Disabling a pipe through the config file vs CLI #13443
Unanswered
kudryk
asked this question in
Help: Coding & Implementations
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Should disabling a pipe in the configuration file be equivalent to do doing so through the
load
function?I'm using 3.7.2.
My config file defines a pipeline as follows:
My model package is
en_tsds_loc_fac
.If I load my model but specify that the
basin_ruler
pipe is to be disabled, I can run text through thenlp
instance, then separately through thebasin_ruler
pipe and all works I intend.If I instead specify in my config file to disable
basin_ruler
so I don't have to do so in load, thebasin_ruler
pipe does not work.Now I run the code:
Passing the doc through the basin_ruler now results in the warning:
Looking at the config system documentation, I get the impression that it should be the same.
Going through the code, when the pipeline is assembled and the component is disabled in the config, it will not be initialized in the assembly CLI code. As a result, that components patterns (it is an instance of an entity ruler) are not loaded in to this entity ruler instance, thus the warning above.
When the documentation says "loaded" but disabled, this means to me that it should still be usable as in the first case.
Should disabling a pipe in the configuration file be equivalent to do doing so through the
load
function?Thank you in advance.
Beta Was this translation helpful? Give feedback.
All reactions