I have read the documentation of ElasticSearch and I am familiar with its endpoints and how to maintain Clusters,Nodes,Indices,etc.
Now I am planning to use it in an Asp.net web api and I am a little bit confused about what is the best way to implement it.
I found that ElasticSearch provides two options to work with .Net application:
1)Nest
2)ElasticSearch.net
So, my first question which way is preferred to integrate my web application with ElasticSearch so I can get advantage of the power of Elasticsearch in searching/analyzing/managing big amount of data and what are the advantages and disadvantages of each one ?
My second question I have read that Nest is a high level client and Elasticsearch.net is a low level one.What does that mean?
My third question do the above libraries expose all ElasticSearch endpoints or are there limitations?
My fourth question is there another way to integrate ElasticSearch in my web application other than the above two libraries?
I would appreciate any references,tutorials so I can get started with(because the documentation of ElasticSearch website is not clear).
Check out the documentation for the .NET clients as I think it answers all your questions.
Both NEST and Elasticsearch.Net expose all of the APIs within Elasticsearch, with the former mapping each request and response to a type so you don't need to do this yourself.
You can of course write your own implementation to interact with Elasticsearch over HTTP. Use whichever you're most comfortable with, but personally I'd recommend starting with NEST as it handles a lot of things for you such as mapping inference, connection behaviour with round-robin, etc. (disclaimer: I work on the Elasticsearch .NET clients).
For an example tutorial of integrating with a .NET (Nancy) web application, check out NuSearch.
Related
I have been working in the Django rest framework since three years on wards. Recently reading about FAST API frame work. FAST API looking like a very good framework for the current REST API service development. Seems like few things are missing from FAST API which are there in django.
Pagination
class based views.
mock testing
class based views.
struggled allot to provide the lookup for encrypted data in the django rest framework.
Could any one help me that how can we achieve above features with FAST API.
Currently in Jan 2020. FastAPI has a much smaller and different featureset to Django Rest Framework (DRF).
DRF is built on Django you have all the features Django available to you as well as convenient tools for building REST API's.
With FastAPI the framework way is more minimal it provides a simple, quick way to create and test a set of OpenAPI compatible endpoints. With a focus on speed and asynchronous style request-response handling.
However does not include ORM or migrations, admin, accounts etc and all the other 'batteries included' stuff you get with Django.
To achieve the functionality you have listed you'd right now you'd have to roll-your-own or use some pre-existing async python packages and adapt for your needs.
I have been getting a bit lost in the creation of my program architecture and I want to take a step back to see if I'm approaching it correctly.
I am wondering if my setup makes sense. I'm starting to think it doesn't.
I am creating intranet applications (We were creating Internet applications, but now the scope has changed). We use an onsite Active Directory (Windows Server 2012 R2). We have a SQL Server Database.
I have been building Front End Angular applications and ASP.NET Web API's to push and pull data. I am now implementing Authentication with Auth0 and it's been a nightmare.
What kind of program architecture would you setup in this scenario?
Much Appreciated.
SQL Server + Asp.Net Web Api + Angular JS forms a perfect architecture for building Single Page Applications (SPAs). This architecture is useful for building desktop like web applications, i.e. apps that runs over web but works like desktop apps.
If you can be more specific about the problem you are facing, you will be able to get better recommendations from so.
This architecture is widely adopted in many scenarios such as SPAs. With it, you will be able to keep your front-end highly decoupled from your backend services being able to support multiple front-ends on the same set of services and run quite a few integration scenarios.
Some of the downsides of such an approach will be the extra layer of complexity added to the application (which might force you to write more tests and handle different failure scenarios that wouldn't happen otherwise, for an example) and authentication routines since you will need to authenticate two heterogeneous environments (the .NET/IIS one and the JS/Angular one).
As for the authentication pain, token-based auth schemes seem the current way to go (such as Auth0) since they let you keep and send an environment-agnostic token which will be used by different layers of your architecture.
In that sense, your architecture makes sense.
However, since you're feeling some pain in its implementation, you might want to ask yourself if you really needed all of these. When you choose an architecture, you do so trying to accomplish some specific goals (multiple front-ends? specific performance requirements? maintainability? auditability?) and the more goals you try to accommodate in your architecture the more complex will become up to a point where the pains start outweighing the benefits.
So, what were you trying to achieve in the first place?
I am working on a REST API to communicate with one of the server which is communicating using oneM2M architecture. Does anybody have an idea how to decode its JSON format in dotnetcore?
If your program needs to communicate with a oneM2M CSE then you need to implement at least parts of the Service Layer Core Protocol. This is specified in oneM2M's TS-0004. You should also have a look at TS-0001, the Functional Architecture document, which beside of the general architecture, describes the individual resources, their structures and their relationships.
You can find these specifications here: http://www.onem2m.org/technical/published-documents.
If you are new to oneM2M then you definitely want to have a look at the "Getting Started" pages and the Application Developer Guide.
