Attempting to Reproduce presto-nrp.yaml

[jahveen]

Hello again,

My colleagues and I have followed the yang2swagger tool instructions to try to replicate the presto-nrp.yaml swagger definition that is published in this repo. However, we cannot generate an identical copy (nor indeed similar) to the file MEF-LSO-Presto-SDK/api/swagger/presto-nrp.yaml. I wonder - is there a specific environment that is required to successfully replicate the result?

In other words, I am saying that the instructions provided do not give enough information to allow us to confidently reproduce the published presto-nrp.yaml swagger definition - so we request some assistance.

Thanks,
Jahveen

[bartoszm]

@jahveen given I am responsible for quite extensive part of yang2swagger I will try to help.
What do you mean by not being able to generate identical copy? Are you able to generate a copy or you are not able to generate and swagger due to problems during generation process?

[tcmoriarty]

@bartoszm @jahveen Hi , I am a colleague of @jahveen ,

I see many warnings and errors when I run -
java -cp yang2swagger-tapi-cli-1.4-cli.jar com.amartus.y2s.Generator -yang-dir api/yang/ -use-namespaces -elements DATA -output ./presto_nrp.yaml

..
2020-09-10 15:43:35,871 [main] INFO c.mrv.yangtools.common.ContextHelper - adding api/yang
2020-09-10 15:43:36,007 [main] INFO c.mrv.yangtools.common.SchemaBuilder - Inspecting all defined yangs [api/yang/mef-common-types.yang, api/yang/mef-common.yang, api/yang/mef-common@2018-09-26.yang, api/yang/mef-types@2018-09-26.yang, api/yang/nrm-connectivity.yang, api/yang/nrm-connectivity@2018-09-26.yang, api/yang/nrm-l-1@2018-09-26.yang, api/yang/nrm-oam@2018-09-26.yang, api/yang/nrp-interface.yang, api/yang/tapi-common@2018-03-07.yang, api/yang/tapi-common@2018-10-16.yang, api/yang/tapi-connectivity@2018-03-07.yang, api/yang/tapi-connectivity@2018-10-16.yang, api/yang/tapi-dsr@2018-10-16.yang, api/yang/tapi-eth@2018-10-16.yang, api/yang/tapi-notification@2018-10-16.yang, api/yang/tapi-oam@2018-10-16.yang, api/yang/tapi-odu@2018-10-16.yang, api/yang/tapi-path-computation@2018-03-07.yang, api/yang/tapi-path-computation@2018-10-16.yang, api/yang/tapi-photonic-media@2018-10-16.yang, api/yang/tapi-topology@2018-03-07.yang, api/yang/tapi-topology@2018-10-16.yang, api/yang/tapi-virtual-network@2018-10-16.yang]
2020-09-10 15:43:37,116 [main] WARN o.o.y.y.p.s.r.AugmentStatementImpl - Failed to add augmentation null:277:12 defined at null:104:4
org.opendaylight.yangtools.yang.parser.spi.meta.InferenceException: An augment cannot add node 'cos-identifier-list' because it is mandatory and in module different than target [at null:146:16]
at org.opendaylight.yangtools.yang.parser.spi.meta.InferenceException.throwIf(InferenceException.java:47)
at org.opendaylight.yangtools.yang.parser.stmt.rfc6020.AugmentStatementImpl$Definition.checkForMandatoryNodes(AugmentStatementImpl.java:301)
at java.util.ArrayList.forEach(ArrayList.java:1259)
....

Many Warnings follow ..

The first ERROR , followed by many more :

2020-09-10 15:43:37,527 [main] ERROR o.o.y.y.p.s.r.BuildGlobalContext - Failed to parse YANG from source SourceSpecificContext [source=null, current=EFFECTIVE_MODEL, finished=FULL_DECLARATION]: Grouping '(urn:onf:otcc:yang:tapi-connectivity?revision=2018-03-07)connection-end-point-ref' was not resolved. [at null:83:16]
2020-09-10 15:43:37,527 [main] ERROR o.o.y.y.p.s.r.BuildGlobalContext - 13 additional errors reported:

Any help would be appreciated , we plan to add to the yang models later , but we are having difficulty getting up and running with default files . I am hoping its something that we can easily address. Thanks -

regards,
Tim

[jahveen]

Hello @bartoszm,

We have performed more extensive testing and here are the details:

Testing Environment

OS: Ubuntu 20.04.1 LTS (GNU/Linux 5.4.0-48-generic x86_64)
java: openjdk version "1.8.0_265"
maven: 3.6.3
yang2swagger: tag/1.1.7
y2s-tapi-cli: tag/v1.0.0
MEF-LSO-Presto-SDK: tags/r3

Command Used

java -classpath /home/jahveen/y2s-tapi-cli/target/yang2swagger-tapi-cli-1.0-cli.jar com.amartus.y2s.Generator mef-common tapi-topology tapi-common tapi-connectivity tapi-path-computation mef-common-types nrm-connectivity nrp-interface -yang-dir /home/jahveen/yangs/ -api-version 2018R2 -output /home/jahveen/presto-nrp.yaml

Notes:
(1) I created the yang-dir /home/jahveen/yangs/ and copied the following YANG modules from /home/jahveen/MEF-LSO-Presto-SDK/api/yang:

nrp-interface.yang, tapi-connectivity@2018-03-07.yang, tapi-common@2018-03-07.yang, nrm-connectivity.yang, tapi-topology@2018-03-07.yang, tapi-path-computation@2018-03-07.yang, mef-common-types.yang, mef-common.yang

