From 2ce2b3a7d7701c38d49f721d23e269a543a6b121 Mon Sep 17 00:00:00 2001
From: Ralph Soika <ralph.soika@imixs.com>
Date: Sun, 21 Aug 2022 22:58:05 +0200
Subject: [PATCH] training config setups

---
 ...config-prod.xml => model-analyze-prod.xml} | 12 ++---
 ...g-config-prod.xml => model-train-prod.xml} |  0
 models/model-valiate-prod.xml                 | 48 +++++++++++++++++++
 3 files changed, 53 insertions(+), 7 deletions(-)
 rename models/{validate-config-prod.xml => model-analyze-prod.xml} (71%)
 rename models/{training-config-prod.xml => model-train-prod.xml} (100%)
 create mode 100755 models/model-valiate-prod.xml

diff --git a/models/validate-config-prod.xml b/models/model-analyze-prod.xml
similarity index 71%
rename from models/validate-config-prod.xml
rename to models/model-analyze-prod.xml
index b38e72c..2da00a2 100755
--- a/models/validate-config-prod.xml
+++ b/models/model-analyze-prod.xml
@@ -2,18 +2,16 @@
 <document xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
 	xmlns:xs="http://www.w3.org/2001/XMLSchema">
 	
-	<!-- Workflow Instance 
-	https://alexander-logistics.office-workflow.de/api/
-	-->
-	<item name="workflow.endpoint"><value xsi:type="xs:string">http://app:8080/api/</value></item>
+	<!-- Workflow Instance -->
+	<item name="workflow.endpoint"><value xsi:type="xs:string">https://alexander-logistics.office-workflow.de/api/</value></item>
 	<item name="workflow.userid"><value xsi:type="xs:string">admin</value></item>
-	<item name="workflow.password"><value xsi:type="xs:string">adminadmin</value></item>
+	<item name="workflow.password"><value xsi:type="xs:string">imixs4.null</value></item>
 
 	<!--   -->	
-	<item name="workflow.query"><value xsi:type="xs:string">($uniqueid:f4eaceda-f28a-4394-928e-1a05daaf2760) OR ($uniqueid:a5d1e7fe-d74a-4fa2-9b06-3221865759fa) OR ($uniqueid:c0290497-f613-4492-81b6-4ae1e1680029) OR ($uniqueid:d8e3fce4-d11c-46a4-8ba0-369b4c6e902f) OR ($uniqueid:8b766f35-828a-41f1-be04-00b4b62da4d0)</value></item>
+	<item name="workflow.query"><value xsi:type="xs:string">($workflowgroup:"Rechnungseingang" OR $workflowgroup:"Sachrechnung") AND ($taskid:5900)</value></item>
 
 	<item name="workflow.pagesize"><value xsi:type="xs:int">100</value></item>
-	<item name="workflow.pageindex"><value xsi:type="xs:int">0</value></item>
+	<item name="workflow.pageindex"><value xsi:type="xs:int">30</value></item>
 	<item name="workflow.entities">
 		<value xsi:type="xs:string">cdtr.name</value>
 		<value xsi:type="xs:string">cdtr.iban</value>
diff --git a/models/training-config-prod.xml b/models/model-train-prod.xml
similarity index 100%
rename from models/training-config-prod.xml
rename to models/model-train-prod.xml
diff --git a/models/model-valiate-prod.xml b/models/model-valiate-prod.xml
new file mode 100755
index 0000000..d5ccb47
--- /dev/null
+++ b/models/model-valiate-prod.xml
@@ -0,0 +1,48 @@
+<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
+<document xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
+	xmlns:xs="http://www.w3.org/2001/XMLSchema">
+	
+	<!-- Workflow Instance -->
+	<item name="workflow.endpoint"><value xsi:type="xs:string">https://alexander-logistics.office-workflow.de/api/</value></item>
+	<item name="workflow.userid"><value xsi:type="xs:string">admin</value></item>
+	<item name="workflow.password"><value xsi:type="xs:string">imixs4.null</value></item>
+
+	<item name="workflow.entities">
+		<value xsi:type="xs:string">cdtr.name</value>
+		<value xsi:type="xs:string">cdtr.iban</value>
+		<value xsi:type="xs:string">cdtr.bic</value>
+		<value xsi:type="xs:string">invoice.total</value>
+		<value xsi:type="xs:string">invoice.date</value>
+		<value xsi:type="xs:string">invoice.number</value>
+	</item>
+	<item name="workflow.locale">
+		<value xsi:type="xs:string">en_GB</value>
+		<value xsi:type="xs:string">en_US</value>		
+		<value xsi:type="xs:string">de_DE</value>
+	</item>
+	
+	<!-- Tika OCR Server  -->
+	<item name="tika.ocrmode"><value xsi:type="xs:string">OCR_ONLY</value></item>
+	<item name="tika.options">
+		<value xsi:type="xs:string">X-Tika-OCRLanguage=eng+deu</value>
+		<value xsi:type="xs:string">X-Tika-PDFocrStrategy=OCR_ONLY</value>		
+	</item>
+
+	<!-- ML spaCy Server -->
+	<item name="ml.training.endpoint"><value xsi:type="xs:string">http://imixs-ml-spacy:8000/</value></item>
+	<item name="ml.training.model"><value xsi:type="xs:string">invoice-de-0.2.0</value></item>
+	<item name="ml.training.filepattern"><value xsi:type="xs:string">.pdf|.PDF</value></item>
+        <!-- LOW | GOOD -->
+	<item name="ml.training.quality"><value xsi:type="xs:string">LOW</value></item>
+
+	
+	<!-- Define the training set and taining mode -->
+	<item name="workflow.query"><value xsi:type="xs:string">($workflowgroup:"Rechnungseingang" OR $workflowgroup:"Sachrechnung") AND ($taskid:5900)</value></item>	  
+	<item name="workflow.pagesize"><value xsi:type="xs:int">100</value></item>
+	<item name="workflow.pageindex"><value xsi:type="xs:int">40</value></item>
+
+
+</document>
+
+
+