sure how the procedure below solves the problem. It gives the user
access to
an X window display in their browser, but how do you connect xgobi or other
processes
which are initiated by the  web server?

From what I can tell, all of the security checking for VNC is done in the
"vncserver" script then
it starts "Xvnc". Perhaps another approach is to replace the vncserver
script with a variant that
handles the check on the user appropriately for the GeneX situation.
I gather that the problem is that the user ends up with shell access to the
server, so you want
them to be appropriately constrained.

Perhaps we should step back and identify the security concerns, and ask the
question whether they can be properly addressed with a locally modified
"vncserver" that could be incorporated into GeneX.

Regards,

Michael
----- Original Message -----
From: "Harry Mangalam" <man...@ho...>
To: "genexdev at SF" <gen...@li...>
Sent: Tuesday, May 08, 2001 7:37 PM
Subject: [GeneX-dev] VNC server/Java client working - some questions..


> After spending an egregious amount of time trying to get the whole thing
to run automatically & anonymously as 'nobody' and finally getting it to
more or less work (with a lot of security breaches and
> dubious work-arounds), I've backed off and am hoping that the following is
still acceptable.
>
> it's considerably less fidgety (for me), more secure (for everyone), but
more fidgety for users.  Id this too intrusive or is it still below the
horizon of fatal annoyance for a biologist?
>
> The user has to:
>
> 1) Use telnet or ssh to log into the GeneX server with real
>     username / password.
>
> 2) start vncserver with no parameters:
>     'vncserver'
>
>     if it's the 1st time that you're using it, it will ask you to generate
a
>     password - do a good one (it's stored encrypted in ~/.vnc/passwd)
using
>     the vncpasswd util - just type 'vncpasswd' and enter a good password 2
X
>     in response to the prompts.
>
>     then run vncserver again as before.
>
> 3) Note the display that vncserver will write to:
>
>     New 'X' desktop is [hostname]:5
>                                   ^
>
> When you fill out the CyberT form, fill out the xgobi parameter as normal
> "View top [    ] results by 'p'."  and then select that display number
above
> to write to at the bottom.  "VNC Display"
>
> Once the analysis is done, a URL will be generated that, once clicked,
will start the Java applet and connect to the VNC server, once you supply
the password that you set
> with the vncpasswd application.
>
> Is that an acceptable procedure?  If no horrible objections, I'll move the
changes into CVS and let people play with them.
>
> --
> Cheers,
> Harry
>
> Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... ||
man...@ho...
>
> _______________________________________________
> Genex-dev mailing list
> Gen...@li...
> http://lists.sourceforge.net/lists/listinfo/genex-dev

Re: [GeneX-dev] VNC server/Java client working - some questions..

[<ja...@op...>]

"Harry Mangalam" <man...@ho...> writes:

> Is that an acceptable procedure?  If no horrible objections, I'll
> move the changes into CVS and let people play with them.

No useful comments from me. I've got a working X server (and it only
took me 6 months to get it to display at 1200x1024 on my 21" monitor,
so I must be smart ... ;-)

jas.

[GeneX-dev] VNC server/Java client working - some questions..

[Harry M. <man...@ho...>]

After spending an egregious amount of time trying to get the whole thing to run automatically & anonymously as 'nobody' and finally getting it to more or less work (with a lot of security breaches and
dubious work-arounds), I've backed off and am hoping that the following is still acceptable.

it's considerably less fidgety (for me), more secure (for everyone), but more fidgety for users.  Id this too intrusive or is it still below the horizon of fatal annoyance for a biologist?

The user has to:

1) Use telnet or ssh to log into the GeneX server with real 
    username / password.

2) start vncserver with no parameters:
    'vncserver'
  
    if it's the 1st time that you're using it, it will ask you to generate a 
    password - do a good one (it's stored encrypted in ~/.vnc/passwd) using
    the vncpasswd util - just type 'vncpasswd' and enter a good password 2 X
    in response to the prompts.

    then run vncserver again as before.

3) Note the display that vncserver will write to:

    New 'X' desktop is [hostname]:5 
                                  ^

When you fill out the CyberT form, fill out the xgobi parameter as normal 
"View top [    ] results by 'p'."  and then select that display number above
to write to at the bottom.  "VNC Display"

Once the analysis is done, a URL will be generated that, once clicked, will start the Java applet and connect to the VNC server, once you supply the password that you set
with the vncpasswd application.

