[MOBY-guts] biomoby commit
Mark Wilkinson
mwilkinson at dev.open-bio.org
Sun Jan 12 18:40:00 UTC 2003
Sun Jan 12 13:39:59 EST 2003
Update of /home/repository/moby/moby-live/UseCases/mw/UseCases
In directory dev:/tmp/cvs-serv8907/UseCases
Modified Files:
AutoUpdate_Citation.html
Log Message:
removing Scenarios. now hosted in Twiki
moby-live/UseCases/mw/UseCases AutoUpdate_Citation.html,1.1,1.2
===================================================================
RCS file: /home/repository/moby/moby-live/UseCases/mw/UseCases/AutoUpdate_Citation.html,v
retrieving revision 1.1
retrieving revision 1.2
diff -u -r1.1 -r1.2
--- /tmp/T04haiBr 2003-01-12 13:39:59.490004807 -0500
+++ /tmp/T15haiBr 2003-01-12 13:39:59.490004807 -0500
@@ -2,10 +2,10 @@
<html>
<head>
<title>AutoUpdate_Citation</title>
-
+
<meta http-equiv="content-type"
content="text/html; charset=ISO-8859-1">
-
+
<meta name="author" content="Mark Wilkinson">
</head>
<body>
@@ -21,7 +21,7 @@
<p> </p>
<h3><a name="PrimaryActor_"> </a> <a href="/bin/view/VO/PrimaryActor">PrimaryActor</a>:</h3>
- Script running on a MOD host<br>
+script "agent" containing rules for determining new relevant citation information<br>
<br>
<p> </p>
@@ -31,7 +31,7 @@
<hr width="10%" align="left">
<h3><a name="EndResult_"> </a> <a href="/bin/view/VO/EndResult">EndResult</a>:</h3>
<ol>
- <li> automatic database inserts of new information relevant to that database.</li>
+ <li> automatic database inserts of new Citations relevant to that database.</li>
<li> alerts to MOD curator of potentially relevant information that must
first be checked by eye.</li>
</ol>
@@ -40,6 +40,9 @@
<h3><a name="OtherActors_"> </a> <a href="/bin/view/VO/OtherActors">OtherActors</a>:</h3>
<ol>
<li> MOD database running a MOBY interface</li>
+ <li>MOBY Central</li>
+ <li>External MOBY Services<br>
+ </li>
<li>Curator</li>
</ol>
@@ -63,7 +66,7 @@
journal articles, so other resources must be used</li>
<li>Curator does *not* want redundancy in his database</li>
<li>Curator has written MOBY Services running on DragonDB that:</li>
-
+
<ol type="A">
<li>accept Author(s) as input objects to generate relevant Citation objects</li>
<li>accept PubMed/ID(s) as input objects to generate Citation objects</li>
@@ -75,7 +78,7 @@
</li>
<li>accept Citations as input objects to write to database.<br>
</li>
-
+
</ol>
</ol>
@@ -92,83 +95,83 @@
<li>Curator writes script where the DragonDB MOBY Service ('C' above) is
queried with keyword "*" to recover all existing Citation objects in
his database; this call is hard-coded in the script.</li>
-
+
<ol>
<li>Citation objects from this service are in the namespace DragonDB/CitationID
and, where possible, carry cross-references to e.g. PubMed/ID's, or other
citation namespaces in the CrossReferences Block (CRIB)</li>
-
+
</ol>
- <li>Author Object are extracted from Citation objects (somehow; either directly
-in the script, or through another MOBY service)</li>
-
+ <li>Author Object are extracted from Citation objects (somehow; either
+directly in the script, or through another MOBY service)</li>
+
<ol>
<li>Citation HAS-A Author, as determined by the object hierarchy.</li>
<li>Citation objects are also held intact for other queries...<br>
</li>
-
+
</ol>
<li>"Author" Object-Type is presented to MOBY Central as a locateServiceByInput
query</li>
<li>MOBY Central responds that PubMed, TAIR, OpenBQS, InfoTree, and DragonDB
accept Author Objects as input, where the output object types are:</li>
-
+
<ol>
<li>PubMed generates Object and/or Citation Object-Types in the
namespace PubMed/ID</li>
-
+
<ol>
<li>reports itself as authoritative for this service in this namespace.</li>
<li>provides Xrefs to other namespaces if it knows about them<br>
</li>
-
+
</ol>
<li>TAIR generates Object and/or Citation Object-Types in the namespace
