How to use the ms2pip.cython_modules.ms2pip_pyx.get_predictions function in ms2pip

dtargets[i].extend(t)
                            else:
                                dtargets[i].extend(t[::-1])
                        else:
                            if i % 2 == 0:
                                dtargets[i] = [t]
                            else:
                                dtargets[i] = [t[::-1]]
                elif tableau:
                    numby = 0
                    numall = 0
                    explainedby = 0
                    explainedall = 0
                    ts = []
                    ps = []
                    predictions = ms2pip_pyx.get_predictions(
                        peptide, modpeptide, charge, model_id, peaks_version, colen
                    )
                    for m, p in zip(msms, peaks):
                        ft.write("%s;%f;%f;;;0\n" % (title, m, 2 ** p))
                    # get targets
                    mzs, targets = ms2pip_pyx.get_targets_all(
                        modpeptide, msms, peaks, float(fragerror), "all"
                    )
                    # get mean by intensity values to normalize!; WRONG !!!
                    maxt = 0.0
                    maxp = 0.0
                    it = 0
                    for cion in [1, 2]:
                        for ionnumber in range(len(modpeptide) - 3):
                            for lion in ["a", "b-h2o", "b-nh3", "b", "c"]:
                                if (lion == "b") & (cion == 1):

ch = charges[pepid]
        charge_buf.append(ch)

        model_id = MODELS[model]["id"]
        peaks_version = MODELS[model]["peaks_version"]

        # get ion mzs
        mzs = ms2pip_pyx.get_mzs(modpeptide, peaks_version)
        mz_buf.append([np.array(m, dtype=np.float32) for m in mzs])

        # Predict the b- and y-ion intensities from the peptide
        # For C-term ion types (y, y++, z), flip the order of predictions,
        # because get_predictions follows order from vector file
        # enumerate works for variable number (and all) ion types
        predictions = ms2pip_pyx.get_predictions(
            peptide, modpeptide, ch, model_id, peaks_version, colen
        )  # SD: added colen
        prediction_buf.append([np.array(p, dtype=np.float32) for p in predictions])

        pcount += 1
        if (pcount % 500) == 0:
            sys.stdout.write("(%i)%i " % (worker_num, pcount))
            sys.stdout.flush()

    return mz_buf, prediction_buf, peplen_buf, charge_buf, pepid_buf