Mapping Unqueue to the corresponding DE_QUEUE action

[noreallyimfine]

Hi, I'm working on parsing the match to draw out some summary statistics. I have run into an issue where I can't tell how to associate the queueing action to an unqueueing one, to know the villager wasn't created.

I see this object showing a villager queued up:

{'timestamp': '0:07:54.729000',
  'type': 'DE_QUEUE',
  'payload': {'object_ids': [1933],
   'amount': 1,
   'unit_id': 83,
   'sequence': 474729,
   'unit': 'Villager'},
  'player': 1}

then there's this one showing something unqueued:

{'timestamp': '0:07:58.425000',
  'type': 'SPECIAL',
  'payload': {'order_id': 4,
   'slot_id': 0,
   'target_id': -1,
   'x': 0.0,
   'y': 0.0,
   'object_ids': [494848],
   'sequence': 478425,
   'order': 'Unqueue'},
  'player': 1}

I am as sure as i can be that this unqueue is the unqueueing of the villager queued in the first object. But I can't figure how what value in the unqueue object can be used to identify where/what was unqueued. I suspect it should be the object_id but I don't see anywhere that 494848 is identified as the TC.

Any help greatly appreciated

[denniske]

If you want to be 100% sure you can record your own game against AI and then look at the recording

[noreallyimfine]

It is the unqueuing of that villager, but what ID connects the 2? or is supposed to connect the 2?

I want to use this in a script so am looking to connect events in a reliable automatable way.

[happyleavesaoc]

I suspect the unqueue object_id is not being correctly parsed, at least based on the example you gave.

[noreallyimfine]

oh that is entirely possible.

I am using the parse_match() func to process the .aoe2record file. Then I'm parsing the actions to track villagers created, looking for DE_QUEUE for queued vils and SPECIAL with the order: 'Unqueue' for unqueued.

Am I missing something?

this is my code for reference

`def summarize_match(match_obj):
match_summary = {
'players': {}
}

for player in match_obj['players']:
    match_summary['players'][player['number']] = {
        'name': player['name'],
        'civ': player['civilization'],
        'winner': player['winner'],
        'eapm': player['eapm'],
        'dark_age_vils': 0,
    }

age_research_actions = [e for e in match_obj['actions'] if e['type'] == 'RESEARCH' and e['payload']['technology'] in AGES]
for act in age_research_actions:
        player = act['player']
        civ = match_summary['players'][player]['civ']
        age = act['payload']['technology']
        age_landing_time = calculate_age_landing_time(age, act['timestamp'], civ)
        age_name = (age.split(' ')[0]).lower()
        match_summary['players'][player][f"{age_name}_research_time"] = parse_to_timedelta(act['timestamp'])
        match_summary['players'][player][f"{age_name}_landing_time"] = age_landing_time

for player in match_summary['players']:
    player_feudal_research_time = f"{match_summary['players'][player]['feudal_research_time']}"
    player_civ = match_summary['players'][player]['civ']

    active_tc_time = 0
    for act in match_obj['actions']:
        if act['player'] == player and act['timestamp'] < player_feudal_research_time:
            if act['type'] == 'DE_QUEUE' and act['payload']['unit'] == 'Villager':
                match_summary['players'][player]['dark_age_vils'] += 1
                active_tc_time += get_civ_research_time(player_civ, act['payload']['unit'])

            elif act['type'] == 'RESEARCH' and act['payload']['technology'] == 'Loom':
                active_tc_time += get_civ_research_time(player_civ, act['payload']['technology'])
            
            elif act['type'] == 'SPECIAL' and act['payload'].get('order') == 'Unqueue':
                print(act)
                ...

    match_summary['players'][player]['dark_age_idle_tc_time'] = format_timedelta(abs(parse_to_timedelta(player_feudal_research_time) - timedelta(seconds=active_tc_time)))

pprint(match_summary)
return format_timedeltas(match_summary)`

[happyleavesaoc]

I'm saying there might be a bug in mgz, not that you are doing something wrong.