Hello 👋! Question about rebuilding invocations if no code has changed. It seems that for me,

pants package

caches the state of the environment to decide when to re-run. The following holds

./pants package <targetX>  # proper building happens
./pants package <targetX>  # no building happens 
# close and open terminal, no code changes
./pants package <targetX>  # again rebuilding

is that expected? I'm on 2.16.0dev7 and quite an old

./pants

launcher script (no scie-pants yet). Thanks!

When you

pants run

a docker image target with

restartable=True

, the created container doesn't get stopped if some source files changed. a build is triggered and a new container is created and run but the old one keeps running. Is that intended behaviour?

How can I debug a python requirement that should be inferred but isnt? For example:

$ ./pants --no-pantsd dependencies integrate_ai_sdk/src/integrate_ai_sdk/base_class.py
09:49:13.34 [WARN] Pants cannot infer owners for the following imports in the target integrate_ai_sdk/src/integrate_ai_sdk/base_class.py:

  * opacus.validators.ModuleValidator (line: 17)
  * torch (line: 13)
  * torch.nn.Module (line: 14)
  * torch.utils.data.Dataset (line: 15)

If you do not expect an import to be inferrable, add `# pants: no-infer-dep` to the import line. Otherwise, see <https://www.pantsbuild.org/v2.14/docs/troubleshooting#import-errors-and-missing-dependencies> for common problems.
integrate_ai_sdk/setup_requirements.txt:reqs
integrate_ai_sdk/src/integrate_ai_sdk/exception.py
integrate_ai_sdk/src/integrate_ai_sdk/utils/class_util.py
integrate_ai_sdk/src/integrate_ai_sdk/utils/util.py
integrate_ai_sdk:reqs#boto3
integrate_ai_sdk:reqs#clkhash
integrate_ai_sdk:reqs#matplotlib
integrate_ai_sdk:reqs#numpy
integrate_ai_sdk:reqs#opacus
integrate_ai_sdk:reqs#pandas
integrate_ai_sdk:reqs#pillow
integrate_ai_sdk:reqs#pyarrow
integrate_ai_sdk:reqs#pyjwt
integrate_ai_sdk:reqs#pytorch-ignite
integrate_ai_sdk:reqs#s3fs
integrate_ai_sdk:reqs#scikit-learn
integrate_ai_sdk:reqs#setuptools
integrate_ai_sdk:reqs#smart-open
integrate_ai_sdk:reqs#torch
integrate_ai_sdk:reqs#torchvision
integrate_ai_sdk:reqs#wheel

note that

torch

and

opacus

are in the dependencies list. Its probably a module mapping thing, but i’m not sure how to determine the key, since in requirements it’s listed as

torch

and

opacus

as well.

If I run generate-lockfiles on mac and plan to deploy on linux, should I commit my lockfiles? Additionally, are there any resources on package lockfiles and does pants package lockfile follow any standard that I can read up on?

Hi all, I'm trying to update our GHA script to use scie-pants. With

pantsbuild/actions/init-pants@v4-scie-pants

, when running

pants tailor --check ::

I'm seeing this cryptic message:

Would create BUILD:
  - Add shell_sources target root

To fix `tailor` failures, run `pants tailor`.
Error: Process completed with exit code 1.

I don't have anything named

root

in my repo. Also, when running tailor command locally, it succeeds. Is this something GHA creates? Any workaround on this issue?

Hi all, another question regarding my python2 journey, I have a package that I need to be py2/py3 compatible, but I’m getting this error when generating the lockfiles:

./pants generate-lockfiles
19:13:42.49 [INFO] Completed: Generate lockfile for py2-py3
19:13:45.85 [INFO] Completed: Generate lockfile for pytest
19:13:45.85 [ERROR] 1 Exception encountered:

Engine traceback:
  in `generate-lockfiles` goal
  in Generate Python lockfile

ProcessExecutionFailure: Process 'Generate lockfile for pytest' failed with exit code 1.
stdout:

