[Bug]: Build issues

[Johnsd11]

What happened?

Just a heads up, getting several build issues due to the fact that the
5.0.0-SNAPSHOT jars for most of the ctakes-??-models.jar seem to have
disappeared from the snapshot repository.

Manually adjusting those to 5.0.1-SNAPSHOT is working for now, but
thought I would let whoever controls the build process know that a fresh
checkout from github is currently broken.

Relevant log output

Example:

[ERROR] Failed to execute goal on project ctakes-ytex-uima: Could not 
resolve dependencies for project 
org.apache.ctakes:ctakes-ytex-uima:jar:5.0.0-SNAPSHOT: The following 
artifacts could not be resolved: 
org.apache.ctakes:ctakes-dependency-parser-models:jar:5.0.0-SNAPSHOT 
(absent): 
org.apache.ctakes:ctakes-dependency-parser-models:jar:5.0.0-SNAPSHOT was 
not found in https://repository.apache.org/content/groups/snapshots/ 
during a previous attempt. This failure was cached in the local 
repository and resolution is not reattempted until the update interval 
of apache.snapshots has elapsed or updates are forced -> [Help 1]


Verified that core models have disappeared:

https://repository.apache.org/content/groups/snapshots/org/apache/ctakes/ctakes-core-models/5.0.0-SNAPSHOT/

contains no jars

cTAKES.error.log contents

Version

5.1.0

What operating system are you seeing the problem on?

No response

Contact Details

No response

[Johnsd11]

I tried just updating the parent pom.xml to 5.0.1-SNAPSHOT and that
seemed to work for now.

If you did not see, I uploaded my PR requests as well and verified each
one builds independently with the current main branch.

#10

#11

#12

I am computing some large similarity matrices with the updated code
today, and so far so good!

[Johnsd11]

Thanks for the PRs, I did get notifications but haven't yet had time to take a look.

The versioning issue with the models is because I am prepping a 5.0.0 release and had to make some -versioning- with the model artifacts first. It is a long story, but 5.0.0 should be put up for a vote soon. Right now there are technical release issues holding it up.

Because of the 5.0.0 release in-process, I think that we will need to wait until 5.0.1 (or whatever is next) to get your updates and improvements into a release.

Many thanks, and I look forward to seeing what you have done,

[Johnsd11]

Is cTAKES 5.0 ready to be used? or is it still under development?where can we download it?

[Johnsd11]

Is cTAKES 5.0 ready to be used? or is it still under development?where can we download it?

5.0.0 has not yet been released.

It is/will be exactly what is in the github repository currently labeled 5.0.0-SNAPSHOT. https://github.com/apache/ctakes

There is a 5.0.0 tag here https://github.com/apache/ctakes/releases/tag/ctakes-5.0.0 - though the versioning within still indicates -SNAPSHOT.

[Johnsd11]

I updated the poms in 5.0.0-SNAPSHOT to use the correct models. There shouldn't be any problem building.

[Johnsd11]

it is working again with the 5.0.0 models. I'm actually
glad you did not roll in my updates yet, as I found another subtle bug
in the creation of the concept graphs and I have implemented 4 more of
the kernel metrics which we are investigating based on the paper from
Sanchez and Batet:
https://www.sciencedirect.com/science/article/pii/S1532046411000645

I am adding in the Dice, Ochiai, Simpson and Braun-Blanquet methods
shown in Table 3. I discovered some of the path metrics were being
inflated due to invalid max-depth calculations in the original concept
graph creation due to the root concept not being set to a depth of zero
which I have fixed and now all appears to be working. I will cancel
those PR's and update later today.

[Johnsd11]

I've closed 2 of my PRs and pushed 2 new ones to replace those with the
latest code fixes and updates.

#13

#14

Thanks for your time and consideration of my contributions :-) No
changes were required of the MySQL update code.

[Johnsd11]

I think we may end up with a series of pubs out of this work. I just
want to emphasize how much more efficient it is to compute these metrics
using cTakes than the old Perl UMLS::Similarity module. I don't think
anyone realizes the amount of computational time this saves us (days to
compute matrices with Perl versus seconds/minutes with cTakes).
Hopefully we can shed some light on the benefits (even if it is a bit
more complicated to setup and use) :-)

[Johnsd11]

That sounds great - I am happy when ctakes beneficial in any way.

As far as simplicity goes, there is an installation tool in 5.0.0-SNAPSHOT: https://repository.apache.org/content/repositories/snapshots/org/apache/ctakes/ctakes-dockhand/5.0.0-SNAPSHOT/
I have to run some final tests, but I have been thinking of placing it in GitHub as an artifact. The single jar is all that is needed and it is a self-contained executable - as long as java is in the $PATH. It builds an installation with a piper file that can be edited by the user.

We are still trying to make standard installation easier, so if you have any suggestions please post them on the ctakes GitHub repo here: https://github.com/apache/ctakes/issues
with the label "enhancement".

Thank you for all the enhancements and technical information,