Python: sharing common code among a family of scripts

My preference would be a separate "bin" or "scripts" directory, with subprojects as libraries / packages:

projectroot
  |
  |- scripts
  |
  |- lib
  |    |
  |    `- matcher.py
  |    `- merger.py
  |    `- subproject1
  |    `- subproject2
  |    `- subproject3

The idea being your scripts can each reference any subprojects necessary as usual packages. And your subprojects can also reference each other with imports.

You can then also have a main or shared script that sets up the subproject packages for you, if that helps.

I suggest putting trivial "launcher" scripts at the top level of your project, and making each of the subproject folders into packages. The modules in the packages can import each other or common code can be factored out into a common package.

Here's what the structure would look like, if we assume the various merger modules can be refactored into a shared version:

projectroot
  |- script1.py # launcher scripts, see below for example code
  |- script2.py
  |- script3.py
  |
  |- common
  |    |- __init__.py
  |    |- merger.py # from other packages, use from ..common import merger to get this
  |
  |- subproject1
  |    |- __init__.py # this can be empty
  |    |- script1_main.py
  |
  |- subproject2
  |    |- __init__.py
  |    |- script2_main.py
  |    |- script2_matcher.py
  |
  |- subproject3
       |- __init__.py
       |- script3_main.py
       |- script3_converter.py
       |- script3_matcher.py

The launcher scripts can be very simple:

from subproject1 import script1_main

if __name__ == "__main__":
    script1_main.main()

That is, all it does is import the appropriate "scriptN_main" module and run a function in it. Using a simple script may also have some small benefits for script startup speed, since the main module can have its compiled bytecode cached to a .pyc file, while scripts are never cached.

Note: I renamed your modules, swapping _ characters in for the . characters. You can't have a . in an identifier (such as a module name), since Python expects it to indicate attribute access. That meant those modules could never be imported. (I'm guessing that this is an artifact of the example files only, not something that you have in your real code.)