Is that an acceptable procedure?  If no horrible objections, I'll move the changes into CVS and let people play with them.

-- 
Cheers,
Harry

Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho...

Re: [GeneX-dev] Re: Minor thing to fix..

[<ja...@op...>]

"Michael Pear" <mic...@ho...> writes:

> Somehow, I was able to install things in a directory called "perl"
> rather than "perl5" using the GeneX install tools. I'll look for
> where that can be specified the next time I go through a complete
> install and be more specific. If it is possible to use another name,
> then it should be possible to point to it when upgrading.

My appologies. I read my code (how embarassing ;-). I *do* allow users
to override the perl library installation path. 

Does anyone have any reason why we should allow this? Is there a
problem with always installing in LOCAL_ROOT/perl?

jas.

Re: [GeneX-dev] Re: Minor thing to fix..

[Michael P. <mic...@ho...>]

Hi Jason,

 Somehow, I was able to install things in a directory called "perl" rather
than "perl5" using
the GeneX install tools. I'll look for where that can be specified the next
time I go through
a complete install and be more specific. If it is possible to use another
name,
then it should be possible to point to it when upgrading.

Regards,

Michael
----- Original Message -----
From: "Jason E. Stewart" <ja...@op...>
To: "Michael Pear" <mic...@ho...>
Cc: <gen...@li...>
Sent: Tuesday, May 08, 2001 2:48 PM
Subject: [GeneX-dev] Re: Minor thing to fix..


> Hey Michael,
>
> This is a useful issue, if you don't mind, I'll keep it public.
>
> I'm happy to do this. In the beginning, the user was able to specify
> every single path in the installation. This made things a bit
> unwieldy, so we went the opposite route to some extent, and hard-coded
> (relatively speaking) a number of the installation
> paths. LOCAL_ROOT/perl5 is one of them.
>
> Can you give me a clear need why it needs to be anything
> different. Are you hand-install Genex.pm into a different location? I
> would just like to understand the need before I make the change,
> because there are already so many different things the installer can
> configure, I'd like to not add more unless there's a clear need.
>
> Thanks,
> jas.
>
> "Michael Pear" <mic...@ho...> writes:
>
> > When updating to a new release, when prompted for the LOCAL_ROOT to
> > find the Config.pm file, the path is assumed to be
> > "LOCAL_ROOT/perl5/Bio/Genex/Config.pm".  This doesn't allow for
> > circumstances where the Bio directory is in a directory other than
> > "perl5", like "perl". Easily circumvented with a ln -s perl perl5,
> > but next time you are working in the install area, it would be good
> > to allow specification of the path down to the "Bio" directory for
> > maximum flexibility.
> >
> > Regards,
> >
> > Michael Pear
>
> _______________________________________________
> Genex-dev mailing list
> Gen...@li...
> http://lists.sourceforge.net/lists/listinfo/genex-dev

[GeneX-dev] Re: Minor thing to fix..

[<ja...@op...>]

Hey Michael,

This is a useful issue, if you don't mind, I'll keep it public.

I'm happy to do this. In the beginning, the user was able to specify
every single path in the installation. This made things a bit
unwieldy, so we went the opposite route to some extent, and hard-coded
(relatively speaking) a number of the installation
paths. LOCAL_ROOT/perl5 is one of them.

Can you give me a clear need why it needs to be anything
different. Are you hand-install Genex.pm into a different location? I
would just like to understand the need before I make the change,
because there are already so many different things the installer can
configure, I'd like to not add more unless there's a clear need.

Thanks,
jas.

"Michael Pear" <mic...@ho...> writes:

> When updating to a new release, when prompted for the LOCAL_ROOT to
> find the Config.pm file, the path is assumed to be
> "LOCAL_ROOT/perl5/Bio/Genex/Config.pm".  This doesn't allow for
> circumstances where the Bio directory is in a directory other than
> "perl5", like "perl". Easily circumvented with a ln -s perl perl5,
> but next time you are working in the install area, it would be good
> to allow specification of the path down to the "Bio" directory for
> maximum flexibility.
> 
> Regards,
> 
> Michael Pear

Re: [GeneX-dev] Embperl pages for GeneX available through CVS

[<ja...@op...>]

"Michael Pear" <mic...@ho...> writes:

>  cvs -d :pserver:ano...@ge...:/home/cvsroot/opengenex \
>    checkout  -r mpear_experimental_1_02 genex-server
                                    ^^^^^
Minor correction: mpear_experimental_1_0_2

Thanks for checking this in Michael, I'm looking at it right now. 
jas.

