No processor type exists with name [opennlp]

[mattiabaldari]

Hi Spinscale,

sorry to bother you again, but I really don't get how to install your this repo.
I am in the situation where I installed ElastcSearch and Kibana (I know they are now required, but I need them) and I want to give the following command:

curl -X PUT "localhost:9200/_ingest/pipeline/opennlp-pipeline" -H 'Content-Type: application/json' -d'
{
  "description" : "A pipeline to do named entity extraction",
  "processors" : [
    {
      "opennlp" : {
        "field": "my_field",
        "ignore_missing": true
      }
    }
  ]
}'

this way I should add data to your plugin.
Sadly when I give that command I receive the following error:

{"error":
  {"root_cause":
    [{ "type":"parse_exception",
       "reason":"No processor type exists with name [opennlp]", 
       "header":{"processor_type":"opennlp"}
    }],
  "status":400}

So if no processor exists with that name I thought maybe the installation hasn't finish correctly.
To install it I followed the following steps:

1. Download it and from within the downloaded folder I ran:
   ./gradlew clean check

2. Then I tried to give:
   bin/elasticsearch-plugin install file://./Downloads/elasticsearch-ingest-opennlp-master/build/distributions/ingest-opennlp-6.5.0.1-SNAPSHOT.zip

but I receive an error:

ERROR: plugin directory [./Downloads/elasticsearch-ingest-opennlp-master/build/cluster/integTestCluster node0/elasticsearch-6.5.0/plugins/ingest-opennlp] already exists; if you need to update the plugin, uninstall it first using command 'remove ingest-opennlp'

so I really don't know what to do. Do you have some clue of where the problem is?

Note: Just to be sure, the bin/elasticsearch-plugin folder from where I ran the command was in

elasticsearch-ingest-opennlp-master/build/cluster/integTestCluster node0/elasticsearch-6.5.0

[mattiabaldari]

Please, let me know if I could give you some more information to clearify the situation

[mattiabaldari]

@spinscale nothing?

[spinscale]

can you share the output of plugins/ingest-opennlp (just sth like find plugins -type f should be sufficient.

Also, when elasticsearch is running, can you share the output from the /_nodes/stats endpoint as well.

Lastly, please share the log output when starting elasticsearch, from the first minute.

Thanks!

[spinscale]

can you explain why you tried to install that plugin in the tests directory named elasticsearch-ingest-opennlp-master/build/cluster/integTestCluster node0/elasticsearch-6.5.0? I do not understand the intention behind this. Can you clarify?

[mattiabaldari]

Hello Spinscale,

I just followed the README. In particular I did the following:

1. Downloaded the project, enter the project folder
2. run ./gradlew clean check
3. Then I should install it giving the command:
   bin/plugin install file:///path/to/elasticsearch-ingest-opennlp/build/distribution/ingest-opennlp-X.Y.Z-SNAPSHOT.zip
   but I am not getting how to run it. Where the bin/plugin should be located?

Thank you

[spinscale]

if you only want to use the plugin, you can download it from the github releases.

/bin/elasticsearch-plugin refers to your installation of Elasticsearch (the path is depending on where your unzipped your distribution, or if you installed the debian or RPM package).

[mattiabaldari]

Ok, but if I need to install it on my elasticsearch instance, is that command right?

[mattiabaldari]

@spinscale

[elastickent]

@meru3m If possible, I suggest reinstalling with the bundle (.zip file) plugin instead of building it locally.

To do that for elasticseasrch 6.5.3:

(adjust the path as needed, these are specific to CentOS 7)
export ES_PATH_CONF=/etc/elasticsearch;/usr/share/elasticsearch/bin/elasticsearch-plugin remove ingest-opennlp

export ES_PATH_CONF=/etc/elasticsearch;/usr/share/elasticsearch/bin/elasticsearch-plugin install https://github.com/spinscale/elasticsearch-ingest-opennlp/releases/download/6.5.3.1/ingest-opennlp-6.5.3.1.zip

Make sure you have downloaded (in the right directory) and have appended the model file config to the elasticsearch.yml file.
# OpenNLP
ingest.opennlp.model.file.persons: en-ner-persons.bin
ingest.opennlp.model.file.dates: en-ner-dates.bin
ingest.opennlp.model.file.locations: en-ner-locations.bin

Restart elasticsearch while tailing the logfile to insure the plugin was loaded, it should look like this:

[2018-12-12T10:18:38,694][INFO ][o.e.p.PluginsService ] [master-2] loaded plugin [ingest-opennlp]

Or by the API

curl -XGET "https://localhost:9200/_nodes/ingest"

Hopefully then opennlp processor will then be accessible to your ingest pipeline.

Here's screenshot of using opennlp on a twitter feed displayed through Kibana's graph app.

[mattiabaldari]

Thank you @elastickent I'll try and post the result here

[mattiabaldari]

Hi @elastickent,

just to be sure, when you wrote:

"Make sure you have downloaded (in the right directory) and have appended the model file config to the elasticsearch.yml file...."

could you specify where should be the "right" directory?

[elastickent]

On centos, for me it worked by dropping them into:
/etc/elasticsearch/ingest-opennlp

e.g.

[root@sensor-01 ingest-opennlp]# pwd
/etc/elasticsearch/ingest-opennlp
[root@sensor-01 ingest-opennlp]# tree
├── en-ner-dates.bin
├── en-ner-locations.bin
└── en-ner-persons.bin