Docs: clarify installation workflow and independent project structure

[Apprentice2907]

Summary

This PR improves the root-level documentation and installation workflow
for the DeepLense repository.

DeepLense is a research monorepo containing multiple independent projects.
Previously, the lack of clear installation instructions caused confusion
around dependency setup and execution (e.g. attempts to use pip install .
or assuming a single global train.py).

This PR clarifies the correct usage pattern and improves reproducibility,
especially for fresh Windows environments.

---

What this PR changes

• Adds a root-level requirements.txt containing common dependencies
  shared across multiple DeepLense projects
• Updates the root README.md to clearly explain:
  • The monorepo structure (independent subprojects)
  • The correct installation workflow using a virtual environment
  • That the repository is not installable via pip install .
  • How to identify and run entry scripts or notebooks per subfolder
• Adds a short verification step to confirm dependency installation

---

How to install and run

1. Create and activate a virtual environment:

   python -m venv venv
   venv\Scripts\activate
   

2. Install common dependencies:
   pip install -r requirements.txt

3. Navigate to a project directory:
   cd <project_directory>

4. Identify the entry point (script or notebook) and run it:

python <script_name>.py
or
jupyter notebook

There is no single global execution command for the entire repository.

5. Testing
   Tested on Windows 10 with Python 3.11 in a fresh virtual environment

Verified that core dependencies import successfully

Confirmed that project scripts run without dependency-related errors
when executed from the correct subdirectory