labeler | An action for automatically labelling pull requests | BPM library

At least, you can pre-compile your regular expressions, and not compile them in every loop cycle, which is essentially what the re.* top-level functions do.

First, instead of declaring the pattern string in a variable, compile the pattern string into a regular expression, then and there, in a variable:

You should inform the paths that way:

In the declaration of LInstruction, you specify a lifetime 'a of the &str references stored inside the label HashMap field. In the struct's impl, you call this lifetime 'a (using the syntax impl<'a> LInstruction<'a>). To put it simply, this implies that each &str reference stored in the label HashMap field of every self instance has a lifetime of 'a. In other words, when you try and insert a &str key into self.label (as you do in add_labeler), the reference you add must also have a lifetime of 'a.

In add_labeler, you are attempting to add the label_name to self.label as a key. Since label_name = token.trim_start_matches("(").trim_end_matches(")"), label_name is a slice of the same String that token is a slice of. Because both slices point to data from the same source, they have the same lifetime. As token is an element of tokens, its lifetime is specified in the tokens parameter signature.

Due to lifetime elision, the lifetime of the &str references in token is different than the lifetime of &self. The unelided signature would be fn add_labeler<'b>(&self, tokens: Vec<&'b str>) -- recall that the &str references in self.label (and by extension &self) have a lifetime 'a as defined in the impl. This is now a problem, as token would have a lifetime 'b, which does not satisfy the lifetime requirement 'a needed for it to be safely added to self.label. The solution to this is to specify that the references in tokens must indeed have the same lifetime as the references stored in self.label, by explicitly adding the 'a lifetime to its type (tokens: Vec<&'a str>).

The compiler error tells you exactly where to put the lifetime as well:

The following will allow you to make plots that plot the percentage of group with characteristic variable while plotting the results with the group name and the count.

lambda returns an anonymous function, you just want to evaluate your function to get the result so you should write:

Here's my understanding based on the error message:

Given you are using the LocalModel(manifestPath: manifestPath) API, it is expecting a legacy TFLite model format where the label map is provided through a separate text file and the model.tflite itself does not contain the label map. That's why your file before your model update works.

To use your updated model.tflite (which seems to contain the lab map inside its metadata), I think you can try the following to use the model.tflite file directly with the custom models API without going through the filename.json manifest:

The documentation provided (https://firebase.google.com/docs/ml/ios/label-images-with-automl) contains outdated information. ML Kit has fully deprecated and removed the GoogleMLKit/ImageLabelingAutoML pod in its recent versions. That pod is now replaced by the GoogleMLKit/ImageLabelingCustom pod. The latest version of MLKitImageLabelingCustom pod is 1.2.0. Please refer to the full migration guide here:

https://developers.google.com/ml-kit/vision/image-labeling/automl/migrate-automl

You should really try to use a dedicated yaml parser such as yq but if for what ever reason you cannot use such as tool, you can remove the last line with the following sed statement:

In this specific case this is correct but you need to be carefull with it since there is no consistent way to do this for all types of resources.

Historically kubectl had --export flag that was generating yamls ready to apply, but it got depricated because of many bugs. Check out the issue on k8s github repo for more details.

There is also another way to export the resources if you used kubectl apply to create it.