deepspeech 0.9.3 json to subrip .srt for googlecolab

json_to_srt.py

import json
import datetime
from google.colab import files
import copy
import sys
#sys.setrecursionlimit(30000)

uploaded = files.upload()

#upfilename = 'json.txt'


for fn in uploaded.keys():
    print('User uploaded file "{name}" with length {length} bytes'.format(name=fn, length=len(uploaded[fn])))
    upfilename = fn

def fmttime(seconds):
    secs = seconds #millisecs / 1000.0
    d = datetime.timedelta(seconds=secs)
    t = (datetime.datetime.min + d).time()
    milli = t.strftime('%f')[:3]
    value = t.strftime('%H:%M:%S,') + milli
    return value 

original_stdout = sys.stdout #""" stdout backup """
filename = 'subtitle.srt' #""" print subtitle text to this file """
with open(upfilename, 'r') as up_f:
    line = up_f.read()
    jso = json.loads(line)
    ###print(jso['transcripts'][0]['words'])
    with open(filename,'w',encoding='utf8') as down_f:

        sys.stdout = down_f #""" stdout to file """"

        totaltime = 0
        sentence = []

        endtime = ''
        starttime = ''
        lastword_time = 0
        lineNum = 1    


        def list_copy(n):
            temp = []
            for i in range(n):
                c = copy.deepcopy(jso['transcripts'][i]['words'])
                temp.append(c)
            print_word(temp,0)
        
        #""" check confidence
        def print_word(copy,n) :
            if (len(copy[n])) > 0:
                #print(len(copy[n]))  
                dic = copy[n].pop(0)
               ##for ob in jso['transcripts'][n]['words']:
                key = [v for v in dic.values()]
                print("confidence:",str(n)+':', key[0])
                #print("confidence:",str(n)+':', dic)
                n +=1
                if n > len(copy)-1:
                    n = 0
                
               ## for ob in dic.values():
               ##     print('confidence;',str(n)+':',ob)
               ##     n += 1
               ##     if n > len(copy)-1:
               ##         n = 0
               ##     break
                      
                print_word(copy,n)
            else:
                for k in range(len(copy)):
                    if k > len(copy) - 1:
                        return
                        break
                    else:
                        n += 1
                        if n > len(copy) - 1:
                            n = 0
                        if len(copy[n]) < 1:
                            continue
                        else:
                            print_word(copy,n)
                
               ## n += 1
               ## if n > len(copy)-1:
               ##     n = 0
               ## if len(copy[n]) < 1:
               ##     n += 1
               ##     if n > len(copy)-1:
               ##         n = 0
               ##     if len(copy[n]) < 1:
               ##         return
               ## else:
               ##     print_word(copy,n)  
                            
        #list_copy(3)
        
        #confidence:0
        confidence =jso['transcripts'][0]
        #print(confidence)
        for i,ob in enumerate(confidence['words']):
            #print(i,ob) 
            talk_start = True
            talk_end = False

            for key in ob:
                if key == 'word':
                    ###print(jso['transcripts'][0]['words'][i][key])
                    if ob[key] != '':
                        sentence.append(ob.get(key))
                    ###print(*sentence)

                elif key == 'start_time':
                    ###print(jso['transcripts'][0]['words'][i][key])
                    time = ob[key]
                    if  time - lastword_time < 1:
                        talk_start = False
                        talk_end = False
       
                    elif time - lastword_time >= 1: # 1 secons silence
                        talk_start = False
                        talk_end = True
                        ### block >
                        totaltime = 0
                        endtime = fmttime(lastword_time)
                        if len(sentence) > 1:
                            temp = sentence.pop()
                            print(lineNum)
                            lineNum += 1
                            print(starttime,'-->',endtime2)
                            # this word goes to next caption
                            kotoba = ''
                            for word in sentence:
                                kotoba += word + ' '
                            print(kotoba.rstrip())
                            print()
                            sentence.clear()
                            sentence.append(temp) # new caption
                        ### <  block 

                    if len(sentence) == 1 :
                        talk_start = True
                        talk_end = False
                        starttime = fmttime(time)
                        p_time = time

                elif key == 'duration':
                    ###print(jso['transcripts'][0]['words'][i][key])
                    totaltime += ob[key] 
                    lastword_time = p_time + totaltime

                    endtime2 = fmttime(lastword_time)

                    #print('in :',fmttime(time),'>>',*sentence)
                    #print('end :',fmttime(time+totaltime)) 
                    #print('bt :',fmttime(totaltime))

                    if totaltime >= 4: # 4 seconds speech gose to 1 caption
                        ### block >
                        totaltime = 0    
                        endtime = fmttime(lastword_time)
                        print(lineNum)
                        lineNum += 1
                        print(starttime,'-->',endtime)
                        kotoba = ''
                        for word in sentence:
                            kotoba += word + ' '
                        
                        print(kotoba.rstrip())
                        print()
                        sentence.clear()
                        ### < block
                    elif totaltime < 4 and i + 1 == len(confidence['words']): 
                        ### block >
                        totaltime = 0    
                        endtime = fmttime(lastword_time)
                        print(lineNum)
                        lineNum += 1
                        print(starttime,'-->',endtime)
                        kotoba = ''
                        for word in sentence:
                            kotoba += word + ' '
                        
                        print(kotoba.rstrip())
                        print()
                        sentence.clear()
                        ### < block 
        sys.stdout = original_stdout # stdout back 

files.download(filename) # download .srt file 

https://discourse.mozilla.org/t/do-deepspeech-have-subtitle-srt-output-mode-how-can-i-merge-words-into-the-proper-sentences/68749/12?u=haywhnk