Provide ability to specify GROBID installation directory

[remkop]

From Slack conversation:

Andy Jackson (...) GROBID is currently a bit of a pain and has to be in {users home folder}/workspace/grobid to actually work, which I couldn't do in the Dockerfile for reasons related to how MyBinder works.

Peter Murray-Rust
Agreed. That's because GROBID didn't have an installer, I think. If we create a distro we can probably copy GROBID into it??

Andy Jackson (...) The one thing that would make it easier is being able to point ami at a user-defined folder containing whatever version of GROBID.

Peter Murray-Rust
There's a wider issue of where AMI finds its resources. Will post.

[petermr]

GROBID is a large Java program which most people run as an executable jar or a web service. See https://grobid.readthedocs.io/en/latest/

I tried to link GROBID in at the Javas library level and it was an effort and not worth it (also bloated the distrib). GROBID is usually run as a service, but here we run in in Batch mode.

Batch mode relies on certain system variables being set. See |https://grobid.readthedocs.io/en/latest/Grobid-batch/ so ami3 needs to know these. I think that's what @deadlyvices is talking about

[remkop]

From https://github.com/petermr/openVirus/issues/25, the goal is to

allow AMIGrobidTool to

• set location of the Grobid executable
• set output filename

[petermr]

I think that allowing the user to define where GROBID is , is enough. They have to copy/install GROBID but they're probably capable of that. IOW do we need to do something more ...?

[remkop]

I see now why this was a bit cumbersome: the grovid version appears in both the installation directory and in the path to the grovid jar. I fixed it by adding a number of options to the ami grobit command.

If grobid is installed in $HOME/workspace/grobid/grobid-$GROBID_VERSION, then just specifying --grobid-version=<VERSION> is sufficient.

If grobid is installed elsewhere, users can specify --grobid-install-dir. If they use a grobid version other than the default (0.5.3), they should also specify --grobid-version.

Here is the full list of the new options:

--grobid-home=<PATH>   Optionally, specify the location of the
                         `grobid-home` directory. If not specified, the
                         value is derived from the Grobid install
                         directory.
--grobid-install-dir=<PATH>
                       Optionally, specify the location where grobid is
                         installed. If not specified, the value is
                         derived from the Grobid version as follows:
                       System.getProperty("user.home") +
                         "/workspace/grobid/grobid-" + <grobid-version>
--grobid-jar=<PATH>    Optionally, specify the location of the Grobid
                         jar. If not specified, the value is derived from
                         the Grobid version and Grobid install directory.
--grobid-java-exe=<PATH>
                       Optionally, specify the location of the java
                         executable to use to invoke Grobid. If not
                         specified, the java executable that runs ami is
                         used.
                       NOTE: Grobid supported version is Java 8. More
                         recent JVM version (like JVM 11) might lead to
                         issues.
                         Default: C:\apps\jdk-14/bin/java
--grobid-version=<X.Y.Z>
                       The grobid version to use.
                         Default: 0.5.3

[petermr]

Great ,

So impressed with how well you hAve understood all the problems, thank you. We may need to use grobid in production quite soon

On Mon, 6 Apr 2020, 04:18 Remko Popma, notifications@github.com wrote:

I see now why this was a bit cumbersome: the grovid version appears in both the installation directory and in the path to the grovid jar. I fixed it by adding a number of options to the ami grobit command.

If grobid is installed in $HOME/workspace/grobid/grobid-$GROBID_VERSION, then just specifying --grobid-version= is sufficient.

If grobid is installed elsewhere, users can specify --grobid-install-dir. If they use a grobid version other than the default (0.5.3), they should also specify --grobid-version.

Here is the full list of the new options:

--grobid-home= Optionally, specify the location of the grobid-home directory. If not specified, the value is derived from the Grobid install directory. --grobid-install-dir= Optionally, specify the location where grobid is installed. If not specified, the value is derived from the Grobid version as follows: System.getProperty("user.home") + "/workspace/grobid/grobid-" + --grobid-jar= Optionally, specify the location of the Grobid jar. If not specified, the value is derived from the Grobid version and Grobid install directory. --grobid-java-exe= Optionally, specify the location of the java executable to use to invoke Grobid. If not specified, the java executable that runs ami is used. NOTE: Grobid supported version is Java 8. More recent JVM version (like JVM 11) might lead to issues. Default: C:\apps\jdk-14/bin/java --grobid-version=<X.Y.Z> The grobid version to use. Default: 0.5.3

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/petermr/ami3/issues/18#issuecomment-609544394, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAFTCS75IWUDYGLSXOLPK3DRLFCXFANCNFSM4LXFVLKQ .

[remkop]

The changes in my previous comment have been pushed to master.

We need to build a new binary to make this available to @anjackson.

@petermr, do you feel like publishing a release for this in ami-jars? Alternatively, we can wait until I have the improved deployment mechanism worked out.

[petermr]

"publishing a release" means what? giving it a version? Or is there something that I personally have to do? I can give you access to ami-jars.

On Mon, 6 Apr 2020, 07:32 Remko Popma, notifications@github.com wrote:

The changes in my previous comment have been pushed to master.

We need to build a new binary to make this available to @anjackson https://github.com/anjackson.

@petermr https://github.com/petermr, do you feel like publishing a release for this in ami-jars? Alternatively, we can wait until I have the improved deployment mechanism https://github.com/petermr/ami3/issues/24 worked out.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/petermr/ami3/issues/18#issuecomment-609591929, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAFTCS55ZIXT5GZ7RY4G5ALRLFZOFANCNFSM4LXFVLKQ .

[remkop]

Yes, I meant giving it a version and running ami-jars.sh to upload it to petermr/ami-jars, so that @anjackson can download it and try out the new version.

If you want to delegate this to me, then I would need access to ami-jars. However, I won't be able to do this today, and unsure about tomorrow, because of other commitments. :-)

[petermr]

No problem. I am amazed how much time your have donated. Is there a final jar to upload? Do I build it?

On Mon, Apr 6, 2020 at 8:38 AM Remko Popma notifications@github.com wrote:

Yes, I meant giving it a version and running ami-jars.sh to upload it to petermr/ami-jars, so that @anjackson https://github.com/anjackson can download it and try out the new version.

If you want to delegate this to me, then I would need access to ami-jars. However, I won't be able to do this today, and unsure about tomorrow, because of other commitments. :-)

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/petermr/ami3/issues/18#issuecomment-609627739, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAFTCS7RPN3JHJLHT5HQ3Y3RLGBHJANCNFSM4LXFVLKQ .

-- Peter Murray-Rust Founder ContentMine.org and Reader Emeritus in Molecular Informatics Dept. Of Chemistry, University of Cambridge, CB2 1EW, UK

[remkop]

Is there a final jar to upload? Do I build it?

You would have to build the jar from master (make sure to update first).

[remkop]

@anjackson Can you try the new release (new JAR in https://github.com/petermr/ami-jars2, version is 20200406) and close this ticket if the new options meet your requirements?

[remkop]

Closing as done.