[conda] How to tweak an installer created by conda constructor, advice appreciated

[Propadovic Nenad]

Hello,

I'm not even sure this is the right list to ask about the conda constructor
installer generator, but I hope so, as I see none better.

(Why do I ask for the things I ask for? Because the cusomer asks for them).

Some moths ago I built an installer, constructor based, Windows+Linux,
installing Python, Git, Pip, and some more, then clones a certain
repository with a bunch of Python-files, and then installes some more
python packages, via Pip into the freshly created Python installation.
Works like a charm, with a little help of post_install scripts, for both
Windows and Linux.

Now the focus has moved away from Linux to Windows, and I have new
requirements:
1) I need the Python installation directory not to be the one specified by
the user in the course of the installation, but a subdirectory of it. I
guess I could probably just move this directory in the
post_install-batch-file, but I'm reluctant to do so, it just feels wrong.
2) I need to remove the installer-GUI-step enabling the user to add the
chosen directory to "Add Anaconda to my PATH environment variable" and
"Register Anaconda as my default Python 2.7". The result of the
installation process is meant to be reather isolated and not to interact
with the Windows environment more than necessary.

I see no constructor keywords enabling me to do either. (But maybe I'd just
need to understand conda better, with all of its possibillities).
I assume I'd be able to do both tweaks if I were able to edit the
NSIS-template-file, but is there even a possibility to do so, ideally in a
a way that is not a dirty hack? (It's not enough that the installer
generator works on my machine...)
If there's some third way of doing what I want, I'd be happy to hear about
it.

The versions I'm using: constructor==2.0.2, conda==4.4.10, Python 2.7,
Miniconda 2 4.3.31 (64 bit), on Windows 10 professional.

Thanks in advance,

Nenad