- Zope - Zope lists

RDFSummary weirdness
by David Chandek-Stark 16 Sep '04

16 Sep '04

I am using RDFSummary 2.4p1 on Zope 2.7.0/CMF 1.4.7. I have noticed for along time this product exhibits strange behavior where -- without updating the RSS feed -- the data changes on successive calls. It seems to flip back and forth between the most recent data and the last (cached?) version. Any ideas? -- David Chandek-Stark Web Applications Developer Duke University - Perkins Library (919) 660-5859 dc(a)duke.edu

1 0

Job Posting - Plone/Zope/Python/XML Developer Wanted
by Paul Allen 16 Sep '04

16 Sep '04

(From http://birds.cornell.edu/About/jobs.html) Python Application Programmer Programmer/Analyst III - Posting #02932) DESCRIPTION: The Cornell Laboratory of Ornithology (CLO) is seeking a Programmer/Analyst to collaborate with team members to develop and maintain the Scientific Knowledge and Education Network (SKEN). SKEN will likely be created in the Plone/Zope content management system (CMS) environment, which was developed in the Python programming language. The goal of SKEN is to build an open-source infrastructure for creating knowledge and education networks centered on primary scientific references. Special attention will be given to integration with the National Science Digital Library system, of which the CLO has recently become a member. Team up with members of CLO's IT unit and the editors of the Birds of North America Online, a comprehensive reference covering the life histories of North American birds, to create SKEN. Conceptualize and design the SKEN infrastructure and develop an open-source platform from which scientific communities can be formed around primary references. This is a two year term appointment with possibility of renewal. REQUIREMENTS: Bachelors with 2-3 years experience or equivalent combination. A thorough understanding of object-oriented programming is essential, with knowledge and experience in Python and Java programming preferred. Working knowledge of XML technologies and programmatic XML document manipulation. Demonstrable skills in problem solving, critical thinking, and written and oral communication. Ability to communicate technical language in layman's terms. Illustrate proven record of successfully experience writing application requirements, architecting applications, and documenting code. Demonstrated ability to design and construct substantial computer software. Must work effectively on a team, and learn technical material quickly with the ability to prioritize multiple projects. Establish realistic goals and deliver work by scheduled deadlines. Ability to handle confidential information. Ability to work well with a diverse group in a professional, respectful work environment. Work independently while supporting the efforts of an overall team. Experience working in Plone/Zope environment is preferred. 5-10 years of experience working on substantial computer software and collaborating with a team of programmers on software projects is desired.

1 0

ZCatalog Performance problem
by Jonathan Hobbs 16 Sep '04

16 Sep '04

I am trying to install a new version of our database and have encountered some performance problems. We are using: Zope 2.6.1 Python 2.1 Linux2 ZCatalog 2.2.0 ZCTextIndex (with Lexicon) The 'old' Data.fs file contains 1,830,854 ('objects in the database'). Retrieval times, based on CallProfiler, are typically < 2 secs. The ZCatalog contains 671,873 entries, and the actual documents are stored in two standard Folders. The Lexicon contains 605,984 words. The 'new' version uses DBTab to mount a second .fs file which contains the ZCatalog and corresponding documents (which are now stored in two BTreeFolder2 type folders). The .fs file contains 7,670,410 ('objects in the database'). Retrieval times, based on CallProfiler, are are in the range 1 to 14 secs. The ZCatalog contains 782,216entries. The Lexicon contains 1,741,003 words. In both versions, the Lexicon 'Input Pipeline Stages' are the same (Splitter, CaseNormalizer, StopWordAndSingleCharRemover). In the 'old' version, the documents stored in the ZCatalog (zclass instances) are 'catalog-aware'. In the 'new' version they are not. Also, in the 'old' version the complete full text of the document was compressed and stored in the metadata table (this was eliminated in the 'new' version to reduce the size of the metadata table). Because of the very slow retrieval times, I installed ZopeProfiler and here is an extract of 'typical' results: ncalls tottime percall cumtime percall filename:lineno(function) 2 0.000 0.000 11.462 5.731 Catalog.py:717(searchResults) 2 2.069 1.035 11.461 5.730 Catalog.py:446(search) 1 0.216 0.216 9.320 9.320 OkapiIndex.py:120(_search_wids) 2 0.000 0.000 9.391 4.696 ZCTextIndex.py:167(_apply_index) 1 0.000 0.000 9.390 9.390 ParseTree.py:117(executeQuery) 1 0.000 0.000 9.390 9.390 BaseIndex.py:169(search) 1 0.216 0.216 9.320 9.320 OkapiIndex.py:120(_search_wids) 2085 0.239 0.000 9.174 0.004 Connection.py:499(setstate) 2085 0.045 0.000 8.761 0.004 FileStorage.py:647(load) 2085 8.510 0.004 8.716 0.004 FileStorage.py:620(_load) 2085 0.151 0.000 0.181 0.000 fsIndex.py:61(__getitem__) This profile was generated using a very simple script to access the zcatalog: <dtml-let res="Catalog.searchResults({'all_searchable_text' : all_searchable_text})"> <b>Done Search</b><br> </dtml-let> It appears that the delay is due to the time required to load the zctextindex into memory. This is supported by the fact that subsequent searches for the same search term take less than 1 second and do not generate any FileStorage '_load's, and running iostat during the search shows LOTS of disk activity (cpu usage is minimal). The cache 'target' is set at 100,000 and currently has about 59,000 objects (61,000 ghost objects). So, finally, the questions: 1) What would cause the Lexicon to jump from 605k words to 1741k words (a 187% increase), even though the zcatalog entries only grew from 671k to 782k (a 16.5% increase)? 2) What would cause the .fs file to jump from 1.8m entries to 7.6m entries (a 322% increase), even though the zcatalog entries only grew from 671k to 782k (a 16.5% increase)? 3) Is there any way to preload the entire zctextindex? Any and all ideas, comments and suggestions will be greatfully received! Jonathan

