Hi there,
I'm having a similar problem to the one reported in #8 :
[...]
ESC[92m[Aug 11 06:36 PM] ESC[0mESC[38;20mMemory monitoring data written to: /srv/scratch/ferrari/Metagenomics/xabi/kmr/KMR19/predict/logfiles/predict-abinitio-memory-monitoring.jsonl
ESC[0m
ESC[92m[Aug 11 06:36 PM] ESC[0mESC[33;20mUnable to access JGI taxonomy lookup, reverting to taxonomy from training dataESC[0m
Traceback (most recent call last):
File "/srv/scratch/z3382651/miniconda3/envs/funannotate2/bin/funannotate2", line 10, in <module>
sys.exit(main())
~~~~^^
File "/srv/scratch/z3382651/miniconda3/envs/funannotate2/lib/python3.13/site-packages/funannotate2/__main__.py", line 26, in main
predict(args)
~~~~~~~^^^^^^
File "/srv/scratch/z3382651/miniconda3/envs/funannotate2/lib/python3.13/site-packages/funannotate2/predict.py", line 908, in predict
{"superkingdom": taxonomy.get("superkingdom"), "kingdom": taxonomy.get("kingdom")}
^^^^^^^^^^^^
AttributeError: 'bool' object has no attribute 'get'
My guess is that there is an issue with some taxonomy lookup, basically bc we haven't even published the description paper for the genus. So the whole taxonomy in the training json file is non-existent:
{
"name": "Psychrogloeomyces_antarcticus_KMR19",
"species": "Psychrogloeomyces antarcticus",
"taxonomy": false,
"busco-lineage": "/srv/scratch/z3382651/db/funannotate2/dothideomycetes_odb12.2",
"assembly": {
"n_contigs": 14,
"size": 20786986,
"n50": 1584158,
"n90": 1174648,
"l50": 6,
"l90": 12,
"avg_length": 1484785
},
[...]During training I indicated both the busco lineage and the augustus species for the starting model so I was a bit surprised by this. I managed to make it run by manually adding a taxonomy to the json file, but I imagine this may happen any time there is no public taxonomy available so a setting/option to override this behaviour would be appreciated.
Cheers
Hi there,
I'm having a similar problem to the one reported in #8 :
My guess is that there is an issue with some taxonomy lookup, basically bc we haven't even published the description paper for the genus. So the whole taxonomy in the training json file is non-existent:
{ "name": "Psychrogloeomyces_antarcticus_KMR19", "species": "Psychrogloeomyces antarcticus", "taxonomy": false, "busco-lineage": "/srv/scratch/z3382651/db/funannotate2/dothideomycetes_odb12.2", "assembly": { "n_contigs": 14, "size": 20786986, "n50": 1584158, "n90": 1174648, "l50": 6, "l90": 12, "avg_length": 1484785 }, [...]During training I indicated both the busco lineage and the augustus species for the starting model so I was a bit surprised by this. I managed to make it run by manually adding a taxonomy to the json file, but I imagine this may happen any time there is no public taxonomy available so a setting/option to override this behaviour would be appreciated.
Cheers