search

MindAI / MiPepid

14 stars 5 forks source link

Incompatible with Python 3.5 and beyond? #2

Open mattbcvs opened 3 years ago

mattbcvs commented 3 years ago

Hi there,

Running MiPepid in Python 3.5.2 with Bcbio 1.2.0, currently trying to recapitulate the results obtained on a sample of the Ensembl lncRNAs as follows:

python3 src/mipepid.py MiPepid_prediction_on_lncRNA_sample.fa smORFs_MiPepid_GENCODE.csv

But get massively different coding scores to "MiPepid_predictions_on_uORF_and_lncRNA" as well as the following message:

/exports/igmm/software/pkg/el7/apps/python/3.5.2/lib/python3.5/site-packages/sklearn/base.py:306: UserWarning: Trying to unpickle estimator LogisticRegression from version 0.19.1 when using version 0.21.3. This might lead to breaking code or invalid results. Use at your own risk. UserWarning)

I think LogisticRegression 0.19.1 corresponds to Python 3.4 but the correct version of Bcbio is less clear and not accessible to me. Is there an update required to enable MiPepid to still be an accessible tool to researchers given recent updates to Python and dependent modules?

Thanks in advance, Matt

akramayman commented 11 months ago

Hello @mattbcvs, I have a problem when trying to run the code on the sample_seqs.fasta file, I got an error like this :

logr = pickle.load(f) ModuleNotFoundError: No module named 'sklearn.linear_model.logistic'

despite I have these versions: Python -> 3.8.18 Pandas -> 2.0.3 Numpy -> 1.24.3 Scikit-learn -> 1.3.2

If you can help me to solve this problem, I think the problem in model.pkl file. Thank you. Akram

JC-therea commented 10 months ago

Hi, I was able to solve it but I had to use the previous version of many packages:

python 3.8.18 scikit-learn 0.22.1 numpy 1.19.0 pandas 1.4.1

I hope it helps

akramayman commented 9 months ago

@JC-therea Thank you