[MOBY-guts] biomoby commit

[Paul Gordon]

gordonp
Mon Apr  2 22:37:43 EDT 2007
Update of /home/repository/moby/moby-live/Java/src/main/ca/ucalgary/seahawk/resources
In directory dev.open-bio.org:/tmp/cvs-serv7056/src/main/ca/ucalgary/seahawk/resources

Modified Files:
	mobyBuilderRules.xml 
Log Message:
Added PMID and DOI parsing rules
moby-live/Java/src/main/ca/ucalgary/seahawk/resources mobyBuilderRules.xml,1.2,1.3
===================================================================
RCS file: /home/repository/moby/moby-live/Java/src/main/ca/ucalgary/seahawk/resources/mobyBuilderRules.xml,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -r1.2 -r1.3
--- /home/repository/moby/moby-live/Java/src/main/ca/ucalgary/seahawk/resources/mobyBuilderRules.xml	2006/12/14 19:38:34	1.2
+++ /home/repository/moby/moby-live/Java/src/main/ca/ucalgary/seahawk/resources/mobyBuilderRules.xml	2007/04/03 02:37:43	1.3
@@ -4,6 +4,39 @@
 <prefix value="tigr">http://www.bioxml.info/dtd/tigrxml.dtd</prefix>
 <prefix value="bioseq">http://www.bioxml.info/dtd/Bioseq.dtd</prefix>
 
+<!-- A base object in the NCBI Genbank's GI namespace -->
+<object>
+  <regex>(?:GI|gi)[|:](\d+)</regex>
+  <xpath>substring-before(substring-after(self::tigr:TU//tigr:ACCESSION/@DBXREF, 'gi|'), '|')</xpath>
+  <namespace>
+    <ns value="NCBI_gi">$1</ns>
+  </namespace>
+</object>
+
+<!-- Pubmed ID -->
+<object>
+  <regex>(?:PMID|pmid):?\s*(\d{5,})</regex>
+  <namespace>
+    <ns value="PMID">$1</ns>
+  </namespace>
+</object>
+
+<!-- Digital Object Identifier, with header -->
+<object>
+  <regex>(?:DOI|doi):?\s*(10\.\d+/[^%"\# \t\r\n]+)</regex>
+  <namespace>
+    <ns value="DOI">$1</ns>
+  </namespace>
+</object>
+
+<!-- Digital Object Identifier, likely guess since you don't divide numbers by letters -->
+<object>
+  <regex>(?:\s|\A|/)(10\.\d+/[A-Za-z]+[^%"\# \t\r\n]+)</regex>
+  <namespace>
+    <ns value="DOI">$1</ns>
+  </namespace>
+</object>
+
 <object>
   <regex>\s*(LOCUS\s+(\S+)(?:.*|\n)+?\n//(?:\n|$))</regex>
   <namespace>
@@ -28,15 +61,6 @@
   </namespace>
 </object>
 
-<!-- a base object in the NCBI Genbank's GI namespace -->
-<object>
-  <regex>(?:GI|gi)[|:](\d+)</regex>
-  <xpath>substring-before(substring-after(self::tigr:TU//tigr:ACCESSION/@DBXREF, 'gi|'), '|')</xpath>
-  <namespace>
-    <ns value="NCBI_gi">$1</ns>
-  </namespace>
-</object>
-
 <!-- an enzyme committee number -->
 <object>
   <!-- have the form '6.1.99.-', where '-' is a wildcard -->