Scenario:
Blob storage: contains pdf, word, image files (about 70 files)
I used default fields and predefined skills to create an Azure search instance through the Azure Portal.
But the results for querying any text in these files is not very good. I made content and key phrases as searchable and retrievable. I tried to use Lucene analyzers but was not a great help.
The main concern is if I type even a letter for example "u" in the search explorer, it returns the file. as per my understanding, there is no such word in my files. so what is it doing?
How to refine the search? and also how to manipulate the result?
I am not an expert in document processing. So using the unstructured documents in the blob instead of JSON formatted documents.
another thing, how to define some field in the index, let's say chapter-name or title name which can relate to the PDF chapters/title name?
Please suggest me some ideas or some example links. I am using .net core to develop this.
use custom skill set to extract the fields which you required and make sure those fields are defined in index.
Related
I'm new to Azure Cognitive Services, and while I'm pretty sure it can help me solve my problem, I don't quite understand which part of it to use for it...
Here's what I want to do:
We have blog posts, say ~1k, and those blog posts all have categories and tags (multiple each). What I want to do, is to "guess" the right categories/tags for each article based on the content, and then present that to the editor as a suggestions at the time of input ("looks like this article is about: health, well-being, ..."). The ~1k articles we already have in the system are currently correctly tagged/categorized, so I'd like to use these a data source for this "guessing".
I've used Azure Search before, and it seems like some combination of EntityRecognition and KeyPhraseExtraction might be a way in the right direction? Azure Cognitive Services also seems to have an API that supports TextAnalytics that would do something similar. I'm a bit confused about why these are two different things (or are they not?)
This also seems like an entirely common problem (matching text against pre-defined categories based on other text that is categorized), so I'm wondering if I'm just missing an obvious solution here?
Thanks in advance.
I think the Azure Cognitive Text Analytics API is your best bet as you are looking for real-time analysis prior to tagging/categorizing for storage.
Text Analytics could return a list of named entities that you could map to your available tags/categories and present to the user.
Azure Cognitive Search requires an indexer and skillset to process target text with an end result of storing the processed results to an index specifically for searching.
I am working on file search engine functionality.I need your suggestions in designing my application.
I am using elastisearch as framework to implement my functionality.
My primary feature is to enable file search based on file name , file type, size and date of creation. I also need to enable searching based on content of file.
Please suggest what can be best possible file to do the indexing and extract file data.
Also since file can be deleted/updated so i would need to generate the index again in some time interval so how can i monitor any change in directory.
I am using SAMBA as my file storage system.
To have the search option in file content you need to index the file into elasticsearch index.
Look in to the Mapper Attachment plugin and this will help you to index the files and make it searchable.
Step01: install the plugin in to your elasticsearch cluster
Step02: convert the files as byte[] and sent it to elasticsearch index
Step03: Now you can search using the file content using normal queries.
Note: This will work only for text based files like pdf, word (doc,docx) & text files. if the pdf files contains text in images it will not be searchable.
The problem briefly: I would like Sitecore to index the contents of PDFs using Solr's built in functionality (supplied by Tika). I'm not sure how to configure Sitecore's indexing to use this feature in Solr(Tika). (I think I need to write a custom indexer.)
I'm working with Sitecore 7 (7.1 Update 1) and want to index content from PDFs (or other rich media types). I'd like to index this data for search purposes.
I have Solr (4.6.1) installed and working with Sitecore 7. When I index my site it saves all of the documents to the correct Solr core, and I can successfully retrieve these documents for display.
Using curl, I can send a PDF to my Solr instance and get it indexed.
curl "http://localhost:8983/solr/update/extract?literal._id=doc1&uprefix=attr_&fmap.content=attr_content&commit=true" -F "myfile=#sample.pdf"
This works, and I can read this content in my Sitecore web project and display it in views, so I know I can get access to this data. However, I would like the data to be attached to the items that I have uploaded in Sitecore.
I'd like something like this to happen when I upload a PDF to the Sitecore Media Library and publish the item, or at least when I re-index the site.
I'm currently walking through the following tutorial to learn some things about writing custom indexing (here is a link to part 1):
http://www.sitecore.net/Community/Technical-Blogs/Getting-to-Know-Sitecore/Posts/2013/04/Sitecore-7-Search-Provider-Part-1-Manually-Triggered-Indexing.aspx
Thanks for you patience.
For Sitecore, when handling media data, Lucene and Solr needed to index the content in a consistent way (so that you could switch between them if needed and still get data indexed in the same way). As Tika integration is very much a Solr thing it was decided that both should use the general windows concept of IFilters for indexing (http://en.wikipedia.org/wiki/IFilter)
This means that as long as you have the correct IFilter for that mime type installed on the machine that is doing the indexing, then the '_content' computedfield will be populated with the output.
This doesn't mean you can't use the Solr Tika integration but it is not supported by default and would be a customization.
It would be very simple to:
Disable the '_content' computed field
Set up a publish pipeline processor that looks at each item being published
Check if it is a media item
Check to see if it is a PDF
Issues a command to push the content to your Solr server for indexing by Tika.
You may want to see what results you get by using an IFilter, if the results are close enough to what you want then you can go with that, if Tika is producing better results for you then you should be able to switch to that, although you would probably have your media content indexed in a separate Solr core so you would lose any Sitecore specific metadata around the document.
