Re: [Discuss-gnuradio] Python OOT Block Selectable Types

[Marcus Müller]

What should is if you use strings to initialize the data type,

in_sig = [numpy.dtype(string_passed_via_constructor_argument)]

and use an option field that gives you "uint16", "float32", "byte",
"complex64"…


Cheers,
Marcus


On 13.05.2016 18:13, Martin Braun wrote:
> You can definitely do that. Passing in some numeric or string value as
> the type to your block and then mapping to the correct numpy type would
> definitely work. Maybe you can even select numpy types from a GRC file,
> not sure if a GRC enum will allow you to choose numpy types.
>
> M
>
> On 05/12/2016 03:16 PM, Richard Bell wrote:
>> Martin,
>>
>> I like the first idea. Would it be possible to create a new variable in
>> the xml file that is assigned numpy.float32 when the user selects the
>> float option, or is assigned numpy.complex64 when the user selects the
>> complex option? Another words, I don't want the user to have to select
>> the port type AND have to type in the python type of the port, that
>> should be clear from the port type drop down selection. I'm just not
>> sure how to work this in the xml file, or if its even possible.
>>
>> Rich
>>
>> On Thu, May 12, 2016 at 11:11 AM, Martin Braun <address@hidden
>> <mailto:address@hidden>> wrote:
>>
>>     You can either make this a block argument (i.e., you ask the user to
>>     pass in numpy.float32 or whatever), or, to keep consistent with the C++
>>     way of doing things, you can make a block argument called 'sizeof' and
>>     then create a dict that maps sizes to numpy.<FOO> values, and use that
>>     to set in_sig.
>>
>>     M
>>
>>     On 05/11/2016 02:04 PM, Richard Bell wrote:
>>     > I've created a python based OOT block and I want to give the user the
>>     > ability to select the input/output type. I know how to do this for C++
>>     > based blocks, but I'm stuck on what to do with the python blocks
>>     > signature, as it doesn't ask for the item_size like the C++ blocks
>>     do. I
>>     > have to give it something explicit like
>>     >
>>     > in_sig = [numpy.float32]
>>     >
>>     > How do I write the generic sig that depends on what the user
>>     selected in
>>     > GRC?
>>     >
>>     > Note: The algorithm in the body of this block doesn't change depending
>>     > on type selected. It's a simple block.
>>     >
>>     > Thanks,
>>     > Rich
>>     >
>>     >
>>     > _______________________________________________
>>     > Discuss-gnuradio mailing list
>>     > address@hidden <mailto:address@hidden>
>>     > https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
>>     >
>>
>>
>>     _______________________________________________
>>     Discuss-gnuradio mailing list
>>     address@hidden <mailto:address@hidden>
>>     https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
>>
>>
>
> _______________________________________________
> Discuss-gnuradio mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/discuss-gnuradio