Mapper-attachments

Elastic Stack Elasticsearch
1

I am trying to install mapper-attachments for es 2.3.1. if follow https://www.elastic.co/guide/en/elasticsearch/plugins/current/mapper-attachments.html it would install mapper-attachments 2.3.1 (same as es). but here GitHub - elastic/elasticsearch-mapper-attachments: Mapper Attachments Type plugin for Elasticsearch looks the version are different. if I need to install from the later link, what does this mean?

2.1 elasticsearch repository github?

I tried to install the plugin from the zip file elasticsearch-mapper-attachments-master.zip, it said NOTE: Unable to verify checksum for downloaded plugin (unable to find .sha1 or .md5 file to verify)
ERROR: Could not find plugin descriptor 'plugin-descriptor.properties' in plugin zip

so a little confusing..

thanks!

2

Just read the guide (the first link you provided).

https://www.elastic.co/guide/en/elasticsearch/plugins/current/mapper-attachments.html#mapper-attachments-install

If you read the very first lines of the README in the GitHub repo you would see also the link to this guide.

3

thanks. I got your point.

4

Very wired here.

here is my env:

http://localhost:9200/_nodes/plugins?pretty

{
  "cluster_name" : "elasticsearch",
  "nodes" : {
    "hYCvRc1yQSK3BoClf5C5zQ" : {
      "name" : "Destroyer of Demons",
      "transport_address" : "127.0.0.1:9300",
      "host" : "127.0.0.1",
      "ip" : "127.0.0.1",
      "version" : "**2.3.1"**,
      "build" : "bd98092",
      "http_address" : "127.0.0.1:9200",
      "plugins" : [ {
        "name" : "**mapper-attachments**",
        "version" : **"2.3.1",**
        ...
      } ],

mapping is:

"tech": {
    "mappings": {
      "manual": {
        "properties": {
          "file": {
            "type": "attachment",
            "fields": {
              "content": {
                "type": "string"
              },
              "author": {
                "type": "string"
              },
              "**title**": {
                "type": "string",
                "store": true
              },
              "**name**": {
                "type": "string"
              },
              "date": {
                "type": "string",
                "store": true
              },
              "**keywords**": {
                "type": "string",
                "store": true
              },
              "**content_type**": {
                "type": "string",
                "store": true
              },
              "content_length": {
                "type": "string",
                "store": true
              },
              "**language**": {
                "type": "string",
                "store": true
              }
            }
          }
        }
      }

I am able to see the content of the indexed document:

   "_source": {
      "file": {
        "_content": "JVBERi0xLjcNJeLjz9MNCg==M......=ZW5kc3RyZWFtDWVuZG9iag1zdGFydHhyZWYNCg==MTE2DQo=JSVFT0YNCg=="
        "_name": "12-240.pdf",
        "_title": "12-240.pdf",
        "_keywords": "Div42 Src580 LGE Mechtech",
        "_language": "en",
        "_content_type": "pdf"
}

but when I search

curl -XGET "http://localhost:9200/tech/manual/_search" -d'
{
  "fields": ["file.content_type", "file.title", "file.keywords", "file.language", "file.name"], 
  "query": {
   "match_all": {}
  }
}'

the response has no file.keywords, file.language, and file.name as follow, I also can't search on these fields.

 "hits": {
    "total": 1,
    "max_score": 0.30685282,
    "hits": [
      {
        "_index": "tech",
        "_type": "manual",
        "_id": "AVPx8cRH4bGZ_KKaUS0y",
        "_score": 0.30685282,
        "fields": {
          "file.content_type": [
            "pdf"
          ],
          "file.title": [
            "12-240"
          ]
        }
      }
    ]
  }

what's possible not right?
Thanks!

5

Thank you for reporting.

I was able to reproduce it with a Java test in master branch.

EDIT: actually we never implemented or documented that: https://www.elastic.co/guide/en/elasticsearch/plugins/current/mapper-attachments-usage.html

The only values you can force are _content_type, _name and _language. Not that the later only works if you set index.mapping.attachment.detect_language: true.

6

I added this PR to get this fully covered by a test.

Note that mapper-attachments is deprecated in 5.0.0 so people should then use ingest-attachment plugin instead.

7

Thanks dadoonet! I got it. Yes thought to get things to work before going to 5.0.0.

8

Hi David

I'm using ES 2.3.4 and trying to install mapping-attachments without success (downloaded from internet)! What is the correct zip file i should be using? Could u please post the correct url? I cannot figure it out!

Please help me out!

thanx valerio

9

Why hacking an old thread?

What did you try so far?
Did you read the plugin documentation which explains how to install it?

10

Hi david and thx.

1 I ve tried bin/plugin install mapping-attachements but it cannot find the file to download from different websites.
2 I ve trid to install the mapping-attachments-master.zip file locally(downloaded from github) but i get an error :ERROR: Could not find plugin descriptor 'plugin-descriptor.properties' in plugin zip
3 I've tried to add plugin-descriptor.properties filled-out manually to the zip file in the root directory, but i get another error regarding missing _source file.

I need to show the elastic search to my customer who is interested into full -text search in pdf file....

Any help is appreciated, i am using ES 2.3.4

11

If you copy and paste from the doc it should work.

mapper not mapping is probably your problem here.