Commit 220195ef authored by Hyun Soo Kim's avatar Hyun Soo Kim
Browse files

rename chemcial_formula to molecular_formula

parent bd8408d6

testcases/download_molecules.py

+19 −15
Original line number Diff line number Diff line
@@ -6,9 +6,11 @@ import time 
MIN_ISOMER_COUNT = 100



MAX_CID_SEGMENT_LEN = 4000  # maybe 3997

MAX_MOLECULE_COUNT = 1

MAX_MOLECULE_COUNT = 10_000_000



C_count_max = 10000

TITIN = "C169719H270466N45688O52238S911"



C_count_max = 100

C_count_min = C_count_max // 2



H_C_ratio_max = 48 / 25
@@ -47,18 +49,20 @@ if __name__ == "__main__": 
    molecule_count = 0

    while molecule_count < MAX_MOLECULE_COUNT:



        chemical_formula = ""

        while chemical_formula in previous_formulae:

            chemical_formula = ""

        molecular_formula = ""

        while molecular_formula in previous_formulae:

            molecular_formula = ""

            C_count = random.randint(C_count_min, C_count_max)

            chemical_formula += "C" + str(C_count)

            chemical_formula += add_atom("H", C_count, H_C_ratio_max, H_C_ratio_min)

            chemical_formula += add_atom("N", C_count, N_C_ratio_max, N_C_ratio_min)

            chemical_formula += add_atom("O", C_count, O_C_ratio_max, O_C_ratio_min)

            molecular_formula += "C" + str(C_count)

            molecular_formula += add_atom("H", C_count, H_C_ratio_max, H_C_ratio_min)

            molecular_formula += add_atom("N", C_count, N_C_ratio_max, N_C_ratio_min)

            molecular_formula += add_atom("O", C_count, O_C_ratio_max, O_C_ratio_min)



        print(molecular_formula)



        url = (

            "https://pubchem.ncbi.nlm.nih.gov/rest/pug/compound/fastformula/"

            + chemical_formula

            + molecular_formula

            + "/cids/JSON"

        )

        previous_time = wait_for_api_ready(previous_time)
@@ -69,15 +73,15 @@ if __name__ == "__main__": 


        data = response.json()

        if "Fault" in data:

            print(chemical_formula + ": " + data["Fault"]["Message"])

            previous_formulae.append(chemical_formula)

            print(molecular_formula + ": " + data["Fault"]["Message"])

            previous_formulae.append(molecular_formula)

            continue



        cids = data["IdentifierList"]["CID"]

        isomer_count = len(cids)

        if isomer_count < MIN_ISOMER_COUNT:

            # print(chemical_formula + ": Has only " + str(isomer_count) + " isomers")

            previous_formulae.append(chemical_formula)

            # print(molecular_formula + ": Has only " + str(isomer_count) + " isomers")

            previous_formulae.append(molecular_formula)

            continue



        cid_segment = str(cids[0])
@@ -101,7 +105,7 @@ if __name__ == "__main__": 
                for molecule_struct in data["PropertyTable"]["Properties"]:

                    if "CanonicalSMILES" in molecule_struct:

                        molecule_name = (

                            chemical_formula + "_" + str(molecule_struct["CID"])

                            molecular_formula + "_" + str(molecule_struct["CID"])

                        )

                        smiles = molecule_struct["CanonicalSMILES"]

                        if (