Some blog posts that might be helpful:
http://www.samjgriffin.com/blog/2013/11/06/sitecore-7-pdf-and-document-content-search/
http://www.sitecore.net/Community/Technical-Blogs/John-West-Sitecore-Blog/Posts/2013/04/Sitecore-7-Indexing-Media-with-IFilters.aspx
I second the recommendation to use Sitecore's built-in MediaItemContentExtractor + IFilter approach, unless you've already ruled that out for some reason (IFilter difficulties, perhaps). If an IFilter is not an option, or if you're interested in the other approach regardless, I would integrate Tika a little differently than Stephen suggested, though.
The tutorial you referenced deals with writing your own search provider - i.e. replacing the built-in provider entirely. You should be able to leverage Sitecore's Solr provider and accomplish what you need with something lighter: a computed index field. The built-in media extractor mentioned above uses this approach, which enables you to put just about anything into your index during the normal indexing process. Here is a blog post from John West that walks through creating a basic computed index field: Sitecore 7: Computed Index Fields.
In short, write a class that implements IComputedIndexField and represents content extracted from PDFs or other rich documents. In your implementation of the ComputeFieldValue method:
Call GetMediaStream() on the document.
Pass the stream to Solr in an extract-only command and capture the result.
Return that result to store it in the computed index field.
If you need the media content to be in a particular existing index field, then configure the computed field as a copy field (see 3.3.3) into the existing field. Otherwise, configure your search to reference the computed field.
The main drawback here is the expense of passing the extracted content back and forth, rather than committing it directly to the index in a single step. Depending on your index size and contents, that may not be an issue for you.
One other potential option would be a post-rebuild task to add media content to the existing indexed documents. I am not certain this would work. It depends on knowing the IDs of the media items' documents and committing the rich document content in partial document updates, which this person was unsuccessful in attempting. If you try this, be sure to execute it in the indexing:end event prior to HTML cache clearing.
Whatever approach you take, if you want to work with Tika on a higher level than cURL, have a look at SolrNet's implementation of ExtractCommand and related classes.
If you could upgrade your site into sitecore 7.2 in which the media items content will be indexed automatically and there is no need to install the related IFilter, You should read the following:
Media Content Indexing in sitecore 7.2
I'll try to briefly describe my problem and task.
My task is to create search engine for different types of file (only text file types) pdf, word, odf, xml but not html.
I have got little experience with lucene about year ago i wrote simple full text search using lucene and hibernate search. That was simple project. But now i have got very difficult task with searching.
We are using java 1.7 and glassfish 3 and i have to concentrate only server side approach not client ui. Ther is my three major problem :
1) All files is stored on webdav server, but information about file name , id file typ etc are stored into database (postgresql) so when i creating index i need to use both information. As a result of query i need only return file id from database. Summary content of file is stored in server but information about file is stored in database so we must retrieve both.
2) Secondary problem it that each file has a level of secrecy. But major problem is that this level is calculated dynamically. When calculating level of security for file we considering several properties. The static properties is files location, the folder in which the file is, but also dynamic information user profiles user roles and departments . So when user "Maggie" is logged she can search only files "test.pdf" , "test2.doc" etc but if user "Stev" is logged he have got different profiles such a Maggie so he can only search some phase in file "broken.pdf", "mybook.odt". test2.doc etc ..... . I think that when for example user search phase "lucene +solr" we search in all indexed documents and after that filtered result. But i think that solution is is not very efficient. What if results count 100 files , so what next we filtered step by step each files ? But i do not see any other solution. Maybe you can help me and lucene or solr have got mechanism to help.
3) Last problem is that some files are encrypted. So that files must be indexed only once before encryption ! But i think that if we indexed secure files so we get security issue. Because all word from that file is tokenized.
I have not got any idea haw to secure lucene documents and index datastore ? its possible ...
Also i have got question that i need to use Solr for my serarch engine or using only lucene and write own search engine ? So as you can see i have not got problem with indexing , serching but with security files and files secured levels.
Thanks for any hints and time you spend for me.
For Indexing both the File and Metadata of the file from DB check ExtractRequestHandler
You can pass the metadata attributes and the file to be indexed as a single request and it would be stored as a single document in lucene index.
For Security, One of the options is to store the Users/Roles who have access to the Files/Documents within the Solr index.
So you can always filter the results with the user/role to retrieve only the those results.
Make you Solr url secured so that Users don't have a direct access to the documents.
Also check for SOLR-1872
For encryption, Solr and underlying Parser Tika does provide handling for the Encrypted files by providing additional parameters.
Apache Solr uses the Apache Tika which uses the Bouncy Castle generic encryption libraries for extracting text content and metadata from encrypted PDF files. See http://www.bouncycastle.org/ for more details on Bouncy Castle.
Maybe I am missing something but is there any way to use the new text search features as described in the 2011 presentation http://www.youtube.com/watch?v=7B7FyU9wW8Y (approx. 30min mark) with Objectify, Entities, and Java? I realize it is an experimental release but the text search features that are present don't seem to cover the full extent they discussed in the presentation. I don't want to have to write my own code to manage the creation, updates to documents. But I don't currently see another way??
Currently, you can't use Full Text Search to search through entities in Datastore; you'll need to create search documents in a search index to use the Full Text Search API, as described in these docs.