Error when updating using Poetry: Broken dependencies in versions >=7.1

[theodotk]

trafficstars

Description

When I update notebook to version 7.1 or higher, my poetry (ver 1.3.2) does not recognize the dependencies on jupyterlab and jupyter-server, and deletes them:

> poetry add notebook="^7.1"
                                                                                                    
Updating dependencies
Resolving dependencies...

Writing lock file

Package operations: 0 installs, 1 update, 28 removals

  • Removing anyio (4.1.0)
  • Removing argon2-cffi (23.1.0)
  • Removing argon2-cffi-bindings (21.2.0)
  • Removing arrow (1.3.0)
  • Removing async-lru (2.0.4)
  • Removing fqdn (1.5.1)
  • Removing isoduration (20.11.0)
  • Removing json5 (0.9.14)
  • Removing jsonpointer (2.4)
  • Removing jupyter-events (0.9.0)
  • Removing jupyter-lsp (2.2.1)
  • Removing jupyter-server (2.12.1)
  • Removing jupyter-server-terminals (0.5.0)
  • Removing jupyterlab (4.0.9)
  • Removing jupyterlab-server (2.25.2)
  • Removing notebook-shim (0.2.3)
  • Removing overrides (7.4.0)
  • Removing python-json-logger (2.0.7)
  • Removing pywinpty (2.0.12)
  • Removing rfc3339-validator (0.1.4)
  • Removing rfc3986-validator (0.1.1)
  • Removing send2trash (1.8.2)
  • Removing sniffio (1.3.0)
  • Removing terminado (0.18.0)
  • Removing types-python-dateutil (2.8.19.14)
  • Removing uri-template (1.3.0)
  • Removing webcolors (1.13)
  • Removing websocket-client (1.7.0)
  • Updating notebook (7.0.8 -> 7.2.1)

And without those one can't launch notebook.

I see that some of those dependencies are present in the pyproject.toml in this repository, so the problem might be elsewhere.

Context

• Operating System and version: Windows 10 ver 22H2
• poetry version: 1.3.2
• notebook version: 7.1 and 7.2

[RRosio]

Thank you for your contribution @theodotk! Are you able to install Jupyter Notebook via the recommended instructions?

[RRosio]

Hi @theodotk just following up with you, were you able to try installing as recommended in the documentation?

[theodotk]

Hi @RRosio, sorry for the delay. Yes, installing with pip works perfectly, and I didn't check it with Anaconda.

My issue is not installation per se, but maintaining a system of packages, for which I use poetry dependency manager, and switching to Anaconda is not a viable short-term option. I guess it's not the intended way of installation, but it used to work