candidates_to_design.py

#!/usr/local/bin/python3
# Copyright (C) 2020 TeselaGen Biotechnology, Inc.
# License: MIT
import math
import json
import textwrap
from pathlib import Path
from typing import List
import argparse
import sys

import pandas as pd
"""
This module provides a command line tool that generates a json design
that can be imported into TeselaGen's DESIGN module. As an input, this
tool uses a csv data table as the one can be exported from the *evolutions*
algorithm's result, at EVOLVE. Some usage examples can be found below
"""

def build_design_from_candidates(
    candidates_data: List[dict],
    bin_cols: List[str],
    #lab_id: str,
    #user_id: str,
    name: str)->dict:
    ''' Builds design dict from candidates data

    Args:
        candidates_data: Records object from EVOLVE output data csv
        bin_cols: List containing the names of the bins to be read from data
        name: Name of design as will be seen on DESIGN.

    Returns:
        A dict that contains design information. This dict can be serialized as
        json to be used within the DESIGN module.

    '''
    designs = {'design':{}}
    # Use just the elements with valid priority (suggested candidates)
    data = [d for d in candidates_data if not math.isnan(d['Priority'])]
    print(f"Generating design using {len(data)} candidates")
    # Create bins
    designs['design']['bin'] = {f"{i+1}":{'id':f"{i+1}",
                                          'direction': True,
                                          'name': name,
                                          'iconId':"1"} for i,name in enumerate(bin_cols)}
    # Create Bincards
    designs['design']['binCard'] = {f"{i+1}":{'id': f"{i+1}",
                                          'index': i,
                                          'cardId': "1",
                                          'binId': f"{i+1}"} for i, name in enumerate(bin_cols)}
    # Create card
    designs['design']['card'] = {"1": {"id": "1",
                                       "isRoot": True,
                                       "circular": True,
                                       "name": "Target Construct"}}
    # Create design
    designs['design']['design'] = {"1":{"id": "1",
                                        #"userId": user_id,
                                        #"labId": lab_id,
                                        "name": name,
                                        "type": "grand-design",
                                        "layoutType": "list",
                                        "numRows": len(data)}}
    # Create elements
    designs['design']['element'] = {}
    el_i = 0
    for i_bin, bin_name in enumerate(bin_cols):
        for i_el, el in enumerate(data):
            el_i += 1
            designs['design']['element'][f"{el_i}"] = {
            	"id": f"{el_i}",
                "name": el[bin_name],
                "index": i_el,
                "binId": f"{i_bin+1}"}
    # Set icon
    designs['design']['icon'] = {"1": {"id": "1",
                                 "isSbol": True,
                                 "name": "USER-DEFINED",
                                 "path": "M 5 40 L 5 60 L 45 60 L 45 40 Z"}}
    # Maybe this is not needed
    designs["hdeDesignExportVersion"] = 3
    #print("Done!")
    return designs


def candidates_to_design(input_file: str, bin_cols: List[str], name: str, output: str):
    ''' Read candidates file (csv) and generates json file that can be read from DESIGN
    module.

    Args:
        input_file: Path to the input csv file as string.
        bin_cols: A list with the names of the bin columns at the csv file
        name: Name of output design
        output: Path of the output file as string.
    '''
    # Get arguments
    if 'csv' in input_file:
        # Loads data from a csv file downloaded from EVOLVE UI
        data_df = pd.read_csv(input_file)
        data = data_df.to_dict(orient="records")
    # Build dictionary object with design
    design_dict = build_design_from_candidates(
        candidates_data=data,#: List[dict],
        bin_cols= bin_cols,
        name= name)
    # Write design to disc
    #Path('_export/').mkdir(parents=True, exist_ok=True)
    with open(output, 'w') as f:
        json.dump(design_dict, f)
        print(f"Created design at '{output}'")


def args_parser(args_in):
    """Parses arguments from the command line tool"""
    parser = argparse.ArgumentParser(
        prog='candidates_to_design',
        formatter_class=argparse.RawDescriptionHelpFormatter,
        description=textwrap.dedent('''\
            Script for generating a design json from a set of evolution's candidates.

            Example usage:
                Download output data from evoultion result "Teselagen Example Evolutive Model".
                Store this file in this same folder with the name "evolve_exported_results.csv".
                The next line will create a file with the default name "EVOLVE_candidates.json"
                with a design containing valid candidates.

                >>> python3 candidates_to_design.py --bin_cols "Teselagen Enzyme A" "Teselagen Enzyme B" --input "evolve_exported_results.csv"

            '''))
    parser.add_argument('--input',
                        default="input.csv",
                        #metavar='-i',
                        type=str,
                        help='Path to csv input file',
                        )
    parser.add_argument('--bin_cols',
                        #metavar='-i',
                        required=True,
                        type=str,
                        nargs='+',
                        help='List containing the names of the bins to be read from data. These names should be column names at the input',
                        )
    parser.add_argument('--name',
                        default="EVOLVE_candidates",
                        #metavar='-i',
                        type=str,
                        help='Path to csv input file',
                        )
    parser.add_argument('--output',
                        default="EVOLVE_candidates.json",
                        #metavar='-i',
                        type=str,
                        help='Path to json output file',
                        )
    return parser.parse_args(args_in)

# Command line tool, parses arguments and call candidates_to_design
if __name__ == "__main__":
    args = args_parser(sys.argv[1:])
    print(args)
    candidates_to_design(
        input_file=args.input,
        bin_cols=args.bin_cols,
        name=args.name,
        output=args.output
    )
    print("Done!")