Translate and adjust count example to fix #589

I cant adjust the metrics example yet. At least I do not know what the outcome of the old morph is since it uses square.

metafacture-runner/src/main/dist/examples/count/gnd/count-gnd-types.flux

@@ -4,7 +4,7 @@ fileName|
 open-file|
 as-lines|
 decode-pica|
-morph(FLUX_DIR + "gnd-type.xml")|
+fix(FLUX_DIR + "gnd-type.fix")|
 stream-to-triples|
 count-triples(countBy="object")|
 template("${s}\t${o}")|

metafacture-runner/src/main/dist/examples/count/gnd/gnd-type.fix

@@ -0,0 +1,6 @@
+if any_match("[email protected]","...*")
+  replace_all("[email protected]","^(..).*","$1") #only keep the first two letters
+  retain("[email protected]") # only keep the relevent element
+else
+  reject()
+end

metafacture-runner/src/main/dist/examples/count/subjects/references.fix

@@ -0,0 +1,8 @@
+do list(path:"041A*","var":"$i")
+  copy_field("$i.9","relevantField.$append")
+end
+
+trim("relevantField.*")
+uniq("relevantField")
+
+retain("relevantField")

metafacture-runner/src/main/dist/examples/count/subjects/references.flux

@@ -1,5 +1,5 @@
 
-default counts="myflux/counts.dat";
+default counts=FLUX_DIR + "counts.dat";
 default catalogue = FLUX_DIR + "10.pica";
 
 //count references
@@ -10,10 +10,9 @@ open-file|
 as-lines|
 catch-object-exception|
 decode-pica|
-morph(FLUX_DIR + "references.xml")|
+fix(FLUX_DIR + "references.fix")|
 stream-to-triples|
 count-triples(countBy="object")|
-
-write("subjects.dat");
+write(counts);
 
 

metafacture-runner/src/main/dist/examples/morph/count/gnd/count-gnd-types.flux

@@ -0,0 +1,11 @@
+default fileName = FLUX_DIR + "gnd-sample.pica";
+
+fileName|
+open-file|
+as-lines|
+decode-pica|
+morph(FLUX_DIR + "gnd-type.xml")|
+stream-to-triples|
+count-triples(countBy="object")|
+template("${s}\t${o}")|
+write("stdout");

metafacture-runner/src/main/dist/examples/morph/count/metrics/metrics-gnd-subjects.flux

@@ -0,0 +1,14 @@
+
+catalogue|
+open-file|
+as-lines|
+catch-object-exception|
+decode-pica|
+batch-log(batchsize="100000")|
+morph(FLUX_DIR + "subject-cooccurrence.xml")|
+stream-to-triples|
+count-triples(countBy="object")|
+calculate-metrics("X2")|
+template("${s} ${o}")|
+//write("stdout");
+write(FLUX_DIR+"x2.dat");

metafacture-runner/src/main/dist/examples/morph/count/metrics/subject-cooccurrence.xml

@@ -0,0 +1,27 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<metamorph xmlns="http://www.culturegraph.org/metamorph"
+    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" version="1">
+    <rules>
+
+        <data source="041A*.9" name="@subj">
+            <trim />
+            <unique />
+        </data>
+
+        <square delimiter="&amp;" name="">
+            <data source="@subj" name=""/>
+            <postprocess>
+                <compose prefix="2:"/>
+            </postprocess>
+        </square>
+
+        <data source="@subj" name="">
+            <compose prefix="1:"/>
+        </data>
+
+        <data source="@subj" name="">
+            <occurrence only="1" />
+            <constant value="1:" />
+        </data>
+    </rules>
+</metamorph>

metafacture-runner/src/main/dist/examples/morph/count/subjects/references.flux

@@ -0,0 +1,19 @@
+
+default counts=FLUX_DIR + "counts.dat";
+default catalogue = FLUX_DIR + "10.pica";
+
+//count references
+"counting references in " + catalogue | write("stdout");
+
+catalogue|
+open-file|
+as-lines|
+catch-object-exception|
+decode-pica|
+morph(FLUX_DIR + "references.xml")|
+stream-to-triples|
+count-triples(countBy="object")|
+
+write("subjects.dat");
+
+