Redun pipeline is broken with orjson ~3.8

[falexwolf]

Hey Rico,

When downloading, installing, and running the main redun pipeline, one gets this error message:

Error message

``` [redun] Execution duration: 0.84 seconds [redun] *** Execution failed. Traceback (most recent task last): [redun] Job eea2a311: File "/Users/falexwolf/repos/bioinformatics-pipeline-tutorial/wf/workflow.py", line 27, in main [redun] def main( [redun] amino_acid = 'C' [redun] enzyme_regex = '[KR]' [redun] executor = [redun] input_dir = 'fasta/' [redun] max_length = 75 [redun] min_length = 4 [redun] missed_cleavages = 0 [redun] Job 299f076a: File "/Users/falexwolf/repos/bioinformatics-pipeline-tutorial/bioinformatics_pipeline_tutorial/lib.py", line 269, in plot_count_task [redun] def plot_count_task(input_count: File) -> File: [redun] input_count = File(path=data/KLF4.count.tsv, hash=be612cb0) [redun] File "/Users/falexwolf/repos/bioinformatics-pipeline-tutorial/bioinformatics_pipeline_tutorial/lib.py", line 278, in plot_count_task [redun] counts_plot = plot_counts(output_path, counts) [redun] File "/Users/falexwolf/repos/bioinformatics-pipeline-tutorial/bioinformatics_pipeline_tutorial/lib.py", line 108, in plot_counts [redun] fig.write_image(tmp_file.path) [redun] File "/opt/anaconda3/envs/base1/lib/python3.9/site-packages/plotly/basedatatypes.py", line 3829, in write_image [redun] return pio.write_image(self, *args, **kwargs) [redun] File "/opt/anaconda3/envs/base1/lib/python3.9/site-packages/plotly/io/_kaleido.py", line 267, in write_image [redun] img_data = to_image( [redun] File "/opt/anaconda3/envs/base1/lib/python3.9/site-packages/plotly/io/_kaleido.py", line 144, in to_image [redun] img_bytes = scope.transform( [redun] File "/opt/anaconda3/envs/base1/lib/python3.9/site-packages/kaleido/scopes/plotly.py", line 153, in transform [redun] response = self._perform_transform( [redun] File "/opt/anaconda3/envs/base1/lib/python3.9/site-packages/kaleido/scopes/base.py", line 296, in _perform_transform [redun] export_spec = self._json_dumps(dict(kwargs, data=data)).encode('utf-8') [redun] File "/opt/anaconda3/envs/base1/lib/python3.9/site-packages/kaleido/scopes/plotly.py", line 76, in _json_dumps [redun] return pio.to_json(val, validate=False, remove_uids=False) [redun] File "/opt/anaconda3/envs/base1/lib/python3.9/site-packages/plotly/io/_json.py", line 199, in to_json [redun] return to_json_plotly(fig_dict, pretty=pretty, engine=engine) [redun] File "/opt/anaconda3/envs/base1/lib/python3.9/site-packages/plotly/io/_json.py", line 126, in to_json_plotly [redun] opts = orjson.OPT_NON_STR_KEYS | orjson.OPT_SERIALIZE_NUMPY [redun] AttributeError: partially initialized module 'orjson' has no attribute 'OPT_NON_STR_KEYS' (most likely due to a circular import) ```

I tried downgrading plotly from 5.11 to 5.8, and it didn't help.

I then downgraded orjson from 3.8.1 to 3.5.4, and it fixed the issue.

If other people want to try out your code, to keep your example running, I'd suggest to pin the plotly and orjson versions for the time being - even if it's not great to pin such fundamental dependencies it's much better than an erroring workflow. ☺️

[ricomnl]

Fixed in https://github.com/ricomnl/bioinformatics-pipeline-tutorial/commit/55409956e5d1e92df62b1d330c4009557b6005ea

[falexwolf]

Great!