stderr:
pid 73099 -> /Users/anler/.cache/pants/named_caches/pex_root/venvs/41379f21c5956cff0d35ab1ef014bb72e1fe14dc/336e97e279ba29b2edc4ace827e24d29cc9b0af8/bin/python -sE /Users/anler/.cache/pants/named_caches/pex_root/venvs/41379f21c5956cff0d35ab1ef014bb72e1fe14dc/336e97e279ba29b2edc4ace827e24d29cc9b0af8/pex --disable-pip-version-check --no-python-version-warning --exists-action a --no-input --isolated -q --cache-dir /Users/anler/.cache/pants/named_caches/pex_root/pip_cache --log /private/var/folders/z6/y9s0fw1s0k1fmn_9cxyfllrm0000gn/T/pants-sandbox-GCDvla/.tmp/pex-pip-log.lij6zsvc/pip.log download --dest /private/var/folders/z6/y9s0fw1s0k1fmn_9cxyfllrm0000gn/T/pants-sandbox-GCDvla/.tmp/tmp8_cj67j2/Users.anler..pyenv.versions.3.10.9.bin.python3.10 pytest==7.0.1 pytest-cov>=2.12,!=2.12.1,<3.1 pytest-xdist>=2.5,<3 --index-url <https://pypi.org/simple/> --retries 5 --timeout 15 exited with 1 and STDERR:
ERROR: Could not find a version that satisfies the requirement pytest==7.0.1
ERROR: No matching distribution found for pytest==7.0.1

Hi all again 😅 I’m having this super weird issue, I had my monorepo configured with a defaUlt interpreter constraints

["==3.11.*"]

in

pants.toml

and a default resolve with some dependencies, all worked perfectly here, the issue comes when adding a py2/py3 library, I changed the interpreter constraints in

pants.toml

to

["==2.7.*", ">=3.9"]

and used it as the interpreter constraints for my newly added package too, created a new resolve called

py2-py3 = resolves/py2-py3.lock

and re-generated lock files. The problem comes when running the

check

goal on my previously working python3 only code, it takes quite a lot of time and finally errors. My question is if someone knows how adding a package with mixed interpreter constraints can cause this check goal to fail in other packages with different constraints that were previously working?

./pants check src/python::
20:19:50.13 [INFO] Completed: Building 4 requirements for requirements.pex from the resolves/python-default.lock resolve: aiohttp[speedups], prometheus-client, pyarrow, pydantic
20:19:50.13 [ERROR] 1 Exception encountered:

Engine traceback:
  in `check` goal
  in Typecheck using MyPy - mypy

ProcessExecutionFailure: Process 'Building 4 requirements for requirements.pex from the resolves/python-default.lock resolve: aiohttp[speedups], prometheus-client, pyarrow, pydantic' failed with exit code 1.
stdout:

stderr:
There was 1 error downloading required artifacts:
1. pyarrow 0.16 from <https://files.pythonhosted.org/packages/3e/00/4e099e2f1992c555771090b5971c59751fc152c68270a31bfeabd114c068/pyarrow-0.16.0.tar.gz>
    ERROR: Command errored out with exit status 1: /Users/anler/.cache/pants/named_caches/pex_root/venvs/41379f21c5956cff0d35ab1ef014bb72e1fe14dc/791e686d7c6a49493bab896fbb2645175e97bb31/bin/python /Users/anler/.cache/pants/named_caches/pex_root/venvs/41379f21c5956cff0d35ab1ef014bb72e1fe14dc/791e686d7c6a49493bab896fbb2645175e97bb31/lib/python3.11/site-packages/pip install --ignore-installed --no-user --prefix /Users/anler/.cache/pants/named_caches/pex_root/pip_cache/.tmp/pip-build-env-fa06xvk8/overlay --no-warn-script-location -v --no-binary :none: --only-binary :none: -i <https://pypi.org/simple/> -- 'cython >= 0.29' 'numpy==1.14.5; python_version<'"'"'3.7'"'"'' 'numpy==1.16.0; python_version>='"'"'3.7'"'"'' setuptools setuptools_scm wheel Check the logs for full command output.

Use `--keep-sandboxes=on_failure` to preserve the process chroot for inspection.

Is it possible to ask the

pex_binary

target to build its own wheels when only sdists are available for certain dependencies? I’m passing a specific

platforms

argument to the target and it’s incompatible with a few dependencies that are only available as an sdist and not as a wheel - if possible I’d like to avoid building my own wheels and placing them on our internal PyPI server.

Hi! I am tinkering with PyOxidizer plugin and ran into an issue. Would appreciate help. Details in the thread.

Any pointers on having support publishing docker images to two different aws ecr's and having the authentication work properly. Ideally, I would like the BUILD to dictate where to publish (to dev and prod, or just dev).

Hey there, I was able to use pants to create docker images that function well on AWS Lambda (python 3.10) and follow the best practices outlined in this article (https://blog.pantsbuild.org/optimizing-python-docker-deploys-using-pants/). Just thought I'd post the relevant build file in this thread in case anyone else is facing a similar problem. Big thanks to @enough-analyst-54434 for helping me get started on this!

