Adding dismod-at to conda-forge #29516

bradbell · 2025-03-24T13:20:12Z

Checklist

github-actions · 2025-03-24T13:21:35Z

Hi! This is the staged-recipes linter and your PR looks excellent! 🚀

conda-forge-admin · 2025-03-24T13:21:54Z

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipes/dismod-at/recipe.yaml) and found it was in an excellent condition.

I do have some suggestions for making it better though...

For recipes/dismod-at/recipe.yaml:

ℹ️ Recipes should usually depend on matplotlib-base as opposed to matplotlib so that runtime environments do not require large packages like qt.

_{This message was generated by GitHub Actions workflow run https://github.com/conda-forge/conda-forge-webservices/actions/runs/14091017934. Examine the logs at this URL for more detail.}

…ance_par_xam.

2025.0.1.path: fix bug in get_option_table.cpp. build.bat: chagne ehco -> echo, bld.bat -> build.bat.

conda-forge-admin · 2025-03-26T20:32:12Z

Hi! This is the friendly automated conda-forge-linting service.

I wanted to let you know that I linted all conda-recipes in your PR (recipes/dismod-at/recipe.yaml) and found some lint.

Here's what I've got...

For recipes/dismod-at/recipe.yaml:

❌ If python is a host requirement, it should be a run requirement.

For recipes/dismod-at/recipe.yaml:

ℹ️ Recipes should usually depend on matplotlib-base as opposed to matplotlib so that runtime environments do not require large packages like qt.

_{This message was generated by GitHub Actions workflow run https://github.com/conda-forge/conda-forge-webservices/actions/runs/14092879725. Examine the logs at this URL for more detail.}

conda-forge-admin · 2025-03-26T22:20:29Z

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipes/dismod-at/recipe.yaml) and found it was in an excellent condition.

I do have some suggestions for making it better though...

For recipes/dismod-at/recipe.yaml:

ℹ️ Recipes should usually depend on matplotlib-base as opposed to matplotlib so that runtime environments do not require large packages like qt.

_{This message was generated by GitHub Actions workflow run https://github.com/conda-forge/conda-forge-webservices/actions/runs/14271047660. Examine the logs at this URL for more detail.}

…test.py.

conda-forge-admin · 2025-04-04T21:00:10Z

Hi! This is the friendly automated conda-forge-linting service.

I wanted to let you know that I linted all conda-recipes in your PR (recipes/dismod-at/recipe.yaml) and found some lint.

Here's what I've got...

For recipes/dismod-at/recipe.yaml:

❌ If python is a host requirement, it should be a run requirement.

For recipes/dismod-at/recipe.yaml:

ℹ️ Recipes should usually depend on matplotlib-base as opposed to matplotlib so that runtime environments do not require large packages like qt.

_{This message was generated by GitHub Actions workflow run https://github.com/conda-forge/conda-forge-webservices/actions/runs/14274063004. Examine the logs at this URL for more detail.}

recipes/dismod-at/recipe.yaml

recipes/dismod-at/build.sh

Co-authored-by: Matthew Feickert <[email protected]>

recipe.yaml: Advance to newer upstream source.

windows.

bradbell · 2025-04-12T13:51:44Z

@matthewfeickert
All of the tests seem to be passing now, except for using the dismod-at script after it is installed on Windows.
This is being tested by recipes/dismod-at/run_test.py

The dismod_at script appears in the projects pyproject.toml file as

[project.scripts]
dismod-at = 'dismod_at:main'

There also is a dismod_at executable created from C++ , but it seems to be working fine.

The failure is happening during the run_test.py script. If you search for run_test.py: in test section of
https://dev.azure.com/conda-forge/84710dde-1620-425b-80d0-4cf5baca359d/_apis/build/builds/1223573/logs/28
you will see

run_test.py: platform.system() =  Windows
run_test.py: found \bld\bld\rattler-build_dismod-at_1744462991\work\test\run\Library/bin/dismod_at.exe
run_test.py: found \bld\bld\rattler-build_dismod-at_1744462991\work\test\run\Scripts/dismod-at.exe

So it appears to me that the dismod-at script should be in the executable path.

