Outreachy Code Project: Abdelrahman Nawar about ersilia HOT 20 CLOSED

I'm Interested in Ersilia as I believe in the product and what it delivers, I think I will be learning a lot if I got accepted,
I've applied to this project specially because I think it fits my skills and the benefit are what I need to learn for the career that I want, I've been using Python for the past three years, I freelanced for a couple of months doing Python Automation and Scripting, I've also built multiple projects, in backend, Data science, Desktop automations, etc.

from ersilia.

My first PR here #180

from ersilia.

Here is my suggestion : #211
I'm currently studying solutions for it

from ersilia.

Screenshot of running a model

from ersilia.

Ho @Nawarrr please could you add your name in the Issue title? where it says

from ersilia.

fetched 3 models

from ersilia.

Perdiction using as a Python library

from ersilia.

simple Streamlit app with a couple of models

from ersilia.

Hi @Nawarrr

Thanks! I am tagging @miquelduranfrigola to check on this but looks good! As I suggested to some other applicants, displaying the molecule as an image when it is inputed would be cool

from ersilia.

Hi @Nawarrr

Thanks! I am tagging @miquelduranfrigola to check on this but looks good! As I suggested to some other applicants, displaying the molecule as an image when it is inputed would be cool

I'm working on it using rdkit but I'm a little stuck, I will get back to you if I couldn't do it

from ersilia.

Hi @Nawarrr

Thanks! I am tagging @miquelduranfrigola to check on this but looks good! As I suggested to some other applicants, displaying the molecule as an image when it is inputed would be cool

from ersilia.

thanks @Nawarrr this looks cool with the molecule drawn!

from ersilia.

Indeed this is quite good. Thanks @Nawarrr

from ersilia.

@GemmaTuron @miquelduranfrigola,
Here is the docstring for ErsiliaModel class, but I have multiple inquires

class ErsiliaModel(ErsiliaBase):
    """The ErsiliaModel object works as the main object for 
    ersilia. It takes the identifier of the model as an attribute,
    then it gets used as an interface for fetching, serving and using the model.

    Identifier is used in fetching model information, the avalibilty
    of the model locally will be checked before fetching. The identifier
    and the slug can be used interchangebly to fetch the model information.

    Typical usage example:
    
        from ersilia import ErsiliaModel
        mdl = ErsiliaModel(__identifier__) 
    
    Attributes:
        model: A str of the identifier or the slug of the model.
        save_to_lake (optional): A boolean Indicating whether the calculations
            will be stored to the lake.
        service_class (optional): A str Indicating the service class.
        config_json (optional): A str of the path to a configuration file.
        credentials_json (optional): A str of the path to a credentials file.
        verbose (optional): A boolean to control enabling verbose output.
        fetch_of_not_avalible (optional): A boolean to enable fetching the model
            if it's not avalible locally.
    """

1- The 2nd paragraph is more of about what the class does with the main argument, not the class itself, should that be removed ?
2- The way I've mentioned the type of the Argument is the same as the google style guide, but in ErsiliaBase init's docstring, the notation used is the parentheses() as I've used to mention that the attribute is optional
I've used the parentheses to mention the optionality of the attribute, should I add the type with it instead of what I'm doing ?

from ersilia.

Hi @Nawarrr I think your CLI feature suggestion is a really good idea

from ersilia.

Hi @Nawarrr I think your CLI feature suggestion is a really good idea

Thanks, I really appreciate you helping me in that feature

from ersilia.

Natural Product Fingerprint Model Card

Card as md with References here,
this is just a screenshot to the model card itself without the references.

from ersilia.

@miquelduranfrigola @GemmaTuron
here is my idea to improve ersilia , I also think that automated tests to check PR would accelerate the contribution process ,
Other PRs I've worked on:
#180

from ersilia.

I'm assigned to #211, but I do intend to work on this after the final application, I won't mark the "All PR or issues assigned to me are complete"

from ersilia.

This is posted after the deadline(Wednesday 20 April) Ignore it if the deadline was final

Model 1 :
Machine learning to predict metabolic drug interactions related to cytochrome P450 isozymes
https://jcheminf.biomedcentral.com/articles/10.1186/s13321-022-00602-x
Model 2 : Sequence-based prediction of protein binding regions and drug–target interactions
https://jcheminf.biomedcentral.com/articles/10.1186/s13321-022-00584-w

Model 3 : Chemical toxicity prediction based on semi-supervised learning and graph convolutional neural network
https://jcheminf.biomedcentral.com/articles/10.1186/s13321-021-00570-8

from ersilia.