2 1

Call Zope objects using wget
by Paolo Dina 16 Sep '04

16 Sep '04

Hi all. I need to call a Python script (that simply creates a new folder object with a unique name) via wget. When I manually invoke the command using via shell, wget -d --spider --http-user=myuser --http-passwd=mypassword http://my.url.it/intranet/reset_portal_content/ZScriptHTML_tryForm the folder object isn't created, while calling the script via Test tab from the ZMI works correctly. The thing that makes me dubious against this misbehaviour, is the server response "302 Moved Temporarily". This is the debug message generated by wget -d: DEBUG output created by Wget 1.9.1 on linux-gnu. --12:48:15-- http://my.url.it/intranet/reset_portal_content/ZScriptHTML_tryForm => `ZScriptHTML_tryForm' Resolving my.url.it... www.xxx.yyy.zzz Caching my.url.it => www.xxx.yyy.zzz Connecting to my.url.it[www.xxx.yyy.zzz]:80... connected. Created socket 3. Releasing 0x8083c10 (new refcount 1). ---request begin--- HEAD /intranet/reset_portal_content/ZScriptHTML_tryForm HTTP/1.0 User-Agent: Wget/1.9.1 Host: my.url.it Accept: */* Connection: Keep-Alive Authorization: Basic YWRtaW46a29hbG9uZQ== Pragma: no-cache ---request end--- HTTP request sent, awaiting response... HTTP/1.1 302 Moved Temporarily Date: Wed, 15 Sep 2004 10:42:09 GMT Server: Zope/(unreleased version, python 2.3.3, linux2) ZServer/1.1 CPS/3.0 Bobo-Exception-Line: 56 Bobo-Exception-Value: See the server error log for details Bobo-Exception-File: DT_Raise.py Bobo-Exception-Type: Redirect Etag: Location: http://my.url.it/intranet/reset_portal_content Content-Type: text/html;charset=iso-8859-1 Connection: close Location: http://my.url.it/intranet/reset_portal_content [following] Closing fd 3 --12:48:15-- http://my.url.it/intranet/reset_portal_content => `reset_portal_content' Found my.url.it in host_name_addresses_map (0x8083c10) Connecting to my.url.it[www.xxx.yyy.zzz]:80... connected. Created socket 3. Releasing 0x8083c10 (new refcount 1). ---request begin--- HEAD /intranet/reset_portal_content HTTP/1.0 User-Agent: Wget/1.9.1 Host: my.url.it Accept: */* Connection: Keep-Alive Authorization: Basic YWRtaW46a29hbG9uZQ== Pragma: no-cache ---request end--- HTTP request sent, awaiting response... HTTP/1.1 200 OK Date: Wed, 15 Sep 2004 10:42:10 GMT Server: Zope/(unreleased version, python 2.3.3, linux2) ZServer/1.1 CPS/3.0 Accept-Ranges: none Ms-Author-Via: DAV Last-Modified: Wed, 15 Sep 2004 06:51:45 GMT Etag: Content-Type: application/octet-stream Connection: close Length: unspecified [application/octet-stream] Closing fd 3 200 OK From Zope mailing lists I have seen that many Zope users have never had problems using wget with Zope.. even if someone has suggested to use zeo + wget in case of problems. Any pointer to solve this issue would be highly appreciated. TIA, Paolo

