I have heard that it should be possible to add a dynamic dictionary, holding a set of terms or word list when you train your model. This should be possible in the Translation Hob today.
How to do this using V3 or NMT in portal.customtranslator.azure.ai using custom translator project?
/Mats
At this time the dictionary feature is not available in the NMT custom translator but it is a feature we are working on and will release in a future update.
Related
I already own a translation engine I created some years ago. I would like to know if I can use that engine to train a new engine using Microsfot Custom Translation.
I haven't bought Translator Text yet or any Azure producsts, so knowing whether if I can 'recycle' my old engine or not is something I will consider before buying any product.
You can call the translator API v2.0 but it is not recommended. At the same time the translator hub had been retired since May. Which means you cannot improve the result with your document.
You train a Custom Translator system with previously translated documents. If your previous engine was a statistical system or neural translation system, you can use the same training material. If not, you could train on material that you have translated by hand or with your previous engine, or a mix of both. Custom Translator languages are limited to the languages Microsoft provides a base engine for.
I am an App developer with no experience with AutoCAD at all, and for my current project I need to convert DWG files into Spatial Data - ESRI Shapefile / GeoJSON, etc.
Given that DWG is a proprietary format owned by Autodesk, the Autodesk Forge API becomes my only option to interpret such file. I have been evaluating the Model Derivative API for a week now, and what I found is that it won't work for my purposes. It has only 3 output formats (f2d, f3d, rvt) for DWG files and none of them fits my purpose.
For this reason, I started to look at others API's in the platform and ended up finding the Design Automation API which might be helpful, however it lacks documentation.
One of the examples in the API is the PlotToPdf activity that exports a DWG into a PDF. I wonder if there is a way to export other formats other than PDF.
I tried to find the possible "commands" in the interface by listing all the activities, workItems and AppPackages but none of these end-points returned me any useful information.
So here are my questions?
How can I can convert a DWG into a non Autodesk format? Being more specifically, the output must be an open format that I can interpret without using proprietary tools.
If the answer for the first question is Autodesk Forge API, which one should I use?
In case I should be using Design Automation API for that, where can I find complementary documentation once the one provided is lacking details?
I would say that Design Automation API is the best option if you don't have (or don't want) AutoCAD running (or any other DWG compatible desktop app).
To be complete: Design Automation includes an instance of AutoCAD running on the cloud, where you can upload a .DWG with a sequence of commands (script) and a .NET custom app to execute more advanced routines. Then download the results. By default, neither AutoCAD nor Design Automation can export to GIS formats (like those you mentioned), only some verticals, like Civil 3D or Map 3D, but these are not available on Design Automation. As you mentioned, Model Derivative API (via REST calls) don't have what you need, it's focused on metadata, not the specifications of the geometry.
Design Automation API supports the same commands AutoCAD Console does. For DXF creation, you can use DXFOUT command, your script would be something like (not tested):
FILEDIA
0
DXFOUT
outputfilename.dxf
QUIT
As per comment, it's a huge effort to just convert to .DXF. I'm not familiar with other libraries do perform the same task.
Now I'm not an expert on GIS data formats, but you may consider move from .DWG to .DXF then an approach like described here.
I don't quite understand this point that you are making in the comments:
I would only use the API to convert DWG to DXF, which means a huge waste of resources considering the monthly cost of the API and the time to develop and maintain the interface.
Can you clarify? Are you saying that transaction cost for Design Automation is too high? Or are you saying that development cost associated with deploying the script that Augusto gave you is too high?
Thanks
Right now the derivativeApi is the way to go for this: You can convert any input format from this list (https://developer.autodesk.com/en/docs/model-derivative/v2/overview/supported-translations/) to the "intermediate format" svf and the convert the svf file to obj for example. This will also create a material file which can be downloaded seperately. I am evaluation the forge api right now and can produce obj from dwg files right now.
Is there a way to be able to employ Xamarin Maps in MVVM pattern by frameworks like Prism? For e.g. to be able to capture pinning on the map in the view model without having to breaking MvvM architecture? Or type in an address to reflect on the map?
Following this article about bindable maps may help until Xamarin folks implement it as a built-in feature.
I am currently implementing a set of web services intended to be used by JavaScript clients and am using Google Gson for JSON serialization/deserialization. My service contracts are thus defined in Java classes. I now need a way of generating documentation of these message interfaces in a way intended to be used by JavaScript developers.
Ideally, I would like a tool along the lines of JavaDoc that generated JSON examples, JSON schemas or other Web developer friendly documentation of my types, that could easily be integrated with Maven to produce HTML based documentation for types with a certain annotation.
I can point them to the generated JavaDoc, but it will require them to understand and hunt through all private fields to work out how to format requests, and also does not take into account Gson type converters etc.
Are there any relevant tools available in this area?
I have used Swagger in the past. It is simple to implement, yet generates powerful API doc.
Please look at this documentation. I think this is what you are looking for.
http://petstore.swagger.wordnik.com/
Is it possible to convert a standard PHP based widget integrated with google maps API to a flash based widget? If yes, what are the ways?
I expect it will generally be a manual process, because the code styles are so different, code-to-code conversion would be likely to be difficult/fragile/error-prone. The only automatic way that I could conceive of is a html-to-flash-ui conversion, which would get you a static look or prototype to start with, but wouldn't help you in the creation of the active/reactive/dynamic elements of the new flash widget.
A quick google search brings up a few paid references to html to flash conversion, as well as a few pages mentioning some libraries that might be worth pursuing, if just as a shortcut to creating a prototype that you would have to refine:
http://drawlogic.com/2008/01/11/as3-flash-to-html-conversion-library-htmlwrapper/
and here:
http://osflash.org/flashml
Since there is one free library, there may quite likely be others as well.