See also: IRC log
<Susie> SenseLab URL - http://senselab.med.yale.edu/SenseLab Public Snapshot 20071119.zip
<scribe> scribe: ericP
<Susie> SenseLab Conversion into Access tables - Luis Marenco (zip file)
<Susie> SenseLab URL - http://senselab.med.yale.edu/SenseLab Public Snapshot 20071119.zip
-> http://senselab.med.yale.edu/ SenseLab Homepage
-> http://senselab.med.yale.edu/site/dsArch/?lb=tree SenseLab Archicture Page
<mscottm> yes
-> http://senselab.med.yale.edu/site/dsarch/images/visio-eavcr_physical_schema_021205.png SenseLab ER diagram
Luis: pink boxes, internal
tables, are there now
... blue box also done
... senselab_db table corresponds to this diagram
... generally, remove the "senselab_"
... re senselab_classes (called "classes" in the ER):
class_db_owner tells you to which db that class belongs
... class_type tells you the type of class
... .. E for entity
... .. A for association
... A is just a bridge class
alanr: example?
luis: later on
... ER dump is from Oracle
... will add links to access database
... attributes class includes the attributes defined for each
class
... attribute caption is a human-readable name
... attribute description
... attribute datatype:
... .. S string -- up to 255 characters
... .. N numeric -- any real/integer/whatever
... . M memo -- up to two G
... .. can have as many memo fields as you want for a
particular table
... .. D datetime -- short or long
... .. Y binary -- binary string stored in the DB
ericP: both M memo and Y binary are BLOBs?
luis: depends on impl
susie: you are using Oracle 10G?
luis: yes
alanr: distinction between memo and binary?
luis: [@@ missed]
kei: you lost binary fields?
alanr: lost them in the MySQL,
but preserved in the XML
... if you do a separate query fro the memofields and save them
as, say, Base64...
susie: careful, LONG is likely to be lost from the Oracle support
luis: have lots of legacy
code
... I integer
... C object reference
... .. H for hierarchy
... attributes can have 1:1 or 1:many mapping
<alanr> "DATATYPE" "NAME" "DESCRIPTION" "XML_DATATYPE" "TABLE_NAME"
<alanr> "C" "Object/Ref" "Object reference (pointer)" "xs:long" "EAV_OBJECTS"
luis: hierarchy is a variation of C, but used for references to the same class
<alanr> "D" "Datetime" "Datetime" "xs:dateTime" "EAV_DATE"
<alanr> "F" "Formula(only)" "Formula (only)" "xs:anyType" "EAV_STRING"
<alanr> "H" "Hierarchy(p)" "Hierarchy (to parent object)" "xs:long" "EAV_OBJECTS"
<alanr> "I" "Integer" "Integer" "xs:long" "EAV_INT"
<alanr> "M" "Memo" "Free Text" "xs:string" "EAV_MEMO"
<alanr> "R" "Real" "Floating point values" "xs:double" "EAV_REAL"
<alanr> "S" "String" "String" "xs:string" "EAV_STRING"
<alanr> "V" "Object/Value" "Object/Value (elaborate)" "xs:long" "EAV_OBJECTS"
<alanr> "Y" "Binary" "Any binary string (files)" "xs:base64Binary" "EAV_BINARY"
<alanr> "B" "Boolean" "Boolean" "xs:boolean" "EAV_INT"
<alanr> hierarchy H can omit the class compared to C. Same class as referrer
luis: examing objects table:
<alanr> SENSELAB_OBJECT - Headers: "OBJECT_CLASS" "OBJECT_ID" "OBJECT_NAME" "OBJECT_DESCRIPTION" "CREATED" "MODIFIED_BY" "PRIVATE" "OWNER" "ASSOCIATION" "VER_NUMBER" "VER_DATE"
luis: you only get the entries
where OBJECT.PRIVATE=FALSE
... VALUEs tables:
... .. EAV_STRING table
<alanr> SENSELAB_EAV_STRING: "OBJECT_ID" "ATTRIBUTE_ID" "VALUE_"
luis: .. real, date, memos...
<alanr> e.g. "4085" "158" "Neuroscience"
luis: serial number allows
ordering of 1:many references
... example: a serial number will tell you in which order you
wish to see the receptors of a neuron
<alanr> SENSELAB_EAV_OBJECT, serial number - when there are multiple objects, how to order them.
<alanr> For display onyl
<alanr> SENSELAB_SEMRELA : "ID" "NAME" "INVERSE" "UMLS_ID" "ABR"
luis: semantic table start with senselab_
<alanr> "0" "(none) Top"
<alanr> "1" "is a" "T186" "IS"
<alanr> "2" "is of type"
<alanr> "3" "is named"
<alanr> "4" "is also named"
<alanr> "5" "described in Genbank as"
<alanr> "6" "whose gene is on"
<alanr> "7" "information obtained from"
<alanr> ...
<alanr> "20" "Biological Properties"
<alanr> "19" "is expressed in"
alanr: "biological property" is a parent of "expressed in"
<alanr> SENSELAB_SEMRELA_HIERARCHY:
<alanr> "PARENT_ID" "CHILD_ID" "STN_RTN"
<alanr> ....