3 3

Re: [Zope] passing parameter problem
by Alexis Roda 16 Sep '04

16 Sep '04

Laura McCord wrote: > I tried what you suggested and that did not work either. actStatus will > not be null. > tal:define="actStatus result/Active_Inactive" returns either active or > inactive and that is working correctly. From what I know from your page template it is equivalent to: <div tal:define="actStatus template/title; colorStat python:actStatus + ' and this is foo'" tal:content="colorStat"> </div> get a value from somewhere and calculate a second value based on first value. That works ok so your problem should be with actStatus initialization or with statChk script. > This is what my python script looks like which includes a parameter, > actStatus: > if actStatus == 'active': > return "#000" > if actStatus == 'inactive': > return "#b7b7b7" If actStatus is not 'active' nor 'inactive' the script will return None. IIRC None is tal Nothing and you're using colorStat in a tal:attributes, which interprets Nothing as "remove the attribute". Try adding a raise ValueError('actStatus="%s"' % repr(actStatus)) statement at the end of the script, just to be sure that the script gets a valid actStatus and returns a color. If statChk is just the four lines you send it could be replaced by something like: colorStat python:test(actStatus == 'active', '#000', '#b7b7b7') HTH -- //// (@ @) ----------------------------oOO----(_)----OOo-------------------------- <> Ojo por ojo y el mundo acabara ciego /\ Alexis Roda - Universitat Rovira i Virgili - Reus, Tarragona (Spain) -----------------------------------------------------------------------

1 0

Keyword indexes
by Garito 16 Sep '04

16 Sep '04

Hi all I have a problem with Keyword index search If I search something like context.Catalog({'keywordindex': ['Value1', 'Value2']}) Catalog returns me an OR search instead of an AND one I find on the web that this is a know bug of keywordindex In the same message someone recomends managableindex from Dieter Maurer but when I try to install it I could see that OFolder is a requierement I need to reduce the number of installeds products For that, these is my question: I need some ZCatalog Index with these features: TAL Expresions (like managebleindex or proxyindex) KeywordIndex capable Any recomendation? Thanks PD: If these is a know bug why don't you (Zope people) correct it? PD: To Dieter: If Zope has Ordered Folder why don't you update managebleindex to use it instead of your OFolder?

2 2

stylesheets
by John Poltorak 15 Sep '04

15 Sep '04

Does Zope come with stylesheets and if so where do I find them? -- John

2 1

Re: [Zope] passing parameter problem
by Laura McCord 15 Sep '04

15 Sep '04

I tried what you suggested and that did not work either. actStatus will not be null. tal:define="actStatus result/Active_Inactive" returns either active or inactive and that is working correctly. This is what my python script looks like which includes a parameter, actStatus: if actStatus == 'active': return "#000" if actStatus == 'inactive': return "#b7b7b7" I have been working on this for two days straight and it has me stumped. Any help or suggestions would be great. Thanks, Laura Laura McCord wrote: > It seems like I have done this hundreds of times and it worked. Can some > point out what I am doing wrong? > > <div tal:define="actStatus result/Active_Inactive; > colorStat python:here.statChk(actStatus)"> That seems correct to me. actStatus can be None? > I also tried this: > > <div tal:define="colorStat python:here.statChk('${actStatus}')"> IIRC you should user statChk(string('${actStatus}')) or statChk(path('actStatus')) HTH --

1 0

Re: DBTAB mounting ZCLASS as container
by Arenz, Ralph 15 Sep '04

15 Sep '04

Harald Koschinski wrote at 2004-1-13 10:36 +0100: >is there a way to mount a Database with a ZClass as the container object ? > >In theory there is the parameter "container_class" in dbtab.conf to >specify which class will taken. But as I can see it's only accepting >PythonClasses - not ZClasses. > >Any hints?? >This will not be possible without major work: > You would need to implement a new type of mount point. > Its class key would allow a ZClass reference. > The corresponding datatype would be a factory that > later instantiates the ZClass. -- >Dieter hi dieter, can you be a little more concrete. I spent a look over the sources from dbtab and found some fragments which gives a hint that dbtab can work with ZKlasses, but i can't figure out how do arrange this. Do i really have to implement a new mount point although i need a ZClass as a container for it? Ralph

2 1

Script Security
by Jason Leach 15 Sep '04

15 Sep '04

hi, Is it possible to have a script that can not be called by users: http://www.foo.com/myscript but can be called by other scripts? J.

2 1