|
|
|
@ -174,7 +174,7 @@ As a general rule, it tends to happen when the gibberish-level is set to one of
|
|
|
|
|
It is caused when the start-word is itself a sentence, and the text generation loop has not started. |
|
|
|
|
For more information, see the GenerateMarkovText function in DataServices.fs. |
|
|
|
|
The reason you would use these functions ove the "Clean Text Generation" function below is these are faster to finish. |
|
|
|
|
If you can tolerate the occasional extra bit of text and the functions below are too slow, you use these functions. |
|
|
|
|
If you can tolerate the occasional extra bit of text and the functions below are too slow, use these functions. |
|
|
|
|
*) |
|
|
|
|
let ps_result1 = Pancake.RequestRandomTextAsync() |
|
|
|
|
let ps_result2 = Pancake.RequestTextAsync 5 10 |
|
|
|
@ -202,10 +202,10 @@ printfn "CLEANED TEXT RESULT: %s" ps_cleanText
|
|
|
|
|
|
|
|
|
|
(* Clean Text Generation |
|
|
|
|
====================================================================================================================== |
|
|
|
|
These functions produce text which have gone through the "clean" process. |
|
|
|
|
These functions produce text which have gone through the "cleaning" process. |
|
|
|
|
This means the results will go through an extra step and remove any erroneous sentences. |
|
|
|
|
This is unlike the "Basic Text Generation" functions above. |
|
|
|
|
The trade-off here is these functions can be slower to finish executing. |
|
|
|
|
The trade-off here is these "cleaning" functions can be slower to finish executing. |
|
|
|
|
If you need your results to match the number of sentences you requested exactly, you should use these functions. |
|
|
|
|
This, also, applies if you can tolerate the extra processing time. |
|
|
|
|
*) |
|
|
|
|