TAIR/CitationID</li>
-
+
<ol>
<li>reports to be authoritative for this service in this namespace</li>
<li>PubMed/ID and/or other cross-references appear in the CRIB where
available<br>
</li>
-
+
</ol>
<li>DragonDB (i.e. self) generates Object and Citation Object-Types in
the namespace DragonDB/CitationID</li>
-
+
<ol>
<li>Reports to be authoritative for this service in this namespace.</li>
<li>PubMed/ID and/or other cross-references appear in the CRIB where
available<br>
</li>
-
+
</ol>
<li>OpenBQS generates Object and/or Citation Objects-Type in the namespace
OpenBQS/ID</li>
-
+
<ol>
<li>reports to be authoritative for this service in this namespace.</li>
<li value="2">PubMed/ID and/or other cross-references appear in the
CRIB where available<br>
</li>
-
+
</ol>
<li>InfoTree generates Object and/or Citation objects in the namespace
ScienceCitationIndex/ID</li>
-
+
<ol>
<li>reports to be authoritative for this service in this namespace.</li>
<li>PubMed/ID and/or other cross-references appear in the CRIB where
available<br>
</li>
-
+
</ol>
-
+
</ol>
<li>Script choses only 'authoritative' services (in this case, all of them
are authoritative)</li>
<li>Script goes through the following loop:</li>
-
+
<ol>
<li>passes list of all Author objects to each service in turn (5.1, 5.2,
5.3, 5.4, 5.5); requests Citation object-type as output</li>
@@ -177,12 +180,12 @@
<li>foreach namespace/id combination, including those in the CRIB Xrefs,
compare it with the namespace/id combinations of the Citations & CRIB
Xrefs it retrieved in step 1.1</li>
-
+
<ol>
<li>remove the ones that it already knows about.</li>
-
+
</ol>
-
+
</ol>
<li>The remaining Citation objects are new to the database, but some may
still not be relevant.</li>
@@ -190,10 +193,10 @@
(Author objects)</li>
<li>compare all remaining Citations for those that have two or more Authors
that match the Author objects obtained from DragonDB</li>
-
+
<ol>
<li>send these Citations to DragonDB Service 'E'</li>
-
+
</ol>
<li>All remaining should be stored in a file and the curator alerted by
email.<br>
@@ -254,15 +257,15 @@
<li>Namespace of the output objects is not registered in MOBY Central<br>
</li>
</ol>
-It was interesting to me that there is so often a necessity to pass back objects
-in your own namespace. This is because, for example, not all literature
-references are in PubMed, so I can not *consistently* guarantee that any
-object I pass back to you is going to have a PubMed/ID. Therefore it
-is more appropriate for me to pass you back a Citation object using my own
-namespace (since that namespace consistently represents everything that I
-know for sure), and give you a foreign namespace as an Xref if I happen
-to know it. This actually doesn't make a single bit of difference to
-MOBY, however, since services which operate on "data" will not care about
+It was interesting to me that there is so often a necessity to pass back
+objects in your own namespace. This is because, for example, not all
+literature references are in PubMed, so I can not *consistently* guarantee
+that any object I pass back to you is going to have a PubMed/ID. Therefore
+it is more appropriate for me to pass you back a Citation object using my
+own namespace (since that namespace consistently represents everything that
+I know for sure), and give you a foreign namespace as an Xref if I
+happen to know it. This actually doesn't make a single bit of difference
+to MOBY, however, since services which operate on "data" will not care about
the namespace, and you should be passing Xrefs for any data that you give
out in your own namespace (if they exist). Thus, if everyone is playing
the game properly, we do not limit interoperability by letting this happen.<br>
More information about the MOBY-guts
mailing list