I do not know why it fails to execute. If you look at the original test file
https://github.com/bradbell/dismod_at/blob/master/test/user/db2csv.py
The sandbox version is running as can be seen by searching for db2csv.py: OK in the build section of
https://dev.azure.com/conda-forge/84710dde-1620-425b-80d0-4cf5baca359d/_apis/build/builds/1223573/logs/28

There are some replacements made by run_test.py to test the installed instead of sandbox version.

matthewfeickert · 2025-04-12T20:23:54Z

 │ Installing test environment
 │  Successfully updated the test environment
 │ Testing commands:
 │ (base) %PREFIX%\etc\conda\test-files\dismod-at\0>IF "" == "" (
 │  
 │  call %SRC_DIR%\build_env.bat 
 │ ) 
 │ (base) %PREFIX%\etc\conda\test-files\dismod-at\0>python "D:\\bld\\bld\\rattler-build_dismod-at_1744462991\\work\\test\\run\\etc/conda/test-files/dismod-at/0\\conda_build_script.py" 
 │ run_test.py: prefix with / and \ repalced by |
 │ D:|bld|bld|rattler-build_dismod-at_1744462991|work|test|run
 │ run_test.py: sys.path =
 │ ['D:\\bld\\bld\\rattler-build_dismod-at_1744462991\\work\\test\\run\\etc\\conda\\test-files\\dismod-at\\0', 'D:\\bld\\bld\\rattler-build_dismod-at_1744462991\\work\\test\\run\\python39.zip', 'D:\\bld\\bld\\rattler-build_dismod-at_1744462991\\work\\test\\run\\DLLs', 'D:\\bld\\bld\\rattler-build_dismod-at_1744462991\\work\\test\\run\\lib', 'D:\\bld\\bld\\rattler-build_dismod-at_1744462991\\work\\test\\run', 'D:\\bld\\bld\\rattler-build_dismod-at_1744462991\\work\\test\\run\\lib\\site-packages']
 │ run_test.py: sys.argv[0] = D:\\bld\\bld\\rattler-build_dismod-at_1744462991\\work\\test\\run\\etc/conda/test-files/dismod-at/0\\conda_build_script.py
 │ run_test.py: platform.system() =  Windows
 │ run_test.py: found \bld\bld\rattler-build_dismod-at_1744462991\work\test\run\Library/bin/dismod_at.exe
 │ run_test.py: found \bld\bld\rattler-build_dismod-at_1744462991\work\test\run\Scripts/dismod-at.exe
 │ python test/user/db2csv.py
 │ test/user/db2csv.py
 │ dismod_at example.db set option hold_out_integrand 
 │ dismod_at example.db init
 │ dismod_at example.db bnd_mulcov 0.0
 │ dismod_at example.db fit fixed
 │ dismod_at example.db fit fixed warm_start
 │ dismod-at build/test/user/example.db db2csv
 │ × error Script failed with status 1
 │ × error Work directory: 'D:\bld\bld\rattler-build_dismod-at_1744462991\work\test\run\etc/conda/test-files/dismod-at/0'
 │ × error To debug the build, run it manually in the work directory (execute the `./conda_build.sh` or `conda_build.bat` script)
 │
 ╰─────────────────── (took 71 seconds)
Error:   × failed to run test: Script failed

Traceback (most recent call last):
  File "D:\a\1\s\.ci_support\build_all.py", line 390, in <module>
    build_all(os.path.join(root_dir, "recipes"), args.arch)
  File "D:\a\1\s\.ci_support\build_all.py", line 206, in build_all
    build_folders_rattler_build(recipes_dir, platform, arch, channel_urls)
  File "D:\a\1\s\.ci_support\build_all.py", line 321, in build_folders_rattler_build
    subprocess.run(args + ["--variant-config", fp.name], check=True)
  File "D:\Miniforge\Lib\subprocess.py", line 571, in run
    raise CalledProcessError(retcode, process.args,
subprocess.CalledProcessError: Command '['rattler-build', 'build', '--recipe-dir', 'D:\\a\\1\\s\\recipes', '--target-platform', 'win-64', '-c', 'D:\\bld', '-c', 'conda-forge', '--variant-config', 'C:\\Users\\VSSADM~1\\AppData\\Local\\Temp\\tmpx7hetfvh']' returned non-zero exit status 1.
##[error]Cmd.exe exited with code '1'.