<alanr> "20" "19"
<alanr> ....
luis: we have XML files with
<semantics/><schema/><data/>
... wanted this readable in the future
... semantic relationship attributes can connect two
attributes
<alanr> SEMRELT: "ID" "NAME" "UMLS_NAME" "UMLS_ID"
luis: semantic relationship types: modeled the same as the UMLS
<alanr> "ID" "NAME" "UMLS_NAME" "UMLS_ID"
<alanr> "1" "Class qualifies Attribute" "1st concept qualifies 2nd" "QB"
<alanr> "2" "Class is parent of Attribute" "1st is parent of 2nd" "CHD"
<alanr> "3" "Attribute has relationship other than synonomous" "2nd concept has relationship other than synonomous" "RO"
<alanr> "4" "Attribute is broader than Class" "2nd concept is broader than 1st" "RB"
<alanr> "5" "Attribute is narrower than Class" "2nd concept is narrower than 1st" "RN"
<alanr> "6" "Attribute qualifies Class" "2nd concept qualifies 1st" "AQ"
<alanr> "7" "Attribute is parent of Class" "2nd is parent of 1st" "PAR"
<alanr> "8" "sibling relationship" "sibling relationship" "SIB"
<alanr> "9" "Similar or alike concepts" "Similar or alike concepts" "RL"
luis: semantic type attribute:
mammal, human, ...
... is many to many
... dictionary table:
<alanr> DICTIONARY:
<alanr> "ELEM_TYPE" "ELEM_ID" "UMLS_ID" "UMLS_NAME"
<alanr> "A" "3" "C0441469"
<alanr> "O" "3542" "C0030708"
<alanr> "C" "19" "C0026336"
<alanr> ...
luis: .. type: object, attribute, class
<alanr> key into corresponding table
luis: attributes ID points to PK
of object, attribte or class
... no one used universal ID
alanr: will be simpler in RDF 'cause these will be URIs
luis: we may have URIs for these [objects, attributes, classes]
-> http://senselab.med.yale.edu/site/a_raf/controls/treeFrame.asp?fs=/site/dsArch/wAdmTree.asp whatever
[expanding Schema Manager/Databases/neurondb]
<alanr> http://senselab.med.yale.edu/site/dbMeta/eavMD_Classes.asp?cl=18
<alanr> http://senselab.med.yale.edu/site/dbMeta/eavMD_Classes.asp?cl=25
-> http://senselab.med.yale.edu/site/dbMeta/eavMD_Attributes.asp?cl=25 attributes for NuuronDB ncr
<alanr> property_present 0 = negation
luis: in http://senselab.med.yale.edu/site/dbMeta/eavMD_Classes.asp?cl=25 , Extensds Class should include Neuron
<alanr> association class = extra properties of a class. For sharing among several classes. Kind of a mixin?
Susie: Luis, thanks for this tour
<matthiassamwald> http://esw.w3.org/topic/HCLS/Senselab_Conversion?action=show
matthiassamwald: recently
expanded Experiences with the conversion of SenseLab databases
to RDF/OWL
... added links to source files
... and liks to the ontology files
... still missing:
... .. some snapshots of the dev process
... .. last automated RDF conversion
... in which form coud this description become and official
HCLS document?
<alanr> IG Note
<alanr> Magic HTML format
<mscottm> http://www.w3.org/2005/07/pubrules
susie: some W3C docs hilight the format for publishing IG Notes (and others)
alanr: mscottm will own the HTML template
<mscottm> http://www.w3.org/Guide/transitions
-> http://www.w3.org/Guide/transitions Organize a Recommendation Track Transition
<matthiassamwald> http://esw.w3.org/topic/HCLS/Semantic_Web_Step_By_Step
-> http://esw.w3.org/topic/HCLS/Semantic_Web_Step_By_Step Step by step: Creating interoperable Semantic Web resources
matthiassamwald: how to phrase tutorial
alanr: seems safer to say "here's how we did it"
matthiassamwald: may be worth saying "this is the best way"
next meeting: next week
susie: will discuss documentation on the next call
This is scribe.perl Revision: 1.128 of Date: 2007/02/23 21:38:13 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/EAV:OBJECT/SENSELAB_OBJECT/ Succeeded: s/semantic_/senselab_/ Found Scribe: ericP Inferring ScribeNick: ericP WARNING: No "Present: ... " found! Possibly Present: Alan_Ruttenberg DICTIONARY MIT-G346 P4 SEMRELT SENSELAB_EAV_STRING SENSELAB_SEMRELA_HIERARCHY Susie alanr ericP kei luis matthiassamwald mscottm You can indicate people for the Present list like this: <dbooth> Present: dbooth jonathan mary <dbooth> Present+ amy WARNING: No meeting title found! You should specify the meeting title like this: <dbooth> Meeting: Weekly Baking Club Meeting WARNING: No meeting chair found! You should specify the meeting chair like this: <dbooth> Chair: dbooth Got date from IRC log name: 26 Nov 2007 Guessing minutes URL: http://www.w3.org/2007/11/26-BioRDF-minutes.html People with action items:[End of scribe.perl diagnostic output]
