Feature: Support standard pyproject.toml dependency formats in flet build

[karim-moon]

Duplicate Check

• I have searched the opened issues and there are no duplicates

Describe the requested feature

When running flet build, the build process fails to correctly parse dependencies specified in pyproject.toml in several scenarios:

1. List format dependencies cause errors
   e.g.

    torch = [
  { version = ">=2.7.0", platform = "linux", markers = "extra == 'cuda'"}, 
  { version = ">=2.7.0", platform = "linux", markers = "extra != 'cuda'"},
  { version = ">=2.7.0", platform = "darwin"},
  { version = ">=2.7.0", platform = "win32"},
]

2. Private PyPI repositories are not supported
   e.g.

some_dep = { version = ">=0.1.6", source = "private-pypi" }
[[tool.poetry.source]]
name = "private-pypi"
url = "{private pypi url}/pypi/pypi/simple"

3. Wildcard versioning (*) causes failures

4. Are there any plans to address these issues in an upcoming release?

5. In the meantime, are there any alternative workarounds we could use besides the simple string version specification approach?

Suggest a solution

No response

Screenshots

No response

Additional details

No response

[FeodorFitsner]

That's a good one. We need to investigate that. It's obviously impossible to replicate this entire functionality in Flet.

[yst-one]

That's a good one. We need to investigate that. It's obviously impossible to replicate this entire functionality in Flet.

Hope flet can come out a dedicated flet toml configuration list