Chat with Mentor

menu
Status #openmrsX surangaX #3X
[22:27] <Rahul> Hi suranga :)
[22:29] <suranga> heloo !
[22:29] <suranga> great to see u r intrested in patient matching !
[22:29] <Rahul> :)
[22:30] <suranga> hows the proposal going ?
[22:30] <suranga> you should post it to melange, that way we get to comment on it :-)
[22:30] <Rahul> i have some doubts about the task .. want to discuss with you :)
[22:31] <suranga> sure, go ahead !
[22:32] <Rahul> regarding the task #1 what would be the invalid matching config ?
[22:33] <suranga> aha, an invalid matching config is something that doesnt have at least one must match and one shod match
[22:33] <suranga> u need at least a must match and a should match for it to be valid
[22:33] <Rahul> oh okay okay :)
[22:34] <Rahul> should put validation when creating a stratergy rite?
[22:34] <suranga> yep, thats right.. any possibly display a good notification altering the use
[22:35] <suranga> *user
[22:35] <Rahul> yup totally fine with task #1 :)
[22:40] <Rahul> now regarding the task #2   i think there are 3 to 4 algorithms to find duplicate patients  the module ... which i can find the sloution rite ?
[22:40] <Rahul> * in the module
[22:43] <suranga> you're right... but in this case, we will be focussing on the web interface deelopment
[22:44] <suranga> so since the web interface uses the exact match algorithem, I think we should go withthat for a start...
[22:45] <Rahul> ya .. :) this would be great to avoid out of memory and browser crash and bandwidth problems :) by calculating the matched record pairs :)
[22:46] <suranga> yep, actually that idea came from Dr. Shaun himself...
[22:46] <Rahul> ok  .. Fisrt step in this Task #2 would be calculate the no of pairs
[22:47] <Rahul> seond : Alert the user the count of the matched patient records
[22:47] <Rahul> third display according to the end user options :) rite ?
[22:48] <suranga> exactly
[22:48] <suranga> actually, James tells me that this a relatively easy. Apparently, this is happening somewhre in the code, so all we need to do it to figure out where to access the data
[22:49] <suranga> I need to talk to james and find out exactly where it happens :-)
[22:49] <Rahul> oh cool :) ..
[22:49] <suranga> how are tasks 3-5 ? :)
[22:49] <Rahul> yup i am on the way to discuss with you ;)
[22:50] <suranga> sure, go ahead :)
[22:51] <Rahul> now regarding the task #3 , to convert the the xml file to Db  model ... would you provide a data model for it ...?
[22:52] <suranga> not really... the idea is that some of the data stored in the xml are duplicates, and dont really need to be stored.
[22:52] <suranga> we are already storing patientmatchingconfigurations in the db... what we need to do is to figure oiut a way how to store the entire reports as well...
[22:53] <suranga> did u see the page I have linked to ?
[22:53] <suranga>  it
[22:53] <suranga> s https://wiki.openmrs.org/display/docs/Patient+matching+-+Moving+flat+files+into+a+database
[22:56] <suranga> in a report, basically we need to store, 1) which configurations were used to create it
[22:56] <Rahul> cool :)  now the reports are storing in the Application/openmrs/patientmatchig/ some*/ de-deup-*   would convert into the Db rite ?
[22:56] <suranga> 2) some basic metadata
[22:57] <suranga> yep, thats right... we are not getting rid of the flat files right now, we just need to introduce a db so that we can work with this data
[22:57] <suranga> in that link, Ive attached a pic of some dummy data
[22:58] <suranga> basically in the case of the report, what u need to save is 1) the configurations used to create it
[22:58] <suranga> 2) multiple sets of similart patients that are potential matches
[22:58] <suranga> ) some metadata
[22:59] <Rahul> :) ya ...
[23:00] <Rahul> the configuration include shud match or must match field .. every run indicates a stratergy rite ?
[23:01] <suranga> sorry, did not understand you there...
[23:01] <suranga> a report can contain n many configurations :-)
[23:02] <Rahul> i mean the configuation i.e config.xml each run tag reprrsents a stratergy rite ?