[GeneX-dev] Embperl pages for GeneX available through CVS

[Michael P. <mic...@ho...>]

I have checked in a collection of pages using Embperl to access the GeneX database.
The pages are intended to provide processing functions for microarray data such as
normalization, ratio computations, averaging, etc. They also provide the ability to
screen experiments that are viewable according to a user login, based on the
user login created for GeneX. That is, only experiments that are loaded into the
database for a particular user are viewable.

The code is available from the cvs tree as follows:

 cvs -d :pserver:ano...@ge...:/home/cvsroot/opengenex \
   checkout  -r mpear_experimental_1_02 genex-server
[server responds with]
(Logging in to ano...@ge...)
CVS password: [you hit Return or Enter]

The pages and a number of sql functions for the database are in a sub directory
"embperl". There is a very brief INSTALL file describing the steps to install that
I will flesh out in the future.

A few caveats:

1) The pages will only work with the V1 database schema, not the most recent
version 1.1 Jason is working on.
2) the pages that deal with slide data (have "slide" in the name; accessed by
"Display Slides" button) rely upon a particular organization of slide data and their
intra-slide replicate spots to work properly. Until I have time to supply documentation
and the corresponding loading utilties, they won't work.
3) There is no link from the genex pages to these pages at this time. You will need to
access them directly, starting with "display-experiments.plhtml".
4) The installation is very manual at this point. The pages are not installed as part of the genex
installation script (Jason I'll ask for your guidance in how we could/should do that) and require
manually running a database script to load the sql functions. You will need to set up a directory
under your html directory for genex and adapt the apache configuration file
supplied to your local site.

Before attempting using them, I would suggest configuring Embperl and do a simple
test page with it to ensure that is working.

I'll keep an eye out for questions on the genex-dev list by my capacity will be limited
for the next couple of days.

Good luck to any brave souls..

Regards,

Michael Pear

Re: [GeneX-dev] Michael's work in CVS?

[Michael P. <mic...@ho...>]

Jason,

 I have just checked in the embperl pages I have worked on 
in branch "mpear_experimental_1_0_2" which branches off
the tag "Rel-1_0_2t1". I'll send out a separate message on this.

My capacity for answering questions will be limited until Wednesday afternoon.

Regards,

Michael

----- Original Message ----- 
From: "Jason E. Stewart" <ja...@op...>
To: <gen...@li...>
Sent: Sunday, May 06, 2001 3:18 PM
Subject: [GeneX-dev] Michael's work in CVS?


> Hey Harry, Michael,
> 
> I find myself in the need of some html forms to add info for the data
> loader. I would like to use Michaels embperl framework if possible (I
> don't want to have to rewrite the wheel). 
> 
> Did it ever make it into CVS? If not, Michael, why don't you send me a
> tarball, and I will include it. I think I'm very near to cutting the
> first public release of Genex 1.1.0. Very soon all the API will work,
> and we will just need to have some test data in the DB.
> 
> jas.
> 
> _______________________________________________
> Genex-dev mailing list
> Gen...@li...
> http://lists.sourceforge.net/lists/listinfo/genex-dev

Re: [GeneX-dev] What's required for Xerces.pm

[<ja...@op...>]

"Harry Mangalam" <man...@ho...> writes:

> OK - downloaded and installed the latest Xerces stuff (1.4), but
> trying to install XML-Xerces-1.3.5, while it did the make's ok,
> resulted in:

This seems good

> t/DOMCount..........Can't load 'blib/arch/auto/XML/Xerces/Xerces.so'
> for module XML::Xerces: blib/arch/auto/XML/Xerces/Xerces.so:
> undefined symbol: _._17SAXParseException at
> /usr/lib/perl5/5.00503/i386-linux/DynaLoader.pm line 169.

This is a dynamic loading error, notice the 'undefined symbol'. This
means that DynaLoader asked ld.so to load Xerces.so and it returned an
undefined symbol exception. 

