I have a requirement for a document management system to handle pdf,word,xls,ppt with semantic search.
I started looking into elasticsearch for the same and stumbled on Apache JacKrabbit and subsequently on OpenKM and Hippo. Even though core features like versioning exists in Jackrabbit, I need some pointers on how to go about this.
I need help navigating through the following concerns:
Should I just use elasticsearch and elasticsearch attachment plugin or use Jackrabbit with MySQL backend and use Elasticsearch to index the documents.
Or should I use OpenKM?
Any pointers would be greatly appreciated. This would finally require App integration.
Update Logically, using ElasticSearch for Search makes sense. But I figure that I cannot use that as primary datasource. What are the best options from storage(primary) Apache JackRabbit with MySQL? As all features are prebuilt in OpenKM, would this be a better option?.
What is it you want to achieve? Are you looking to manage making the documents available, is it about managing the content in documents? ES, or any search engine, is generally not a primary data source.
I can't give you any advice wrt OpenKM (neither for or against). Whether Hippo is a match depends on your case which I need to know more about.
Related
I have 2 questions:
Can I run spring-data-elastic v4.0.1.RELEASE (with org.elasticsearch:elasticsearch 7.6.2 ) with ES client running on 7.4.0??? If not, what combination can I use for 7.4.0 client? We are migrating to AWS and I need to use 7.4.0 version of client.
I have parent/child relationship (configured as join datatype field). Could pls somebody provide a documentation or explain, how to use either ElasticsearchRestTemplate or ElasticsearchOperations to correctly insert/update both parent and child records?
Thank you.
Best regards,
Robert
ad 1): from the Elasticsearch documentation I can't at the moment find anything in the breaking changes sections that would prevent using a 7.4.0 client library, but that does not mean there aren't any. But that does not mean that there aren't any. Recently there was a breaking change in the Java classes (from 7.7 to 7.8) and I got the information:
our compatability focus is on the HTTP APIs and we don’t offer any guarantees on the code itself. There’s more background here: https://github.com/elastic/elasticsearch/issues/22707#issuecomment-274163711
So I'd say, write a small test app and with the corresponding libraries, start a local ES 7.4 and test it.
ad 2): adding the join-type mapping ang implementing the corresponding inserts etc. is currently worked on and will hopefully be available in version 4.1.
I'm new to elasticsearch and am still trying to set it up. I have installed elasticsearch 5.5.1 using default values I have also installed Kibana 5.5.1 using the default values. I've also installed the ingest-attachment plugin with the latest x-pack plugin. I have elasticsearch running as a service and I have Kibana open in my browser. On the Kibana dashboardI have an error stating that it is unable to fetch mappings. I guess this is because I havn't set up any indices or pipelines yet. This is where I need some steer, all the documentation I've found so far on-line isn't particularly clear. I have a directory with a mixture of document types such as pdf and doc files. My ultimate goal is to be able to search these documents with values that a user will enter via an app. I'm guessing I need to use the Dev Tools/console window in Kibana using the 'PUT' command to create a pipeline next, but I'm unsure of how I should do this so that it points to my directory with the documents. Can anybody provide me an example of this for this version please.
If I understand you correctly, let's first set some basic understanding about elasticsearch:
Elasticsearch in it's simple definition is a "Search engine". so you need to store some data, and then elastic will help you to search using a search criteria, and it will retrieve relevant data back
You need a "Container" to save your data to, and elastic has this thing like any database engine to store your data, but the terms are somehow different. for example a "Database" in sql-like systems is called "Index", and what you know as "table" is called "Type" in elastic.
from my understanding, you will need to create your index (with or without mappings) to have a starting point, and I recommend you to start without mappings just to "start" and get things working, but later on it's highly recommend to work with "mappings" if applicable, because elastic is smart, but it cannot know more about your data than you do
Because Kibana has failed to find a proper index to start with, it has complained and asked you to either provide a syntax for index names, or a specific index name so it can infer the inline mappings and give you the nice features of querying, displaying charts, etc of your data, so once you create your index, you will provide that to the starting page of Kibana, and you will be ready to go.
Let me know if you need something more specific to your needs :)
I'm currently working on a POC with Couchbase, using Spring Data to put & get documents on/off a bucket on a cluster.
