[OTDev] Proposal: List of Tasks for the implementation of QPRF/QMRF editors and their integration in OpenTox
Martin Guetlein martin.guetlein at googlemail.comMon Jul 12 13:57:40 CEST 2010
- Previous message: [OTDev] Proposal: List of Tasks for the implementation of QPRF/QMRF editors and their integration in OpenTox
- Next message: [OTDev] R Web Interfaces
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
On Mon, Jul 12, 2010 at 1:04 PM, Nina Jeliazkova <jeliazkova.nina at gmail.com> wrote: > On Mon, Jul 12, 2010 at 1:16 PM, chung <chvng at mail.ntua.gr> wrote: > >> Hi Martin, >> >> >> On Mon, 2010-07-12 at 10:40 +0200, Martin Guetlein wrote: >> >> > On Wed, Jul 7, 2010 at 12:24 PM, chung <chvng at mail.ntua.gr> wrote: >> > > Hi Martin, All, >> > >> > Hi Pantelis, >> > >> > > I more or less agree that we should restrict ourselves into >> > > providing >> > > only some absolutely necessary functionalities for the time and >> > > extend >> > > the editor(s) later. I'd like to outline my perception about the GUI >> > > design >> > > and the use case behind that. >> > > >> > > --> First of all the GUI is not much more than a decorated >> > > user-friendly web client that exploits the QPRF reporting >> > > web service to create/store a QPRF report according to >> > > the JRC specifications. >> > > >> > > I'd like to go into some further detail about the use case and >> > > user requirements about the editor so I summarize a list of >> > > must-have and should-have functionalities for the first version >> > > of the editor: >> > > >> > > MUST-HAVE FEATURES: >> > > 1. Create an empty report (This is the easiest thing to do...). >> > > The user fills out all fields needed for the QPRF report >> > > (see JRC specifications). >> > > Once the user has completed report editing he can: >> > > i. Store it locally (in RDF format) >> > > ii. Store it on a remote server and acquire a report URI. >> > > ii. Export it as PDF (The pdf creation is done using the QPRF >> > > reporting service) >> > >> > You are actually right, that a user could start with a blank report in >> > the editor. From my point of view, the normal use case would be that a >> > report is created via report webservice, and can be viewed/edited with >> > the editor. >> >> Why should the web service create an empty report? I mean this is more >> or less some >> static information. I don't see any obvious reasons for having it >> deployed on the web. >> > > It make sense to have web service functionality to > 1)create empty report > 2)have different sections of the report exposed by URIs > 3)update different sections of the report (by PUT / POST to the sections > URIs) This sounds sensible, and should work rather straightforward, just following the report-(xml)-structure. For example the title for qmrf could be accessed via: http://<report-service>/qmrf/<report-id>/QMRF_chapters/1/QSAR_title This should work as soon as rdf is supported as well. Regards, Martin > > > > Just my 2 cents. > > Nina > >> >> > Again, I do not think RDF is a must have feature (rather than the >> > 'official' no yet defined format). >> >> OK, then let's use the existing XML for the QPRF report. >> >> Best regards, >> Pantelis >> >> > >> > > >> > > 2. User can load a report from the remote server providing the report's >> > > URI. >> > > The report is downloaded in RDF format and it's content is >> > > displayed >> > > in a user-friendly manner. In no way does the user interact with >> > > RDFs. >> > > >> > > 3. Once the user has completed editing a report he can either store it >> > > locally >> > > or in a remote server. In the latter case, a report URI is returned >> > > to the >> > > user. The report is stored in RDF. >> > > >> > >> > Same comments as for point 1. >> > >> > >> > > EXTENSIONS: >> > > >> > > 1. User creates a report providing a compound and/or a model. >> > > The compound can be identified by its name, smiles or even >> > > its URI. The user can choose a model from a list (using the >> > > ontology >> > > service) or provide the model URI manually. Credentials might >> > > need to be provided. User credentials are stored in the current >> > > session so that the user won't have to re-enter them. >> > >> > I agree, we should provide a nice GUI-supported way to