Call Summary: We talked about the mapping of our ODF RelaxNG grammar to some graph data base to analyse it. On the 20th of September will be our next meeting :
https://www.timeanddate.com/worldclock/meetingdetails.html?year=2017&month=09&day=20&hour=14&min=30&sec=0&p1=179&p2=37&p3=136&p4=234&iv=1800 The teleconference login data for next call will be found in the OASIS calendar event:
https://www.oasis-open.org/committees/event.php?event_id=39530 [16:37] Svante Schubert:
https://lists.oasis-open.org/archives/office-collab/201709/msg00000.html [16:38] Svante Schubert: Neo4J should work, all you need is a GraphXML import :) [16:38] Svante Schubert: I have tested it with OrientDB [16:38] Svante Schubert: But was not entirely happy with the visual front-end [16:38] Svante Schubert: It collapses if you have more than 30 nodes.. [16:38] Svante Schubert: likes 60 thousand nodes of the ODF grammar :) [16:41] Svante Schubert: There are some points we might need to improve to ease traversal, like this construct: [16:42] Svante Schubert: <element> <choice> <name>text:page-count</name> <name>text:paragraph-count</name> <name>text:word-count</name> <name>text:character-count</name> <name>text:table-count</name> <name>text:image-count</name> <name>text:object-count</name> </choice> <ref name="common-field-num-format-attlist"/> <text/> </element> [16:42] Svante Schubert: One element might have different names [16:42] Svante Schubert: It seems Neo4J is loosing interoperability [16:43] Svante Schubert: e.g. the gremlin (tinkerpop) installation is not by default in the installation [16:47] Svante Schubert: In the MultiSchemaModel memory dump, where the graph is created from it looks like this: [16:47] Svante Schubert: 36: ELEMENT "text:page-count", "text:paragraph-count", "text:word-count", "text:character-count", "text:table-count", "text:image-count", "text:object-count", 37: SEQUENCE 38: REF 'common-field-num-format-attlist', [16:48] Svante Schubert: the question is if we might split this reference up into many close to identical subgraphs to make traversal scripts easier to write :) [16:54] Svante Schubert: For instance, to find the minimal ODF text document, we would traverse the graph and drop every optional subtrees. [16:55] Svante Schubert: A second step might be to create a document from the given sub-grammar for testing purpose. [16:59] Svante Schubert: BTW similar element choice construct can be used with elements: [16:59] Svante Schubert: <element name="text:bibliography-mark"> <attribute name="text:bibliography-type"> <ref name="text-bibliography-types"/> </attribute> <zeroOrMore> <attribute> <choice> <name>text:identifier</name> <name>text:address</name> <name>text:annote</name> <name>text:author</name> [16:59] Svante Schubert: We should just be aware of this, when traversing the tree/graph [17:00] Patrick: <element> <choice> <name>text:page-count</name> <name>text:paragraph-count</name> <name>text:word-count</name> <name>text:character-count</name> <name>text:table-count</name> <name>text:image-count</name> <name>text:object-count</name> </choice> <ref name="common-field-num-format-attlist"/> <text/> </element> [17:01] Patrick: means to allow all these elements, <ref name="common-field-num-format-attlist"/> [17:01] Svante Schubert: yes and might have the same parent [17:01] Svante Schubert: Like an identical subtree with a different element name as subtree parent [17:04] Patrick: this choice element defines the listed <name> with the common-field-num-format-attlist attribute, that is each of them has that common set of attributes [17:07] Patrick: <element name="text:page-count"> <ref name="common-field-num-format-attlist"/> </element> [17:07] Patrick: and so on for all the choice elements [17:11] Svante Schubert: Back to the GraphXML. When you look at it:
https://issues.apache.org/jira/secure/attachment/12885443/odf12-graph.xml You see, that in the first step it is a very simple graph with 'hasChild' edges. ?