As I'm working in a big company, I'm lucky they gave me a bucket, but still I don't have the admin rights on the cluster, so I only have access to the bucket.
But as I'm digging into the Spring Data documentation, I'm not able to find a way to retrieve documents without creating views on the server. (I'm getting errors like "Unknown query param" ). Nevertheless with couchbase java sdk i'm able to, through n1ql queries, but the use of the Spring data layer is mandatory.
The answers I found always point me to the server-side function direction
ex : https://stackoverflow.com/a/30928169/3744307
What I would like to find, is a way to add a repository method like
List findReceiptByAccount(String Account)
without having to specificly declare the function server-side.
Is this possible, or have I to send a request to the administrators to create functions for me everytime I have to add a findByX method?
Thanks for your time,
What version of CB is it ?
I think that prior to 4.5, a n1ql access (which you seems to have) is enough to build your index yourself !
With Spring Data Couchbase 2.x that would use a N1QL index in the background, and it would work with a single primary index (although having 1 index per repository entity class would be best for performance). Maybe you can ask your admin to create that index once?
I am going to use logstash+ES+kibana for my project. I want to know how to use this framework for multi tenants. Can any one explain me how after the authentication Kibana query the elastic search index and load in Kibana's dashboard? Can I restrict kibana to look for a specifix index of Elastic search for a particular user or some-id? Anybody has tried this?
Thnx
You could, but depending on your use case it is probably not a good idea. There are a few gotchas, particularly regarding security and separating the users. First Kibana is just javascript running in the browser. So whatever Kibana is allowed to do so is your user. You can however have a separate index pattern for each "user", but elastic search does not provide you any ways of authenticating a users or authorizing a user access to a specific index. You would have to use some sort of proxy for this.
I recommend http://www.found.no/foundation/elasticsearch-in-production/ and http://www.found.no/foundation/elasticsearch-security/ for a more in depth explanation.
Create an index for each tenant.
In this way you can use a proxy (like the app the hosts kibana) to intercept the request and return a settings that includes the index to use.
The value that specifies the index to use can be the logged in user or you can get that value somewhere else.
To separate even more the data, you can use a prefix in each index name, and then when you specify an index you can use a pattern to take all the index related to only certain kind of data/entities.
Hope this help.
Elasticsearch announced today a plugin they are working on that should provide security features to ES product. Probably, this will contain ways of restricting access based on roles and users setup at cluster and indices level. If this happens I see no way for them not to extend this security layer to Kibana, as well. Also, it seems this plugin will have a commercial version only.
I am developing an app on java. It has mongo db at the back end which stores files(in gridFS). I use spring framework to interact with mongo db.I want to search for text present in the stored documents(pdf,doc,txt files). I know mongo db supports full text search(from 2.4).My question is
does spring framework support Full text search? or should we take the help of solr or lucene?
If both of the above is possible which is a better option?
Wat about indexing?I dont have much knowledge regarding indexing in full text search
When will 2.4 be available?
1 Spring does not support full text search within its core features, however, within the spring-data project there are two sub-projects that allow the interaction with solr and elasticsearch, both of them are full text search engines built in the top of apache lucene, for detailed information look at these links:
https://github.com/dadoonet/spring-elasticsearch
https://github.com/SpringSource/spring-data-solr
2 Depends of your needs, lucene is a low level library, while elasticsearch and solr are out of the box search engines built in the top of lucene, I think that elasticsearch provides better integration with mongodb, through the mongodb-river which support indexing of gridFS attachments. Look at these links:
http://www.elasticsearch.org/
https://github.com/richardwilly98/elasticsearch-river-mongodb/
3 You need to clarify this question.
4 I don't know when the mongodb version 2.4 will be available, but don't forget that the full text search is still an experimental feature, and also I think that this feature still does not support gridFS.
MongoDB text search will not pull text out of PDF, DOC, or, for that that matter, any files that are stored in GridFS. From the perspective of MongoDB, GridFS files are uninterpreted binary.
If you'd like to use MongoDB's new text search capabilities to search in different file types, you'll need to do the work in your application to extract text from these files and add the text into documents that you explicitly insert into MongoDB. You can use existing libraries such as Apache Tika to do the heavy lifting. Note that Tika is what Solr/Lucene uses do text extraction from rich-text document types.
As for text search indexing in MongoDB, please refer to the release notes here