search and >> > browse the existing reports. >> > >> > > >> > > >> > > On Mon, 2010-07-05 at 17:33 +0200, Martin Guetlein wrote: >> > > >> > >> On Mon, Jul 5, 2010 at 11:53 AM, Pantelis Sopasakis < >> chvng at mail.ntua.gr> >> > >> wrote: >> > >> > Dear all, >> > >> > I'm sending you the brief list of tasks I also presented in today's >> > >> meeting >> > >> > to comment on if you like: >> > >> > >> > >> > Current State of Development: >> > >> > >> > >> > A. QPRF Editor >> > >> > * Initial design of the GUI is available at: >> > >> > http://opentox.ntua.gr/Q-edit/launch.html >> > >> > * Source code available at: >> > >> > http://github.com/alphaville/Q-edit >> > >> >> > >> Hi Pantelis, All, >> > >> >> > >> nice work, you already have been very busy. I had a look at your >> editor and >> > >> a discussion with Andreas Karwath, and we had the impression that your >> > >> editor might be step too far ahead. We do not have much time left, so >> maybe >> > >> we first should agree on the functionality of the OpenTox REACH report >> > >> service and the report editors. >> > >> >> > >> That's what we had in mind (the report service api is defined >> accordingly): >> > >> >> > >> REACH report service features: >> > >> -------------------------------------------- >> > >> * create report >> > >> - internally collect meta-information, >> > > >> > > By 'internally' you mean information retrieval from remote servers? >> > >> > right >> > >> > > >> > >> make validations >> > > >> > > You mean by model validations? >> > >> > right >> > >> > > >> > >> - leave fields that cannot be filled out automatically (need more >> user >> > >> input) >> > > >> > > Indded, there are quite a lot such meta information (e.g. Comments on >> > > the uncertainty of the prediction). >> > > >> > >> * provide already existing qmrf / qprf reports >> > > >> > > Are there any existing QPRF reports in a format different from PDF? >> > >> > I meant that the reports are stored on the server (in an internal >> > format, supported formats for the user are 'official'-xml format, >> > probably rdf. conversion to pdf/other formats will be done in the >> > editors). >> > >> > > >> > >> * save changes to existing reports >> > >> * provide reports in the following formats: >> > >> - MUST HAVE: format accepted by the EC (for qmrf: existing xml, for >> qprf: >> > >> most likely analogous format) >> > > >> > > >> > > As I mentioned in our previous virtual meeting, it might seem more >> > > difficult to work with RDF rather that with some XML of ours but it is >> > > both easier and also better practice than XML. Existing RDF >> > > representations will be easily embedded into the RDF representation of >> > > the QPRF report. The parsing will be carried out using Jena. Also, >> using >> > > RDF uniformness is retained. >> > >> > I know the advantages of rdf, however what we must have at the end is >> > the official EC format, rdf would be nice to have for OpenTox internal >> > usage. >> > However, as stated before, we decided that we will start with rdf as >> > format for QPRF. >> > >> > > >> > > >> > >> - EXTENDED FEATURE: opentox internal rdf-format >> > >> >> > >> QMRF/QPRF editors: >> > >> -------------------------------- >> > >> MUST HAVE: >> > >> * open the report in the EC-accepted format >> > >> * edit single fields by hand >> > >> * save report in EC-accepted format >> > >> * export functions supporting document formats like PDF >> > >> * QPRF and QMRF editors should have a similar GUI and must be easy to >> use >> > >> EXTENDED FEATURES: >> > >> * support opentox internal rdf-format >> > >> * init creation of new report >> > > >> > > I think creation of a new report is something quite easy and >> > > straightforward so it could be a must-have one. >> > >> > What I meant, was that it should be possible to initialize the >> > creation of new report at the webservice (by selecting a model / >> > compound, and sending the post command to the web services). >> > >> > > >> > >> * browse existing reports >> > >> * communicate with various opentox web services to automatically fill >> out >> > >> additional fields >> > > >> > > Such as? >> > >> > For example, the Applicability domain section in the QPRF report is >> > quite detailed (3.3-a-ii, what if the