Where's the latest and greatest on how dependency inference works for files/resources targets? I'd like to try it out, but I don't remember where in the docs/blogs specifies how to use it.

Hello, Am I able to use:

[docker]
run_args = [
  "-p 0.0.0.0:13000:13000"
]

with a

docker_environment

that is used for my tests? As you can see I want the docker container to expose a port

Tried the new python-dump-source-analysis goal, and it doesn't understand use of env(). Bug? `InvalidFieldTypeException: The 'image_tags' field in target apps/_demo_app/docker:demo_app must be an iterable of strings (e.g. a list of strings), but was

['latest', None]

with type

list

.`

Is there an easy way to keep requirements.txt sorted?

Hi! is there an easy way to automate installation of python packages? I had to create a script that does the following commands to add new packages to a repo:

#!/bin/sh -e
poetry add --lock $@
rm -f poetry.lock
pants generate-lockfiles ::
pants export --resolve=python-default --resolve=pytest

Is there a way to streamline this? Or am I doing things wrong?

📣 Interested in public speaking? There are lots of great opportunities to speak at conferences and meetups about build engineering, developer tooling, dev productivity, build systems, devops, CI, Python, Rust, and other Pants Cinematic Universe topics. Watch here for CFP announcements, and DM me anytime for speaker support and assistance. We love to help more people educate folks about our favorite subjects, regardless of whether a given talk is about Pants or not. Cheers!

🌟 Periodic reminder that we really appreciate: 1. Adding your company logo to https://www.pantsbuild.org/page/who-uses-pants (see link at the bottom) 2. Github stars at https://github.com/pantsbuild/pants These help grow the project and the community, which in turn helps everyone get more features and support. So thanks for contributing!

👖 Did you know: we're actively growing the Pants team all the time! We've welcomed many new Maintainers and Contributors in the past year alone. Many of whom have gone on to contribute exciting new features that you use with Pants every day. One way you can help the project continue to flourish is by pointing out folks in our community who show potential for becoming good future team members. The entry level criteria are a demonstrated commitment to Pants, and some track record of contributions to the project — manifested in any of many ways, not necessarily only via code. When someone does standout work, or shows special enthusiasm for giving back to the project, we don't want to let that fall through the cracks. We want to work with them to help meet mutual goals. So feel free to DM me anytime to point out someone to have on the radar. Thanks so much!

Are there known reasons why generating a lockfile would be really slow? I have one that went for 1000s, never returned.. ended up restarting because i accidentally saved a file. Currently it's back up to 400s again

has anyone gotten airflow stand-alone working for local dev with pants?

I'm in the process of incremental adoption of pants in our repo. I have

lint

working and I'm trying to get

check

working. It's failing due to ambiguous dependencies. We have the source code for a few python packages in this repo, so we do have multiple requirements.txt files with overlapping dependencies. We currently build one virtualenv which includes all requirements so I don't think there are any conflicts. We use this virtualenv to do typechecking and run tests. Is it possible to tell pants to use the correct requirements.txt file based on the directory structure so that I can enable

check

without breaking our existing build?

There's two ways to do this: 1. use resolves to have separate requirements files, and set the appropriate resolve per target: https://www.pantsbuild.org/docs/python-third-party-dependencies#warning-multiple-versions-of-the-same-dependency (potentially using `__defaults__`s to set it for a whole subtree) 2. only import one set of requirements and use it for the whole repo, e.g. add a parent requirement file that just has lines like

-r path/to/requirements.txt

,

-r another/path/requirements.txt

, and only one

python_requirements(...)

target in the whole tree The second one is likely to be simpler/smoother (having a single set of dependencies for the majority of the repo).

Thanks! How do I tell pants tailor to ignore some requirements files?

I think you can set some settings in

pants.toml

, either: 1.

[python] tailor_requirements_targets = false

https://www.pantsbuild.org/docs/reference-python#tailor_requirements_targets to ignore all requirements files by default (and then add the targets manually, as required) 2.

[tailor] ignore_adding_targets = [...]

https://www.pantsbuild.org/docs/reference-tailor#ignore_adding_targets to ignore specific ones I'd probably go with 1, I think

With #2 Pants won't parse those included requirements, afaik, so I don't think that will work?

I think Pants ignores

-r

lines because it expects the included requirements files to have their own

python_requirements()

targets

Ah, oops, sorry for the misleading evidence (I’ve recently done a poetry integration rather than a requirements.txt one)

But - Pants 2.16 resolves dep inference ambiguity based on source roots

So code in a given source root will take the dep provider from the same source root