[23:03] <Rahul> <run>*
[23:03] <suranga> oh yes, thats right
[23:03] <suranga> so in the xml, each congifurations keeps tab of which attributes have been set, and which are ignored...
[23:04] <suranga> they have a <include> tag to keep tract of wether they are being used or not...
[23:04] <Rahul> ok ya  :) ..
[23:05] <Rahul> ya :) now the first step for task #3  would be remove the flat file storing of reports and import it to the db rite ?
[23:08] <suranga> no, thats where the isssue is
[23:09] <suranga> see, the patientmatching module is also used as a standalone jar by people who dont know about OpenMRS
[23:09] <suranga> so if we were to remove the flat file entirely, they will be stuck with no way of using it
[23:10] <suranga> plus, peiople who stored their data in the file already are going to loose those too
[23:10] <suranga> so at the moment, the focus is to get flat file and database running in parallel
[23:10] <Rahul> oh yaa ... :) okay  :)
[23:10] <suranga> (as in configurations are both written into the xml and the database
[23:11] <suranga> later on, we are planning to do away with the flat file entirely, but at this stage we are not 100% of the path we will take :)
[23:11] <Rahul> okay :) ...
[23:13] <Rahul> now regarding the task #4 , what is the data quality and information content metrics for patientmatching module ? :)
[23:14] <Rahul> how data fields are characterized for the duplication stratergy ?
[23:14] <suranga> Rahul  its like thiis...
[23:14] <suranga> task 4 was already implemented by Dr. Shaun using the statistical module R
[23:15] <suranga> what needs to be done is to figure out how to use this code to work with the module itself, or to implement  something similar using java etc.
[23:16] <suranga> basically, you wont be asked to do something really complex here... we will probably be asking you to make use of some already generated data for this purpose...
[23:17] <Rahul> cool :) where is that module ? :)
[23:20] <Rahul> i mean statistical module has been implemented totlly in the pateintmatching module ?
[23:20] <suranga> Rahul its in either pairmet.R or sigmet.R files
[23:21] <suranga> no, not really, its that we have a few scripts that make use of the R module
[23:21] <suranga> (R module is a statistical software, its not related to OpenMRS :P)
[23:22] <suranga> James Egg is the ideal one to answer questions on task four, since he has some experiance with that one :-)
[23:23] <Rahul> yup seeing *.R file  :) .. a new language :)
[23:23] <Rahul> * scripting language
[23:23] <suranga> dont worry about R... we wont ask you to script using it ;D
[23:23] <suranga> I tried using R once, it was terrible :D
[23:24] <Rahul> :) :D ..
[23:24] <suranga> we will basically either ask you to 1. implement something similar with java, or 2. figure out a way to use the data retreived from this and use it in the module...
[23:25] <Rahul> ya  okay :) ...
[23:26] <Rahul> now regarding the task #5, Implement  Gui feature  (https://wiki.openmrs.org/display/docs/Patient+Matching+Module) in  web-interface rite ?
[23:31] <Rahul> Convert the swing application (recenlty i worked on swing appliation :) ...) into Jsp page  rite ?
[23:32] <suranga> Rahul  sorry I was away
[23:32] <suranga> thats right
[23:32] <suranga> have u tried out running the standalone ?
[23:32] <Rahul> no probs suranga  :)
[23:32] <suranga> the RecMatch.java class has a mail method
[23:32] <suranga> when u run that, u get a swing ui
[23:32] <Rahul> no not but saw the screenshorts in the wiki link :)
[23:32] <suranga> the ideas is that some of the featues in this should go into the webapp as well
[23:33] <suranga> oh, right, i see...
[23:33] <suranga> basically, you will just need to work on the front end of the code for these
[23:34] <suranga> they have been implemented (and used) by the standalone, so u will just need to call these methods etc. from the controlooer / jsp
[23:34] <Rahul> ya :) .... that's cool :)
[23:36] <Rahul> i am doubts are crystal clear :)
[23:38] <Rahul> thanks alot suranga  , now i will start writing the proposdal  :) ...
[23:39] <suranga> great !
[23:39] <suranga> im looking forward to it !:)
[23:39] <suranga> feel free to buzz me or James, we're happy to cosider your proposal :-)
[23:40] <Rahul> :) thanks ... :):)