locker: move export logic into locker for reuse

[abn]

This change, moves common functionality from exporter into the locker
to allow for sharing of logic that can generate DependencyPackage
instances given a a list of requirements using the lock data.

This change set also resolves #3115.

Resolves: #3112
Resolves: #3115
Resolves: #3160
Closes: #3125

[abn]

Moving to draft as this requires new tests and also existing tests need to be validate.

[marns93]

@abn Thanks for the fix. I've tested it locally in our repositories and it works as expected.
Looking forward to the next Poetry release. :)

[zyv]

Another test case from our users @ moneymeets/python-poetry-buildpack#16 - I didn't check, but that should work properly with this PR because you check whether the locked package is None, but doesn't work with my hack @ #3125 :

[[package]]
category = "main"
description = "Fast and simple WSGI-framework for small web-applications."
name = "bottle"
optional = false
python-versions = "*"
version = "0.12.18"

[[package]]
category = "main"
description = "WSGI HTTP Server for UNIX"
name = "gunicorn"
optional = false
python-versions = ">=3.4"
version = "20.0.4"

[package.dependencies]
setuptools = ">=3.0"

[package.extras]
eventlet = ["eventlet (>=0.9.7)"]
gevent = ["gevent (>=0.13)"]
setproctitle = ["setproctitle"]
tornado = ["tornado (>=0.2)"]

[[package]]
category = "main"
description = "Python client for Redis key-value store"
name = "redis"
optional = false
python-versions = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*, !=3.4.*"
version = "3.5.3"

[package.extras]
hiredis = ["hiredis (>=0.1.3)"]

[metadata]
content-hash = "184994068e190b2720ed44423cc8673e8300ceae582e90301adbe7502e598b42"
lock-version = "1.0"
python-versions = "3.8.6"

[metadata.files]
bottle = [
    {file = "bottle-0.12.18-py3-none-any.whl", hash = "sha256:43157254e88f32c6be16f8d9eb1f1d1472396a4e174ebd2bf62544854ecf37e7"},
    {file = "bottle-0.12.18.tar.gz", hash = "sha256:0819b74b145a7def225c0e83b16a4d5711fde751cd92bae467a69efce720f69e"},
]
gunicorn = [
    {file = "gunicorn-20.0.4-py2.py3-none-any.whl", hash = "sha256:cd4a810dd51bf497552cf3f863b575dabd73d6ad6a91075b65936b151cbf4f9c"},
    {file = "gunicorn-20.0.4.tar.gz", hash = "sha256:1904bb2b8a43658807108d59c3f3d56c2b6121a701161de0ddf9ad140073c626"},
]
redis = [
    {file = "redis-3.5.3-py2.py3-none-any.whl", hash = "sha256:432b788c4530cfe16d8d943a09d40ca6c16149727e4afe8c2c9d5580c59d9f24"},
    {file = "redis-3.5.3.tar.gz", hash = "sha256:0e7e0cfca8660dea8b7d5cd8c4f6c5e29e11f31158c0b0ae91a397f00e5a05a2"},
]

[abn]

@zyv is the pyproject.toml available for that?

[zyv]

@abn

[tool.poetry]
name = "bruh"
version = "0.1.0"
description = "Service that implements REST-ful API service"
authors = ["Ilya Breitburg <me@breitburg.com>"]

[tool.poetry.dependencies]
python = "3.8.6"
redis = "^3.5.3"
bottle = "^0.12.18"
gunicorn = "^20.0.4"

[build-system]
requires = ["poetry>=0.12"]
build-backend = "poetry.masonry.api"

[zyv]

❤️

[marcinzaremba]

This change seems to introduce a serious incompatibility in regards to how an export file looks for local file dependencies. Namely instead of using a path exactly how it is specified in pyproject.toml it is always transformed into an absolute one which makes a resulting requirements.txt file less portable as it in corporates an absolute path that is dependant of a running environment.

For example for Poetry 1.1.2:
Having a local file dependency declaration pyproject.toml:

telemetry = { file = "vendor/telemetry-1.1.1-py3-none-any.whl" }

would yield:

telemetry @ vendor/telemetry-1.1.1-py3-none-any.whl \
    --hash=sha256:777828bea2a9b9b4125cedcb044afdb74b1dc8f3cc35c04b86ab072caf09d9d2

On the other hand after these changes (for Poetry 1.1.3) resulting requirements.txt looks like:

telemetry @ file:///Users/<my-local-path>.../vendor/telemetry-1.1.1-py3-none-any.whl; python_version >= "3.7" and python_version < "4.0" \
    --hash=sha256:777828bea2a9b9b4125cedcb044afdb74b1dc8f3cc35c04b86ab072caf09d9d2

Is this an expected change/behavior that we need to get used to for the following versions or is it a regression in comparison with earlier versions?

[github-actions]

This pull request has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.