Order of monstache golang functions execution

[Joey0538]

Hi @rwynn I'm new to Monstache, its a great tool....
Could you please clarify the order of execution regarding these golang functions for building monstache plugins:

func Map(input *monstachemap.MapperPluginInput) (output *monstachemap.MapperPluginOutput, err error)

func Filter(input *monstachemap.MapperPluginInput) (keep bool, err error)

func Pipeline(ns string, changeStream bool) (stages []interface, err error)

func Process(input*monstachemap.ProcessPluginInput) error

My guess for the call order is : Pipeline (called once for each collection & finding documents based on some criteria) -> Filter (whether to continue processing further for indexing or drop based on doc field criteria) -> Map (transform doc before indexing) -> Process (Maybe log or modify existing doc for processing)

Please let me know if I'm right about this, I plan on writing a go plugin for resuming direct-read-namespaces from a given timestamp (got this idea from you while I was scrolling through the open and closed issues)

And another doubt, if I'm using the same go.mod file from the repo, to be precise same go version 1.19, why can't I directly mount the plugin.so built locally and supply it to the docker image via volume mounts? when I attempt to do this I'm unable to run the plugin with some _vprintf error but, if I build the plugin through the dockerfile, I'm able to run the plugin? Its like for each change in the plugin code, I have rebuild the entire docker image for monstache to be able to run the plugin, is there no other way to directly mount the plugin file to the docker image?