only sentence as boundary. no valid section. automatic deploy.

jeekim · Jul 11, 2017 · f6305bf · f6305bf
1 parent f83272b
commit f6305bf
Show file tree

Hide file tree

Showing 4 changed files with 31 additions and 11 deletions.
diff --git a/build.sbt b/build.sbt
@@ -1,5 +1,5 @@
-/* import AssemblyKeys._
-import sbt.complete.DefaultParsers._*/
+import AssemblyKeys._
+// import sbt.complete.DefaultParsers._
 
 assemblySettings
 
@@ -11,6 +11,9 @@ libraryDependencies += "org.scalacheck" %% "scalacheck" % "1.12.5" % "test"
 scalacOptions in (Compile,doc) := Seq("-groups", "-implicits")
 scalacOptions in Test ++= Seq("-Yrangepos")
 
+name := "AnnotationFilter"
+version := "0.2"
+
 lazy val testERC = taskKey[Unit]("Prints 'ERC test results'")
 testERC := {
   "cat test/ercfunds.txt" #|
@@ -89,3 +92,14 @@ generateMP := {
 		"bin/mp.rb 2" #>
 		file("automata/mp.mwt") !
 }
+
+val deploy = TaskKey[Unit]("deploy", "Copies assembly jar to remote location")
+
+deploy <<= assembly map { (asm) =>
+  val account = sys.env.get("ACCOUNT").getOrElse("")
+	val dpath = sys.env.get("DPATH").getOrElse("")
+	val local = asm.getPath
+	val remote = account + ":" + dpath + asm.getName
+	println(s"Copying: $local -> $account:$remote")
+	Seq("scp", local, remote) !!
+}
diff --git a/src/main/java/ukpmc/AnnotationFilter.java b/src/main/java/ukpmc/AnnotationFilter.java
@@ -109,11 +109,11 @@ public void invoke(StringBuilder yytext, int start, DfaRun runner) {
           String content = map.get(Xml.CONTENT);
           String newoutput;
 
-          if ("TABLE".equals(map.get("type"))) {
+          /* if ("TABLE".equals(map.get("type"))) {
               DfaRun dfaRunEntity = new DfaRun(dfa_entity);
               dfaRunEntity.clientData = map.get("type");
               newoutput = dfaRunEntity.filter(content);
-          } else if ("SENT".equals(map.get(Xml.TAGNAME))) {
+          } else */ if ("SENT".equals(map.get(Xml.TAGNAME))) {
               DfaRun dfaRunPlain = new DfaRun(dfa_plain);
               dfaRunPlain.clientData = map.get(Xml.TAGNAME);
               newoutput = dfaRunPlain.filter(content);
@@ -279,7 +279,7 @@ private static boolean isOnlineValid(String db, String id, String domain) {
          loadPredefinedResults();
 
          Nfa bnfa = new Nfa(Nfa.NOTHING);
-         bnfa.or(Xml.GoofedElement("SecTag"), procBoundary)
+         bnfa // .or(Xml.GoofedElement("SecTag"), procBoundary)
          .or(Xml.GoofedElement("SENT"), procBoundary);
          dfa_boundary = bnfa.compile(DfaRun.UNMATCHED_COPY);
 