Do you have a Windows machine that you can test your build on locally? Can you get more verbose output from the failing test? Have you validated that all files needed for your tests are present?

recipe.yaml: add sqlite3 as a test requirement.

bradbell · 2025-04-13T00:21:18Z

Do you have a Windows machine that you can test your build on locally? Can you get more verbose output from the failing test? Have you validated that all files needed for your tests are present?

Yes, I have been testing on a windows machine, but just directly building and installing in a temp conda environment. I should try and run the build there and see what happens. Thanks.

bradbell · 2025-04-13T02:36:05Z

The tests ran successfully using rattler-build on my windows machine. Here is the script I used to run it

set variant=win64.yaml
set recipe=recipes\dismod-at\recipe.yaml
del /s /q output/bld
rem
rattler-build build ^
        --test native ^
        --keep-build ^
        -m .ci_support/%variant% ^
        --log-style plain ^
        --recipe %recipe%
rem
echo conda_build.sh: Done

…d run two test cases.

matthewfeickert · 2025-04-14T15:35:14Z

By

Do you have a Windows machine that you can test your build on locally?

I mean have you been running the build-locally.py locally and are you able to compare the results?

bradbell · 2025-04-14T15:55:55Z

Where are the instructions for running build-locally.py ?
I get the following in linux:

...
Building all recipes
+ python /home/conda/staged-recipes-copy/.ci_support/build_all.py
Found no recipes to build

matthewfeickert · 2025-04-14T17:38:33Z

Where are the instructions for running build-locally.py ?

In the README: https://github.com/conda-forge/staged-recipes/blob/bbcbafaed55a25652d22630d31110d7a52fcf0e0/README.md#local-debugging-with-build-locallypy

bradbell · 2025-04-14T19:08:08Z

Where are the instructions for running build-locally.py ?

In the README: https://github.com/conda-forge/staged-recipes/blob/bbcbafaed55a25652d22630d31110d7a52fcf0e0/README.md#local-debugging-with-build-locallypy

The only instructions I found on the page above for building were

pixi run build-linux:

This eventually gave the same error as above; i.e. , it does not know were the recipe.yaml is (starting in the staged-recipe.git direcotory).

matthewfeickert · 2025-04-14T19:32:09Z

The only instructions I found on the page above for building were
pixi run build-linux:

staged-recipes/README.md

Lines 48 to 55 in bbcbafa

    
           `build-locally.py` can be run with any recent Python, or via a [`pixi`](#pixi) task: 
        
           * `pixi run build-linux`: will launch a Docker container, provision all the necessary tools and build your recipe for Linux. 
        
           * `pixi run build-osx`: will provision a conda environment with the necessary tools to build your recipe for macOS. This involves fetching and caching the necessary Apple SDKs. 
        
           * `pixi run build-win`: will provision a conda environment with the necessary tools to build your recipe for Windows. 
        
           These tasks will pass any extra arguments to `build-locally.py`, including `--help`. The resulting 
        
           artifacts will be available under `build_artifacts`.

This eventually gave the same error as above; i.e. , it does not know were the recipe.yaml is (starting in the staged-recipe.git direcotory).

oh. I see now that you're also not on a development/feature branch but are on main

staged-recipes/README.md

Line 18 in 83e0a2b

2. Make a new branch from `main` for your package's recipe.

In general avoid doing this for any PR and that's going to potentially cause issues for local debugging as it figures out what recipes to build by what differs from main, which is why there was this mention added in PR #29226. For PRs you're comparing against the target branch, so things will still work, but for local builds it is going to delete the recipe.

bradbell · 2025-04-14T20:17:28Z

Now I see the following instructions:

1. Fork this repository.
2. Make a new branch from main for your package's recipe.
3. Make a new folder in recipes for your package.

Does it matter what I name the branch ?
It seems to me that it does matter what I name the folder (that will be the package name correct ?).

matthewfeickert · 2025-04-14T21:31:34Z

Does it matter what I name the branch ?

Good question. No, that's totally arbitrary. The internal logic is just that it is comparing the recipes that exist on the source branch to the target branch of main.

Mind you that the build failure in CI is something different though, so if you want to continue to debug here that's fine. Though I might suggest that you ask for advice on the conda-forge Zulip, as I don't know if I can provide anything super actionable.

It seems to me that it does matter what I name the folder (that will be the package name correct ?).

Another good question. I think this is also arbitrary, as the recipe YAML file and any other supporting files are the only things that matter in the end, but I'm not 100% sure on this.