The combination of modules that I settled on was because of trial and error selection - there was no clear definition of which revision of each module I should use

(2) I specified all of the YANG modules in the command. The result is no different to me not specifying them at all, since only the specified modules are in the yang-dir

(3) I specified api-version to match the one specified in presto-nrp.yaml

Output

99% the same as the repo version in /home/jahveen/MEF-LSO-Presto-SDK/api/swagger/presto-nrp.yaml. The differences I found are:

(1) description: The repo version is "MEF 60 (PRESTO NRP) swagger definition". My generated version is a list of the modules

(2) title: Both versions are a list of the modules, but in a different order, and the repo version is missing the module name "tapi-path-computation"

(3) basepath: This is included in the repo version only

(4) consumes / produces: The repo version has two ("application/json", "application/xml"), the generated version only allows one to be specified

Questions

(1) Please comment on the differences I found. Are they expected? If so, am I supposed to manually edit my output to match the repo

(2) More a statement than a question - it would be useful to provide clearer instructions on how to reproduce the swagger output so that we can be confident on how to generate it, and more importantly, we can be confident that we can incorporate our own YANG deviations and generate the correct swagger definitions

(3) You have since updated yang2swagger (tag/1.1.7) and y2s-tapi-cli (tag/v1.0.0), and the output I see when using later versions is noticeably different. Should I expect the next release of the SDK (presumably r4) to use the latest versions of y2s-tapi-cli with all its bug fixes, etc.?

[jahveen]

Hello,

Is someone able to assist me with my questions?

Thanks.

Jahveen

[bartoszm]

Hi @jahveen,
I was contributing to the yang2swgger tooling so I can share some information that you might find useful.
I think that there was some change in the tooling as well as in the yang definitions so it might be the case that you are not using the right combination of versions while generating.
It is also possible that you are using a different combination of yang modules than the one which was selected to generate the presto swagger from the repository.

[bartoszm]

Answers: (1) I suppose (since I was not generating the files for the latest releases) that there were some manual modifications applied to generated content (e.g. title basepath etc. ) (2) I would agree with the statement - I am CC Michal to let him take that under consideration for community content improvement (3) I suppose that the new versions will be generated with the latest tool support. The fixes I've been able to introduce were answers to some of the community requests. Unfortunately, there are still issues awaiting resolution or discussion but I am afraid that without additional contribution (tool is an open source project) it will be hard to address these given my limited availability. Cheers, Bartosz

…

On Wed, 21 Oct 2020 at 11:02, jahveen ***@***.***> wrote: Hello @bartoszm <https://github.com/bartoszm>, We have performed more extensive testing and here are the details: Testing Environment OS: Ubuntu 20.04.1 LTS (GNU/Linux 5.4.0-48-generic x86_64) java: openjdk version "1.8.0_265" maven: 3.6.3 yang2swagger: tag/1.1.7 y2s-tapi-cli: tag/v1.0.0 MEF-LSO-Presto-SDK: tags/r3 Command Used java -classpath /home/jahveen/y2s-tapi-cli/target/yang2swagger-tapi-cli-1.0-cli.jar com.amartus.y2s.Generator mef-common tapi-topology tapi-common tapi-connectivity tapi-path-computation mef-common-types nrm-connectivity nrp-interface -yang-dir /home/jahveen/yangs/ -api-version 2018R2 -output /home/jahveen/presto-nrp.yaml Notes: (1) I created the yang-dir /home/jahveen/yangs/ and copied the following YANG modules from /home/jahveen/MEF-LSO-Presto-SDK/api/yang: nrp-interface.yang, ***@***.***, ***@***.***, nrm-connectivity.yang, ***@***.***, ***@***.***, mef-common-types.yang, mef-common.yang The combination of modules that I settled on was because of trial and error selection - there was no clear definition of which revision of each module I should use (2) I specified all of the YANG modules in the command. The result is no different to me not specifying them at all, since only the specified modules are in the yang-dir (3) I specified api-version to match the one specified in presto-nrp.yaml Output 99% the same as the repo version in /home/jahveen/MEF-LSO-Presto-SDK/api/swagger/presto-nrp.yaml. The differences I found are: (1) description: The repo version is "MEF 60 (PRESTO NRP) swagger definition". My generated version is a list of the modules (2) title: Both versions are a list of the modules, but in a different order, and the repo version is missing the module name "tapi-path-computation" (3) basepath: This is included in the repo version only (4) consumes / produces: The repo version has two ("application/json", "application/xml"), the generated version only allows one to be specified Questions (1) Please comment on the differences I found. Are they expected? If so, am I supposed to manually edit my output to match the repo (2) More a statement than a question - it would be useful to provide clearer instructions on how to reproduce the swagger output so that we can be confident on how to generate it, and more importantly, we can be confident that we can incorporate our own YANG deviations and generate the correct swagger definitions (3) You have since updated yang2swagger (tag/1.1.7) and y2s-tapi-cli (tag/v1.0.0), and the output I see when using later versions is noticeably different. Should I expect the next release of the SDK (presumably r4) to use the latest versions of y2s-tapi-cli with all its bug fixes, etc.? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <https://github.com/orgs/MEF-GIT/teams/mef-lso-presto-sdk-community/discussions/7/comments/3>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AALBB3GT45EGO66KPQP5SZTSL2PR7ANCNFSM4RE5SBWA> .

-- -- Bartosz Michalik