diff --git a/src/main/resources/validate.properties b/src/main/resources/validate.properties
@@ -3,6 +3,7 @@ doiblacklist=doi.prefix.1000.tsv
 # boundary=article
 # boundary=text
 # boundary=SENT|text|article
-boundary=table|SENT
+# boundary=SENT
+# boundary=table|SENT
 entity=z:[a-z]+
 # entity=z:acc
diff --git a/src/test/scala/Annotation_test.scala b/src/test/scala/Annotation_test.scala
@@ -19,9 +19,9 @@ class AnnotationSpec extends FlatSpec with Matchers {
       """<article><text><SENT sid="34" pm="."><plain>omim <z:acc db="omim" valmethod="onlineWithContext" domain="omim" context="(?i)(o*mim)" wsize="20" sec="">603878</z:acc></plain></SENT></text></article>""") should be (
       """<article><text><SENT sid="34" pm="."><plain>omim <z:acc db="omim" ids="603878">603878</z:acc></plain></SENT></text></article>""")
     //
-    dfaRun.filter(
-      """<SecTag type="TABLE"><text><SENT sid="34" pm="."><plain>omim <z:acc db="omim" valmethod="onlineWithContext" domain="omim" context="(?i)(o*mim)" wsize="20" sec="">603878</z:acc></plain></SENT><SENT sid="35" pm="."><plain><z:acc db="omim" valmethod="onlineWithContext" domain="omim" context="(?i)(o*mim)" wsize="2000" sec="XXXX">603878</z:acc></plain></SENT></text></SecTag>""") should be (
-      """<SecTag type="TABLE"><text><SENT sid="34" pm="."><plain>omim <z:acc db="omim" ids="603878">603878</z:acc></plain></SENT><SENT sid="35" pm="."><plain><z:acc db="omim" ids="603878">603878</z:acc></plain></SENT></text></SecTag>""")
+    //dfaRun.filter(
+    //  """<SecTag type="TABLE"><text><SENT sid="34" pm="."><plain>omim <z:acc db="omim" valmethod="onlineWithContext" domain="omim" context="(?i)(o*mim)" wsize="20" sec="">603878</z:acc></plain></SENT><SENT sid="35" pm="."><plain><z:acc db="omim" valmethod="onlineWithContext" domain="omim" context="(?i)(o*mim)" wsize="2000" sec="XXXX">603878</z:acc></plain></SENT></text></SecTag>""") should be (
+    //  """<SecTag type="TABLE"><text><SENT sid="34" pm="."><plain>omim <z:acc db="omim" ids="603878">603878</z:acc></plain></SENT><SENT sid="35" pm="."><plain><z:acc db="omim" ids="603878">603878</z:acc></plain></SENT></text></SecTag>""")
     //
     dfaRun.filter(
       """<SecTag type="FIG"><text><SENT sid="34" pm="."><plain>omim <z:acc db="omim" valmethod="onlineWithContext" domain="omim" context="(?i)(o*mim)" wsize="20" sec="">603878</z:acc></plain></SENT><SENT sid="35" pm="."><plain><z:acc db="omim" valmethod="onlineWithContext" domain="omim" context="(?i)(o*mim)" wsize="2000" sec="XXXX">603878</z:acc></plain></SENT></text></SecTag>""") should be (
@@ -31,9 +31,14 @@ class AnnotationSpec extends FlatSpec with Matchers {
       """<SENT sid="0" pm="."><plain>Research was supported by Finnish Academy (141069), ERC (ERC-2009-AdG-<z:acc db="erc" valmethod="context" domain="" context="(?i)(European Research Council|ERC grant|ERC advanced grant|ERC starting grant|ERC consolidator grant|ERC proof-of-concept grant|ERC-20[0-1][0-9]-)" wsize="60" sec="XXXX">250050</z:acc>, FutureGenes) grant, Sigrid Juselius Foundation, Finnish Foundation for Cardiovascular Research (all to S.Y.); the European Research Council (ERC-2010-AdG-<z:acc db="erc" valmethod="context" domain="" context="(?i)(European Research Council|ERC grant|ERC advanced grant|ERC starting grant|ERC consolidator grant|ERC proof-of-concept grant|ERC-20[0-1][0-9]-)" wsize="60" sec="XXXX">268804</z:acc>, VESSEL network), Leducq Transatlantic Network of Excellence on Lymph Vessels in Obesity and Cardiovascular Disease (11CVD03) (all to K.A.), a VENI fellowship of the Netherlands Organization of Scientific research (to J.C.S. 016.116.017); a PhD-student fellowship from the Cardiovascular Research Institute Maastricht (to T.L.T.)  </plain></SENT>""") should be (
       """<SENT sid="0" pm="."><plain>Research was supported by Finnish Academy (141069), ERC (ERC-2009-AdG-<z:acc db="erc" ids="250050">250050</z:acc>, FutureGenes) grant, Sigrid Juselius Foundation, Finnish Foundation for Cardiovascular Research (all to S.Y.); the European Research Council (ERC-2010-AdG-<z:acc db="erc" ids="268804">268804</z:acc>, VESSEL network), Leducq Transatlantic Network of Excellence on Lymph Vessels in Obesity and Cardiovascular Disease (11CVD03) (all to K.A.), a VENI fellowship of the Netherlands Organization of Scientific research (to J.C.S. 016.116.017); a PhD-student fellowship from the Cardiovascular Research Institute Maastricht (to T.L.T.)  </plain></SENT>""")
     //
+    //dfaRun.filter(
+    //  """<SecTag type="ACK"><text><SENT sid="34" pm="."><plain>omim <z:acc db="omim" valmethod="onlineWithContext" domain="omim" context="(?i)(o*mim)" wsize="20" sec="ACK">603878</z:acc></plain></SENT><SENT sid="35" pm="."><plain><z:acc db="omim" valmethod="onlineWithContext" domain="omim" context="(?i)(o*mim)" wsize="2000" sec="ACK">603878</z:acc></plain></SENT></text></SecTag>""") should be (
+    //  """<SecTag type="ACK"><text><SENT sid="34" pm="."><plain>omim 603878</plain></SENT><SENT sid="35" pm="."><plain>603878</plain></SENT></text></SecTag>""")
     dfaRun.filter(
-      """<SecTag type="ACK"><text><SENT sid="34" pm="."><plain>omim <z:acc db="omim" valmethod="onlineWithContext" domain="omim" context="(?i)(o*mim)" wsize="20" sec="ACK">603878</z:acc></plain></SENT><SENT sid="35" pm="."><plain><z:acc db="omim" valmethod="onlineWithContext" domain="omim" context="(?i)(o*mim)" wsize="2000" sec="ACK">603878</z:acc></plain></SENT></text></SecTag>""") should be (
-      """<SecTag type="ACK"><text><SENT sid="34" pm="."><plain>omim 603878</plain></SENT><SENT sid="35" pm="."><plain>603878</plain></SENT></text></SecTag>""")
+      """<SENT sid="1201668" pm="."><plain>Anti-M13K07 monoclonal antibody was purchased from Amersham Biosciences (Piscataway, NJ).Table 1.Antibody specificity.AntibodyProtein specificityReovirus strainT1LT3DT3SA+5C6T1 σ1+−−<z:acc db="pdb" valmethod="onlineWithContext" domain="pdbe" context="(?i)(pdb|(?:protein +data *bank))" wsize="20000" sec="">9BG5</z:acc>T3 σ1−++8H6T1 and T3 μ1+++4F2T3D σ3−+− </plain></SENT>""") should be (
+      """<SENT sid="1201668" pm="."><plain>Anti-M13K07 monoclonal antibody was purchased from Amersham Biosciences (Piscataway, NJ).Table 1.Antibody specificity.AntibodyProtein specificityReovirus strainT1LT3DT3SA+5C6T1 σ1+−−9BG5T3 σ1−++8H6T1 and T3 μ1+++4F2T3D σ3−+− </plain></SENT>"""
+
+    )
   }
   // <z:acc db="omim" ids="603878-603890">603878 to 603890</z:acc>
 }