model doesn't use any structural >> > fragments?). >> > We could leave those sections empty by default, but give the user some >> > guidance if he wants to make more experiments. >> > >> > > >> > >> >> > >> In the last goto-meetings we agreed that we could go "rdf-first" for >> the >> > >> QPRF-report and -editor (because it fits into our framework), but we >> should >> > >> keep in mind that the format accepted by the European Commission will >> > >> probably be similar to the QMRF format. >> > >> For the QMRF, we decided to get it working with the existing qmrf-xml >> format >> > >> first (so that its working in Rhodes), and provide the rdf format >> later on. >> > >> >> > >> Do people agree on the design of the service and editors? Please >> provide >> > >> some feedback, we could than finalize a TASK/TODO list to be assigned >> on >> > >> Friday. >> > > >> > > >> > > We could have a draft todo list by tomorrow to be discussed and >> > > finalized on Friday. >> > >> > As we already have decided on todo list some time ago, maybe we could >> > adjust this list to the new requirements (and shift partner >> > responsibilities if necessary), instead of coming up with a totally >> > new one? >> > I added the old list to the end of the mail. >> > >> > Best regards, >> > Martin >> > >> > P.S.: Sorry for answering late, I was away last week for few days. >> > >> > >> > * by 07.05.2010: >> > Definition of the API >> > This has already be done by Nina (thanks a lot for that!) and can >> > be found here (http://www.opentox.org/dev/apis/api-1.1/Validation) - >> > and Martin and I concur. Probably, others might want to comment (it >> > still depends on future work regarding automatic filling of parameters >> > - see below). >> > The API would allow as parameters a Model URI, a List of >> > (cross-)validation URIs (for the same model) and additional parameters >> > (which cannot be filled automatically, as well as manually constructed >> > reports in qmrf-xml and a to be defined qprf format. That would mean, >> > that complete reports can be stored on the OpenTox servers. >> > >> > * by 01.06.2010: >> > ALU-FR: QMRF: Identify possible attributes for automatic >> preconstruction >> > TUM: QPRF: Identify possible attributes for automatic preconstruction >> > (in close collaboration with NTUA) >> > >> > * by 22.06.2010: >> > ALU-FR: QMRF: Derive data from web services automatically, construct >> > XML for Editor >> > NTUA and SL: QMRF :Adapt QMRF Editor (Java Webstart, OpenTox design) >> > from >> https://ambit.svn.sourceforge.net/svnroot/ambit/trunk/ambit1/src/ambit/data/qmrf >> > (in collaboration with IBMC) >> > NTUA and SL: QPRF : Define (most likely a XML schemata) format for >> > QPRF (in close collaboration with TUM and Ex-ECB (Andrew needs to be >> > contacted)) >> > >> > * by 13.07.2010: >> > IBMC: QMRF: Integrate QMRF-Editor into OpenTox Web services (Get >> > user input, initialize model building and validation, Start QMRF >> > editor) >> > ALU-FR: QPRF : Derive data from web service automatically, construct >> > XML for Editor >> > >> > * by 03.08.2010: >> > NTUA and SL: QPRF: Based on Nina’s QMRF implementation, implement a >> > QPRF editor (including Java-Webstart and OpenTox design) >> > IBMC: Integrate QPRF-Editor into OpenTox Webservices (Get user >> > input, initialize model building and validation, Start QPRF editor) >> > _______________________________________________ >> > Development mailing list >> > Development at opentox.org >> > http://www.opentox.org/mailman/listinfo/development >> > >> >> >> _______________________________________________ >> Development mailing list >> Development at opentox.org >> http://www.opentox.org/mailman/listinfo/development >> > _______________________________________________ > Development mailing list > Development at opentox.org > http://www.opentox.org/mailman/listinfo/development > -- Dipl-Inf. Martin Gütlein Phone: +49 (0)761 203 8442 (office) +49 (0)177 623 9499 (mobile) Email: guetlein at informatik.uni-freiburg.de
- Previous message: [OTDev] Proposal: List of Tasks for the implementation of QPRF/QMRF editors and their integration in OpenTox
- Next message: [OTDev] R Web Interfaces
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Development mailing list