I'm looking at enterprise JavaFX, and how to integrate JavaFX with server-side code. In the last few weeks I've done a certain amount of research in DataFX and Open Dolphin, and downloaded some videos, as well as looking at a couple of other frameworks. For example I've looked at the video on DataFX at:
https://www.youtube.com/watch?v=EN4fo6x0DcQ
However, although this video and others I've looked at explain how to set up a client application that connects with a server, I've found very little information on how to put together some server-side code that the client can connect to. Sure, one could use JAX-RS, but why re-invent the wheel? In the first instance I would like to put together some really simple server-side code that some test client-side code can connect to using DataFX or one of the other frameworks. The aim eventually is to get a client using JavaFX to communicate with a server.
My second question is that of the various frameworks available, is DataFX the best to use for a simple application?
I have experience with a Glassfish server hosting a JSF application, and it may be useful to have such a server hosting a JSF application communicating with a browser as well as communicating with a client JavaFX, as that way I can test out that the communication with the JavaFX application.
The latter is a bit of an aside, and my main questions are where can I get information on server-side programming for this, and the best frameworks to use?
Many thanks in advance.
If you are able to manage yourself the client-server communication you can pick any JavaFX Application Framework listed there:
https://github.com/mhrimaz/AwesomeJavaFX
Any of them allows to separate UI code from communication code.
As I'm author of JRebirth, I can advice you to create some RemoteService (extends Service and providing JAX-RS facilities or whatever) to perform this job.
If you search an all-in-one library managing client AND server side, DataFX + OpenDolphin is probably the most advanced one.
I'm the author of DataFX & Dolphin Platform (https://github.com/canoo/dolphin-platform). Both are valid frameworks that fill fit to your needs. Maybe a combination of both - Dolphin Platform as remoting layer between client and server and DataFX to define the routing and mvc based views on the client.
Some days ago I copied all the DataFX sources to GitHub (https://github.com/guigarage/DataFX) and currently trying to create a new version based on the modules that are maintained by me. Maybe I will extract the MVC related stuff and create a new framework based in it, we will see. What I currently can say is that I plan to work on it the next month next to the Dolphin Platform since I think a combination of both will be a good fit.
I would use this combination today to create applications but yeah, I'm the main developer of both frameworks so the choice is quite easy for me ;)
As you mentioned JSF I think that Dolphin Platform is a perfect match for you since one idea of the framework is to have a modern successor for JSF that can be used to create desktop & web based applications but provide managed controllers on the server. We provide a cool maven based jumpstart (Maven archetype) that will give you a quick introduction and a runnable client / server application with a desktop and web client in 2 minutes: https://canoo.github.io/dolphin-platform/#_dolphin_platform_jumpstart
I have a few projects coming up that have a number of endpoints or clients that can hit the same data. For instance a site might have...
A asp.net MVC based end user facing website
A web based adminitration back end that can allow specific, limited updates from some users in situatiosn where a full client isn't useful (mobile web etc)
A full on rich client for administration so we can use touch and other techniques to really make the user experience for content management shine - these may be silverlight or full on WPF apps, as needed
The question is... whats the best way to connect all that? Right now I use a multiple project split for the MVC.
ProjectName.Core - this project has all the common models, all the repository classes and all the helper classes
ProjectName.Web - an MVC project that references the Core to pull in the repositories and models it needs
ProjectName.Admin.Web - another MVC project dedicated to the admin interface that references the Core to pull in the repositories and models it needs
(which will ultimately live on a seperate subdomain from the end user facing site)
Then the story peters out int he sense of clear guidance. When the time comes to build a WPF / Silverlight project to hit the data I can do one of the following to the best of my understanding now...
Convert the "Core" to provide a RIA style DomainService, and attempt to alter the MVC projects to make use of it. However there is little guidance on using MVC with RIA, RIA is in its infancy AND the WPF to RIA story is also still only thinly documented
Do the same as above, but using WCF. However WCF prings with it async complexity that I really want RIA to hide for me.
Fall back 20 yards and just bolt plain old web services onto the Core on top of the Repositories and Models I already have. This seems... old school :)
Any thoughs and input are welcome including pointers to examples and documentation. I want to make the best decisions I can now, and am coming up to speed fast on RIA and WCF so I can but community input is always helpful.
Thanks!
Take a look at ADO.NET Data Services (formerly known as Project “Astoria”).
On my way to work today I was listening to a .NET Rocks! podcast, "Stephen Forte on Data Access Options", and they very excited about this, especially for scenarios like the one you describe.
It's interesting stuff, and something I would check out sometime very soon.
I think there's something to be said about plain old WCF services, these can make use of your domain services and repositories and expose a model more appropriate for services. Too often I've found that simply exposing the domain model on the wire ends up with a duplication of logic on both the client and the server.
My advice would be for some sort of service layer, this has the logic of shaping your domain model into appropriate types for the wire. \
Ideally I'd like to be able to share my Domain Services between the client (WPF / Silverlight) and the server (ASP.NET MVC) and have different underlaying repositories (Linq to NHibernate / Astoria). Difficult with the asynchonous nature of Silverlight.
For the curious, it certainly looks like RIA Services is the win here. Build a single DomainService then consume it everywhere. Brad Abrams covers a lot of this ground at bit.ly/94fFx - it really helped.