Poetry doesn't check extra sources for package

• I am on the latest Poetry version.
• I have searched the issues of this repo and believe that this is not a duplicate.
• If an exception occurs when executing a command, I executed it again in debug mode (-vvv option).

• OS version and name: Ubuntu 20.04.3 LTS (Focal Fossa)
• Poetry version: 1.1.12
• Contents of your pyproject.toml file:

[tool.poetry]
name = "shit"
version = "0.1.0"
description = ""
authors = ["None"]

[tool.poetry.dependencies]
python = "^3.7"
tensorflow = "1.15.5"

[tool.poetry.dev-dependencies]

[[tool.poetry.source]]
name = "tensorflow"
url = "https://tf.kmtea.eu/whl/stable.html"

[build-system]
requires = ["poetry-core>=1.0.0"]
build-backend = "poetry.core.masonry.api"

Issue

Poetry is incapable of finding a package from a custom source. The package is also part of pypi, but not for arm64 architectures, so I need to add a custom source to search for it, but it doesn't try to look for it there. This results on poetry add tensorflow==1.15.5 failing on arm64 architectures.

I have added a custom source where this is available (and it works if you do pip install tensorflow==1.15.5 -f https://tf.kmtea.eu/whl/stable.html) but poetry never looks in that source for the package, so the build fails with:

Unable to find installation candidates for tensorflow (1.15.5)

  at ~/.poetry/lib/poetry/installation/chooser.py:74 in choose_for
       70│             links.append(link)
       71│
       72│         if not links:
       73│             raise RuntimeError(
    →  74│                 "Unable to find installation candidates for {}".format(package)
       75│             )
       76│
       77│         # Get the best link
       78│         chosen = max(links, key=lambda link: self._sort_key(package, link))

I have also tried with:

[tool.poetry.dependencies]
python = "^3.7"
tensorflow = {version = "1.15.5", source = "tensorflow"}

but it doesn't work either.

Any tips? Thanks!

Start with the package-selection path reported in poetry/installation/chooser.py, then reproduce the failure using the provided pyproject.toml and poetry add tensorflow==1.15.5 command. Confirm that the custom source configuration is considered and that the command finds an installation candidate for tensorflow 1.15.5.