…d_prc). run_test.py: use dismod_at.system_command_prc.

bradbell · 2025-04-16T18:08:20Z

Please note the following comment in run_test.py:

      #
      # skip
      # test/user/db2csv.py tests the installed dismod-at python script.
      # This is failing on the conda-forge test machine, but works
      # doing a local rattler-build on a windows machine. At this point
      # I think it is a problem with the conda test system.
     skip = test_file=='test/user/db2csv.py' and platform.system() =='Windows'

You can set skip equal to False and run the test on your local system as a check; see
#29516 (comment)

bradbell · 2025-04-16T18:26:59Z

@conda-forge/help-python-c, ready for review

xhochy · 2025-04-27T07:45:38Z

recipes/dismod-at/recipe.yaml

@@ -0,0 +1,80 @@
+# conda-verify warnings:


Please remove all the comments here. They seem to be quite "empty"

xhochy · 2025-04-27T07:46:50Z

recipes/dismod-at/build.bat

+%BUILD_PREIX%\Library\lib\pkgconfig;^
+%BUILD_PREIX%\Library\share\pkgconfig;


This is not correct. pkg-config should not point to BUILD_PREFIX. Please remove this.

@xhochy

Please remove all the comments here. They seem to be quite "empty" .
This is not correct. pkg-config should not point to BUILD_PREFIX. Please remove this.

Done in the following commit:
b44d02ff

recipe.yaml: remove empty comment.

bradbell added 6 commits March 23, 2025 05:55

dismod-at: first version that builds on Linux.

e9bc565

dismod-at: use patch file instead of sed to patch source.

8505ff3

dismod-at: use patch (instead of sed) to patch source code.

54a37cc

Merge remote-tracking branch 'refs/remotes/origin/main'

0e2e7fd

dismod_at: convert build to use ninja and add build.bat.

30d8cd0

Merge branch 'conda-forge:main' into main

e8871b1

bradbell added 7 commits March 24, 2025 09:50

dismod-at: 2025.0.1: CMakeLists.txt: splite3.pc is misisng on windows.

b01a5bf

dismod-at: 2025.0.1: init_command.cpp: fix (not -> !).

b7ec52c

dismod-at: fix another if expression, reduce chance of failure in bal…

c9b1132

…ance_par_xam.

dismod-at: run_test.py: list sitepackages directory.

312a5bd

2025.0.1.path: fix bug in get_option_table.cpp. build.bat: chagne ehco -> echo, bld.bat -> build.bat.

dismod-at: build.bat: add c++17 compiler flag.

382e128

dismod-at: advance to 2025.0.2 to fix some conda install problems.

9aa32a3

dismod-at: put python in host section (try to fix pip install).

5b43460

dismod-at: Trying to get proper version of pytyhon to run pip.

e57fa7b

bradbell added 8 commits April 3, 2025 08:54

dismod-at: 2025.0.2.patch: C++: close open sqlite databases.

645b954

dismod-at: dismodat.py.bat: wrapper for dismodat.py on DOS.

12a7ef5

dismod_at: build.bat: change python3 -> python.

a8a6315

dismod-at: another attempt to get windows sandbox test to run.

8479c6b

dismod-at: 2025.0.2.patch: fix path in dismodat.py, fix exit in user_…

3e91ff4

…test.py.

dismod-at: build.sh: change python -> $PREFIX/bin/python.

421350f

Merge branch 'conda-forge:main' into main

333fc53

dismod-at: put python in host:, print python version and sys.path.

0485e15

matthewfeickert suggested changes Apr 5, 2025

View reviewed changes

recipes/dismod-at/recipe.yaml Outdated Show resolved Hide resolved

recipes/dismod-at/recipe.yaml Outdated Show resolved Hide resolved

recipes/dismod-at/build.sh Outdated Show resolved Hide resolved

bradbell and others added 2 commits April 7, 2025 06:09

Update recipes/dismod-at/recipe.yaml

f19a555

Co-authored-by: Matthew Feickert <[email protected]>

Update recipes/dismod-at/recipe.yaml

92a2650

Co-authored-by: Matthew Feickert <[email protected]>

bradbell added 3 commits April 11, 2025 17:25

recipe.yaml: balance_pair_xam.cpp: try 3 times instead of 2.

