Textgrid to .lab, take 2

from praatio import textgrid

from pathlib import Path

def get_combined_words_and_phones(filename):
    from praatio import textgrid

    tg = textgrid.openTextgrid(filename, False)
    if not tg.tierNameList:
        return []
    if tg.tierNameList == ['words', 'phones']:
        word_tier = 'words'
    elif tg.tierNameList == ['Word', 'phones']:
        word_tier = 'Word'

    word = tg.tierDict[word_tier]
    phones = tg.tierDict['phones']
    
    i = 0
    j = 0
    out = []

    def it_to_dict(it):
        ret = {}
        ret['start'] = it.start
        ret['end'] = it.end
        ret['label'] = it.label
        return ret

    while i < len(word.entryList) and j < len(phones.entryList):
        cur_word = it_to_dict(word.entryList[i])
        cur_word['phones'] = []
        while j < len(phones.entryList) and phones.entryList[j].end <= cur_word['end']:
            end_time = phones.entryList[j].end
            tmp_phone = it_to_dict(phones.entryList[j])
            cur_word['phones'].append(tmp_phone)
            j += 1
            if end_time == cur_word['end']:
                i += 1
                out.append(cur_word)
                continue
    return out

This phone merging is only intended to merge a silence or spoken noise 'phone' to the left, but for the most part this doesn't do what I'd wanted, as it often means a silence 'word' has been inserted.

def merge_phones(word):
    outphones = []
    if len(word['phones']) == 1:
        return word['phones']
    for i in range(0, len(word['phones'])):
        if i > 0 and word['phones'][i]['label'] in ["", "sil", "spn"]:
            outphones[-1]['end'] = word['phones'][i]['end']
        else:
            outphones.append(word['phones'][i])
    return outphones

def tg_to_lab(filename, target="phones"):
    combined = get_combined_words_and_phones(filename)
    merged = [merge_phones(x) for x in combined]
    flattened = [item for sublist in merged for item in sublist]
    
    out = []

    for tmp_phone in flattened:
        start = int(tmp_phone['start'] * 10000000)
        end = int(tmp_phone['end'] * 10000000)
        label = tmp_phone['label']
        out.append(f"{start} {end} {label}")

    return out

inpath = Path("/home/jim/Playing/mfa_alignments/snc-out")
outpath = Path("/home/jim/Playing/mfa_alignments/snc-lab")

for filename in inpath.glob("*.TextGrid"):
    out = outpath / f"{filename.stem}.lab"
    lab = tg_to_lab(filename)

    with open(out, "w") as outf:
        for line in lab:
            outf.write(line + "\n")