This is in libxerces.so:

~/work/xml-xerces/c/lib $ nm libxerces-c1_4.so |grep SAXParseException
001437c8 T _._17SAXParseException

or if we demangle the names:

~/work/xml-xerces/c/lib $ nm --demangle libxerces-c1_4.so |grep SAXParseException
001437c8 T SAXParseException::~SAXParseException(void)

So it looks like you don't have your LD_LIBRARY_PATH set up to find
libxerces? Maybe running ldconfig will help?

Did you get a warning when you ran 'perl Makefile.PL' about not
finding libxerces?

jas.

Re: [GeneX-dev] What's required for Xerces.pm

[Harry M. <man...@ho...>]

OK - downloaded and installed the latest Xerces stuff (1.4), but trying to install XML-Xerces-1.3.5, while it did the make's ok, resulted in:


May 07@10:25am /tmp/XML-Xerces-1.3.5 
7 % make test
make[1]: Entering directory `/d/tmp/XML-Xerces-1.3.5/New_classes_perl'
make[1]: Leaving directory `/d/tmp/XML-Xerces-1.3.5/New_classes_perl'
make[1]: Entering directory `/d/tmp/XML-Xerces-1.3.5/DOMParse'
make[1]: Leaving directory `/d/tmp/XML-Xerces-1.3.5/DOMParse'
PERL_DL_NONLAZY=1 /usr/bin/perl -Iblib/arch -Iblib/lib -I/usr/lib/perl5/5.00503/i386-linux -I/usr/lib/perl5/5.00503 -e 'use Test::Harness qw(&runtests $verbose); $verbose=0; runtests @ARGV;' t/*.t
t/DOMCount..........Can't load 'blib/arch/auto/XML/Xerces/Xerces.so' for module XML::Xerces: blib/arch/auto/XML/Xerces/Xerces.so: undefined symbol: _._17SAXParseException at
/usr/lib/perl5/5.00503/i386-linux/DynaLoader.pm line 169.

 at blib/lib/XML/Xerces.pm line 7
BEGIN failed--compilation aborted at t/DOMCount.t line 14.
dubious
        Test returned status 255 (wstat 65280, 0xff00)
Undefined subroutine &Test::Harness::WCOREDUMP called at /usr/lib/perl5/5.00503/Test/Harness.pm line 288.
make: *** [test_dynamic] Error 255


the supposedly missing file 'blib/arch/auto/XML/Xerces/Xerces.so' is where it's supposed to be so this seems odd.


-- 
Cheers,
Harry

Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho...

Re: [GeneX-dev] What's required for Xerces.pm

[Harry M. <man...@ho...>]

"Jason E. Stewart" wrote:
> > Also, we should talk before we dive into the forms - that is a bit
> > where I might be able to help, tho it woul dprobably be best if we
> > could autogenerate a minimal set of forms based on the XML table
> > descriptors and perhaps an additional tag such as 'forms_required'
> 
> Hmmm... Hadn't thought about the auto-generation idea. Seems
> workable.
> 
> jas.

Hmm - yeah, I had originally thought about this as a way to generate skeleton forms, but it occurs to me that this process would also helpus with the ILSI custom form generation or indeed ANY
customization.  If we can come up with a std way to generate forms based on the XML and your tdscript instantiates the DB based on it, then ANY customization of the XML definition can auto-generate
the rest of the forms, instantiatiation, and other stuff.  Neat stuff if we can genrate nice forms...

> _______________________________________________
> Genex-dev mailing list
> Gen...@li...
> http://lists.sourceforge.net/lists/listinfo/genex-dev

-- 
Cheers,
Harry

Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho...

Re: [GeneX-dev] What's required for Xerces.pm

[<ja...@op...>]

"Harry Mangalam" <man...@ho...> writes:

> Just starting to dive into the genex XML stuff.  What's requqired
> for Xerces.pm?

Um. Good catch. I need to add this information to the GeneX README,
thanks. 

> obviously:
> http://xml.apache.org/dist/xerces-p/XML-Xerces-1.3.5.tar.gz

Yes, thats the current stable release.

> Do we also need the Xerces-C 1.3 distro?  and if so where do we get
> it?

No, you can use Xerces-C 1.4, which is the stable distribution:

http://xml.apache.org/dist/xerces-c/stable/

Xerces.pm 1.3.x runs fine with Xerces-C 1.4, it just doesn't expose
any of the (minor) API changes 1.4 makes available. Xerces.pm 1.4 is
still a moving target, and I won't have a non-beta version ready for a
week or two.

