Migrate flowrep.tools and parts of flowrep.workflow

[liamhuber]

Depends on pyiron/flowrep#185.

With this, all remaining external references to flowrep go through flowrep.models.api, and we can remove flowrep.workflow (parts of which now live here) and flowrep.tools (all of which now lives here) and move models up to the root. @samwaseda, I copied the functions with minimal changes, brought their tests with them, and tagged you as a co-author.

I had to touch a bunch of the semantikon tests, but only in two trivial ways pertaining to running with the flowrep toy WfMS:

1. It imports stuff from the recipe, so I had to move these out of <locals>
2. It only takes **kwargs not *args, **kwargs, so I added keywords to all the run calls

To put this in context, I've been working so far from the bottom-up with structuring the unstructured data. Here we have a structured-to-dict conversion point. Now, I would work top-down inside semantikon to have the callers of the dictionaries start to consume structured data instead (either flowrep structured data directly, or new classes defined here), until we don't have any raw dictionaries left. I've found this and my last couple PRs here very promising that this can be done with good continuity in the structure of the test suite and how things look and behave for users -- i.e. I think we can get this migration done without breaking anything at any point along the road.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[github-actions]

👈 Launch a binder notebook on branch pyiron/semantikon/flowrep-dict

[liamhuber]

@samwaseda, on my local machine visualize, qudt, ontology, and analysis are all throwing mypy errors too. TBH I'm neither super concerned nor super motivated to fix them here.

Otherwise this is working nicely, and relies only on the flowrep.models.api interface now.

[codecov]

Codecov Report

❌ Patch coverage is 99.70149% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 97.64%. Comparing base (70b1133) to head (be6d36d).
⚠️ Report is 45 commits behind head on main.

Files with missing lines	Patch %	Lines
semantikon/flowrep_dict.py	99.69%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #410      +/-   ##
==========================================
+ Coverage   97.47%   97.64%   +0.16%     
==========================================
  Files          10       11       +1     
  Lines        1584     1908     +324     
==========================================
+ Hits         1544     1863     +319     
- Misses         40       45       +5     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Copilot]

Pull request overview

Migrates Semantikon’s remaining direct dependencies on flowrep.tools / flowrep.workflow to the flowrep.models.api surface, introducing an in-repo compatibility layer (semantikon.flowrep_dict) to convert Flowrep live nodes into the legacy nested-dict format consumed by Semantikon.

Changes:

• Add semantikon/flowrep_dict.py to convert Flowrep live node trees to the legacy workflow dict format and to host hashing/metadata utilities previously sourced from Flowrep.
• Update Semantikon workflow decorator/runtime path to use flowrep.models.api (parsers, schemas, wfms) and the new converter layer.
• Update unit tests, example notebook, and environment/dependency pins to Flowrep 0.3.0 and keyword-only run calls.

Reviewed changes

Copilot reviewed 14 out of 14 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
semantikon/flowrep_dict.py	New compatibility layer: live-node → legacy dict conversion + hashing/metadata utilities
semantikon/workflow.py	Switch workflow decorator/run to Flowrep v0.3 APIs and new dict conversion
semantikon/ontology.py	Update hashed-node import source to semantikon.flowrep_dict
semantikon/converter.py	Update function-metadata import source to semantikon.flowrep_dict
tests/unit/test_flowrep_dict.py	New tests covering flowrep_dict conversion, graph helpers, hashing, and utilities
tests/unit/test_workflow.py	Update serialization import and adjust workflow validation expectations
tests/unit/test_ontology.py	Move workflow fixtures out of locals and convert .run() calls to keyword args
tests/unit/test_analysis.py	Convert .run() calls to keyword args; move workflow fixtures out of locals
pyproject.toml	Pin flowrep==0.3.0
notebooks/knowledge_graph.ipynb	Remove Flowrep legacy dict example; update .run() call style
docs/environment.yml	Pin flowrep=0.3.0
.ci_support/lower_bound.yml	Pin flowrep=0.3.0
.ci_support/environment.yml	Pin flowrep=0.3.0
.binder/environment.yml	Pin flowrep=0.3.0

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[liamhuber]

@samwaseda I made some small changes to get it mypy compliant and extended the coverage far enough to get the green ticks -- there are some untested logical branches inside the digraph conversions that still aren't hit, but I don't grok those well enough to see how to construct examples that reach them.

[samwaseda]

@samwaseda I made some small changes to get it mypy compliant and extended the coverage far enough to get the green ticks -- there are some untested logical branches inside the digraph conversions that still aren't hit, but I don't grok those well enough to see how to construct examples that reach them.

@copilot Address the last part of Liam's comment

[Copilot]

@samwaseda I've opened a new pull request, #414, to work on those changes. Once the pull request is ready, I'll request review from you.

[samwaseda]

It looks like knowledge_graph.ipynb was not executed. Is there a particular reason for this?

[liamhuber]

It looks like knowledge_graph.ipynb was not executed. Is there a particular reason for this?

Looks like it ran fine to me. From the latest CI build notebooks:

[samwaseda]

No I mean you uploaded an unexecuted notebook

[liamhuber]

No I mean you uploaded an unexecuted notebook

Oh, that is just a user error. Or something that happened during merge conflict resolution. Either way, I'll execute and upload. I also find it nice to see the executed one when browsing on GitHub.

[samwaseda]

Here Copilot correctly pointed out that memory locations etc should not be included in the notebooks because they change after each execution. You can delete print(workflow_dict) from the notebook and execute it again

Edit: I did it myself