Currently diva-windup rule (DivaLauncher.perform) is very slow. We guess it's due to our construction of janus-graph. The arquillian test log (*) repeats the following message while running the diva rule.

Sep 14, 2021 9:33:05 AM org.janusgraph.graphdb.transaction.StandardJanusGraphTx$3 execute
WARNING: Query requires iterating over all vertices [()]. For better performance, use indexes
Sep 14, 2021 9:33:06 AM org.janusgraph.graphdb.transaction.StandardJanusGraphTx$3 execute
WARNING: Query requires iterating over all vertices [()]. For better performance, use indexes
Sep 14, 2021 9:33:06 AM org.janusgraph.graphdb.transaction.StandardJanusGraphTx$3 execute
WARNING: Query requires iterating over all vertices [()]. For better performance, use indexes
...

Diva standalone runs faster in few seconds.

Here are steps to run the arquillian test.

$ git clone https://github.com/WASdev/sample.daytrader7.git
$ cd tackle-diva
$ ./gradlew publishToMavenLocal -x test
$ cd windup
$ JAVA_HOME=<path-to-java8-runtime> mvn verify -Dtest=DaytraderTest

*) https://github.com/konveyor/tackle-diva/runs/3597051370

Implement preliminary function to support PL/SQL, which is a format used in Oracle DBMS.

• As a first step, DOA structurally parses PL/SQL using our parser and provides assessment results for automatic/manual conversion.
• In the next step, DOA performs minimum conversion of source SQLs so that psql accepts converted files without errors.
  • #72
  • #75

• Remove usage of Makefile
• Introduce template-based document/code generation (using jinja2)
• Change name and structure of the generated files
• Update documents to match the implementation
• Brush up documents

• added the Jitpack registry
• declared the diva-core dependency following the Jitpack's naming convention

Consider Jitpack works also if the version tag references a commit or a tag so we can use it in a very productive way: right now it's using the snapshot version based on the latest main branch.

• Implement DOA frontend in Python and package it in python-doa-cli directory.
• Enhance DOA for Oracle including semantic verification and package it in python-doa-lib directory.
• Deploy the built DOA Docker image on quay.io.

Current DOA frontend (that invokes DOA Docker images) are implemented as bash scripts. Using bash to implement complex CLI, including handling of CLI options, whitespace in arguments and exceptions, is error-prone.

Thus we plan to rewrite them as a Python script, that adds Python to dependencies to use DOA.

commit 1383ebb6fd77643a802897199cb5be60739d5201 Author: Shin Saito [email protected] Date: Wed May 11 20:10:39 2022 +0900

fix #80. version of DOA bumped to 2.2.1

commit bb34f3c304c524841878b6979a73fd526c7e89c0 Author: Shin Saito [email protected] Date: Wed Apr 20 04:48:22 2022 +0000

update document

commit 0655f9a1e0247ac21021be6153d5b91370d5a56d Author: Shin Saito [email protected] Date: Wed Apr 20 04:43:06 2022 +0000

update document

commit 3aa1599e0214a02ec1d6940d8196d9b3355bec8d Author: Shin Saito [email protected] Date: Wed Apr 20 04:09:54 2022 +0000

update document

commit ac0a661e0cf3a19ecacde22bc219b3f3c9ffb22b Author: Shin Saito [email protected] Date: Wed Apr 20 04:06:25 2022 +0000

update document

commit 7f6cb4bbfcd961da08b1501c6300ca17d7e7cb3a Author: Shin Saito [email protected] Date: Tue Apr 19 16:09:00 2022 +0000

update document

commit 2078edff2745783ce8a26b0cba0462ee96e6d934 Author: Shin Saito [email protected] Date: Tue Apr 19 16:00:09 2022 +0000

add rew document on Oracle dialects

DOA (run-doa.sh) fails when -f is specified and the input directory contains whitespaces:

Example:

bash ./run-doa.sh -l oracle -f "/tmp/a b/plsql-example"

It makes docker run, especially mounting of the input directory, fail.

See for -f and -l usage: https://github.com/konveyor/tackle-diva/blob/main/doa/README-Oracle.md

Users of Tackle DiVA should not have to clone a repo and build their own Docker image. We should host an "official" image on quay.io with all of the other Konveyor tools.

I propose that we call the image tackle-diva and change the documentation and scripts to point to: quay.io/konveyor/tackle-diva for all commands related to running DiVA stand-alone. This should simplify its use.

Along with this the README.md and ./distrib/bin/diva_docker script should be updated to use the new image.

While testing the Windup Diva addon, it turned out the console output has a lot of System.out.println [link to one sample below] entries that would be great if moved into a logging framework at (my suggestion) DEBUG level.

https://github.com/konveyor/tackle-diva/blob/ea087c515c50b1832abc73dc8b14dfeddf4895d5/core/src/main/java/io/tackle/diva/analysis/JPAAnalysis.java#L110

Hi , @YasuKatsuno @akihikot I'm onboarding to the Tackle project along with @jortel and @mansam right now.

(Note: I'm not sure if tackle-diva is meant to be integrated with tackle-inventory or if it's more tied to windup. Looking for information on this)

We are working to learn how the Tackle App Inventory can interface with other tools in the Tackle suite, and trying to define the requirements of the App Inventory data model.

It would be helpful to have a reference doc on the public interface of tackle-diva (TD) so that we can get an idea of how TD this would tie in with Tackle Inventory, and what changes might be needed to the Tackle Inventory data model:

• Expected inputs / outputs of TD (examples would be useful)
• Methods for invoking TD (REST? CLI? Java API? Something else?)
• Is there a Quay.io container auto-build set up so we can consume TD easily running as a Pod on OpenShift?
• Is there a deploy YAML available with a deployment definition for TD and any required auxiliary OpenShift resources (e.g. namespaces, services, routes). This would let us quickly play around with TD using other info provided. We may be able to assist here if needed.

Do documents detailing the public interface of TD exist? Even some simple examples would be useful to get started.

cc @rromannissen @jwmatthews @PhilipCattanach