> Do we also need the SWIG stuff?

No, SWIG is needed only if you want to make code changes to the C++
code. 

> Also, we should talk before we dive into the forms - that is a bit
> where I might be able to help, tho it woul dprobably be best if we
> could autogenerate a minimal set of forms based on the XML table
> descriptors and perhaps an additional tag such as 'forms_required'

Hmmm... Hadn't thought about the auto-generation idea. Seems
workable. 

jas.

[GeneX-dev] What's required for Xerces.pm

[Harry M. <man...@ho...>]

Hi Jason,

Just starting to dive into the genex XML stuff.  What's requqired for Xerces.pm?

obviously:
http://xml.apache.org/dist/xerces-p/XML-Xerces-1.3.5.tar.gz

Do we also need the Xerces-C 1.3 distro?  and if so where do we get it?
Do we also need the SWIG stuff?

I can't find any of this info in the docs yet...

Also, we should talk before we dive into the forms - that is a bit where I might be able to help, tho it woul dprobably be best if we could autogenerate a minimal set of  forms based on the XML table
descriptors and perhaps an additional tag such as 'forms_required'


-- 
Cheers,
Harry

Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho...

[GeneX-dev] Michael's work in CVS?

[<ja...@op...>]

Hey Harry, Michael,

I find myself in the need of some html forms to add info for the data
loader. I would like to use Michaels embperl framework if possible (I
don't want to have to rewrite the wheel). 

Did it ever make it into CVS? If not, Michael, why don't you send me a
tarball, and I will include it. I think I'm very near to cutting the
first public release of Genex 1.1.0. Very soon all the API will work,
and we will just need to have some test data in the DB.

jas.

Re: [GeneX-dev] Question on CyberT changes

[Harry M. <man...@ho...>]

Sure, Michael

I didn't mean to be obscure, but it's explained a bit more fully in the help page section:

http://genex.ncgr.org/genex/cybert/CTHelp.html#PairedExprCol

The upshot is that the current data in the Version 1 database has no provision for providing the needed data dor the Bayesian version of the Paired analysis so I eliminated it.  The upload - data
version does have this ability so if you calculate this column locally and feed it to the analysis it will work.  This should be possible to support in the version 2 DB, along eith a number of other
things.

This was an error that allowed it to be released at all in the Paired version (we were working with C+E data and rolled it over to the paired version without considering what it actually required. 
Our apologies...

Harry



Michael Pear wrote:
> 
> Harry,
> 
> I see in your change log for Cyber T a notation for V1.0.2
> 
>   * slight mod of the form to disallow Bayes analysis on data from the DB as
>     there's no way to provide the expression levels required.  Only can do Bayes
>     with uploaded data at this point.
> 
> As we really want to use CyberT from the database, what does this mean?
> Does this affect the paired and control vs experiment modes of Cyber T?
> Could you be more specific regarding the issue of "expression levels" vs.
> "magnitude of the ratios" which apparently prompted some fixes?
> 
> Regards,
> 
> Michael Pear
> 
> _______________________________________________
> Genex-dev mailing list
> Gen...@li...
> http://lists.sourceforge.net/lists/listinfo/genex-dev

-- 
Cheers,
Harry

Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho...

Re: [GeneX-dev] Trouble with genex installation

[Harry M. <man...@ho...>]

Hi there,

Thanks for getting back in touch with us.. 

Could you please send us the URL for the status page for the installation:

http://[the.genex.hostname.here]/cgi-bin/genex/samples/status.pl

or if it's unavailable to the outside world, could you run that URL and send us the output?

as well as the files:

Option.Reminders 
&
genex-errors.txt

Cheers
harry



"Silvera, Damian" wrote:
> 
> Good evening,
> 
> my name is Damian Silvera. I am contacting you from the University of Miami
> School of Medicine. We have implemented (or are trying to) GeneX on a Sun
> Ultra 1 running Solaris 7 (a secure, hardened version of the OS). At
> present, we have all the prerequisites installed (I am working on a web page
> that will step through the recipe) and are ready to run the install-all.pl
> script. We are able to get through 90% of the script but get stuck at the DB
> Updater part. We are getting the following error:
> 
> Can't locate Bio/Genex.pm in @INC (@INC contains:
> /usr/local/lib/perl5/5.00563/sun
> 4-solaris /usr/local/lib/perl5/5.00563
> /usr/local/lib/site_perl/5.00563/sun4-solar
> is /usr/local/lib/site_perl .) at ./db-update-0.8.1-0.9.pl line 13.
> BEGIN failed--compilation aborted at ./db-update-0.8.1-0.9.pl line 13.
> 
> The Genex.pm perl module is in the /usr/local/lib/perl5/5.00563. I can't see
> why the script is dying, however. Any ideas?
> 
> _______________________________________________
> Genex-dev mailing list
> Gen...@li...
> http://lists.sourceforge.net/lists/listinfo/genex-dev

-- 
Cheers,
Harry

Harry J Mangalam -- (949) 856 2847 (v&f) -- hj...@nc... || man...@ho...

[GeneX-dev] Question on CyberT changes

[Michael P. <mic...@ho...>]

Harry,

I see in your change log for Cyber T a notation for V1.0.2


  * slight mod of the form to disallow Bayes analysis on data from the DB as
    there's no way to provide the expression levels required.  Only can do Bayes
    with uploaded data at this point.

As we really want to use CyberT from the database, what does this mean?
Does this affect the paired and control vs experiment modes of Cyber T?
Could you be more specific regarding the issue of "expression levels" vs.
"magnitude of the ratios" which apparently prompted some fixes?

Regards,

Michael Pear

Re: [GeneX-dev] Trouble with genex installation

[Michael P. <mic...@ho...>]

Genex.pm should be in a sub directory "Bio" in one of the perl lib
directories, which should automagically be created during the install
step. Is it there or actually directly in "/usr/local/lib/perl5/5.00563"?

Regards,

Michael Pear

----- Original Message ----- 
From: "Silvera, Damian" <DSi...@me...>
To: <gen...@li...>
Sent: Thursday, May 03, 2001 4:21 PM
Subject: [GeneX-dev] Trouble with genex installation


> Good evening,
> 
> my name is Damian Silvera. I am contacting you from the University of Miami
> School of Medicine. We have implemented (or are trying to) GeneX on a Sun
> Ultra 1 running Solaris 7 (a secure, hardened version of the OS). At
> present, we have all the prerequisites installed (I am working on a web page
> that will step through the recipe) and are ready to run the install-all.pl
> script. We are able to get through 90% of the script but get stuck at the DB
> Updater part. We are getting the following error:
> 
> Can't locate Bio/Genex.pm in @INC (@INC contains:
> /usr/local/lib/perl5/5.00563/sun
> 4-solaris /usr/local/lib/perl5/5.00563
> /usr/local/lib/site_perl/5.00563/sun4-solar
> is /usr/local/lib/site_perl .) at ./db-update-0.8.1-0.9.pl line 13.
> BEGIN failed--compilation aborted at ./db-update-0.8.1-0.9.pl line 13.
> 
> The Genex.pm perl module is in the /usr/local/lib/perl5/5.00563. I can't see
> why the script is dying, however. Any ideas?
> 
> _______________________________________________
> Genex-dev mailing list
> Gen...@li...
> http://lists.sourceforge.net/lists/listinfo/genex-dev
> 

[GeneX-dev] Trouble with genex installation

[Silvera, D. <DSi...@me...>]

Good evening,

my name is Damian Silvera. I am contacting you from the University of Miami
School of Medicine. We have implemented (or are trying to) GeneX on a Sun
Ultra 1 running Solaris 7 (a secure, hardened version of the OS). At
present, we have all the prerequisites installed (I am working on a web page
that will step through the recipe) and are ready to run the install-all.pl
script. We are able to get through 90% of the script but get stuck at the DB
Updater part. We are getting the following error:

Can't locate Bio/Genex.pm in @INC (@INC contains:
/usr/local/lib/perl5/5.00563/sun
4-solaris /usr/local/lib/perl5/5.00563
/usr/local/lib/site_perl/5.00563/sun4-solar
is /usr/local/lib/site_perl .) at ./db-update-0.8.1-0.9.pl line 13.
BEGIN failed--compilation aborted at ./db-update-0.8.1-0.9.pl line 13.

The Genex.pm perl module is in the /usr/local/lib/perl5/5.00563. I can't see
why the script is dying, however. Any ideas?