934b9e0

run_test.py: Improve messaging, change test_file in place.

2e0e8ea

recipe.yaml: Advance to newer upstream source.

run_test.py: Trying to figure out why dismod-at script not found on

446f069

windows.

run_test.py: add default exception message.

f540fa2

recipe.yaml: add sqlite3 as a test requirement.

bradbell added 3 commits April 13, 2025 17:06

run_test.py: change run_test: OK -> run_test.py: OK.

c08da51

run_test.py: fix error message, make sandbox2installed a funciton, an…

5ac590f

…d run two test cases.

remove spaces at end of lines

767de5e

Merge branch 'main' into main

9f13f30

bradbell added 4 commits April 15, 2025 02:48

run_test.py: system_command: better error messaging.

70d8475

recipe.yaml: advance to dismod_at-2025.4.16 (for better system_comman…

5212040

…d_prc). run_test.py: use dismod_at.system_command_prc.

run_test.py: skip test that fails on windows.

1f3b380

run_test.py: fix platform.system -> platform.system().

822098f

github-actions bot added python-c review-requested labels Apr 16, 2025

xhochy requested changes Apr 27, 2025

View reviewed changes

bradbell added 2 commits April 29, 2025 07:00

build.bat: remove setting of PKG_CONFIG_PATH.

b44d02f

recipe.yaml: remove empty comment.

Merge branch 'main' into main

d445682

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Adding dismod-at to conda-forge #29516

Adding dismod-at to conda-forge #29516

bradbell commented Mar 24, 2025

github-actions bot commented Mar 24, 2025

conda-forge-admin commented Mar 24, 2025 •

edited

Loading

conda-forge-admin commented Mar 26, 2025

conda-forge-admin commented Mar 26, 2025 •

edited

Loading

conda-forge-admin commented Apr 4, 2025

bradbell commented Apr 12, 2025

matthewfeickert commented Apr 12, 2025

bradbell commented Apr 13, 2025

bradbell commented Apr 13, 2025 •

edited

Loading

matthewfeickert commented Apr 14, 2025

bradbell commented Apr 14, 2025 •

edited

Loading

matthewfeickert commented Apr 14, 2025

bradbell commented Apr 14, 2025

matthewfeickert commented Apr 14, 2025

bradbell commented Apr 14, 2025

matthewfeickert commented Apr 14, 2025

bradbell commented Apr 16, 2025 •

edited

Loading

bradbell commented Apr 16, 2025

xhochy Apr 27, 2025

xhochy Apr 27, 2025

bradbell Apr 30, 2025

		%BUILD_PREIX%\Library\lib\pkgconfig;^
		%BUILD_PREIX%\Library\share\pkgconfig;

Adding dismod-at to conda-forge #29516

Are you sure you want to change the base?

Adding dismod-at to conda-forge #29516

Conversation

bradbell commented Mar 24, 2025

github-actions bot commented Mar 24, 2025

conda-forge-admin commented Mar 24, 2025 • edited Loading

conda-forge-admin commented Mar 26, 2025

conda-forge-admin commented Mar 26, 2025 • edited Loading

conda-forge-admin commented Apr 4, 2025

bradbell commented Apr 12, 2025

matthewfeickert commented Apr 12, 2025

bradbell commented Apr 13, 2025

bradbell commented Apr 13, 2025 • edited Loading

matthewfeickert commented Apr 14, 2025

bradbell commented Apr 14, 2025 • edited Loading

matthewfeickert commented Apr 14, 2025

bradbell commented Apr 14, 2025

matthewfeickert commented Apr 14, 2025

bradbell commented Apr 14, 2025

matthewfeickert commented Apr 14, 2025

bradbell commented Apr 16, 2025 • edited Loading

bradbell commented Apr 16, 2025

xhochy Apr 27, 2025

Choose a reason for hiding this comment

xhochy Apr 27, 2025

Choose a reason for hiding this comment

bradbell Apr 30, 2025

Choose a reason for hiding this comment

conda-forge-admin commented Mar 24, 2025 •

edited

Loading

conda-forge-admin commented Mar 26, 2025 •

edited

Loading

bradbell commented Apr 13, 2025 •

edited

Loading

bradbell commented Apr 14, 2025 •

edited

Loading

bradbell commented Apr 16, 2025 •

edited

Loading