Game Events Guide

AI Summary

Dota 2 replays contain 364 Source 1 legacy game events with typed fields. Use parse_game_events() to capture events like player kills, rune pickups, tower destructions, and combat log entries. Filter by event name substring or specific names. Events have tick timestamps and typed field dictionaries. Common events: dota_player_kill, dota_combatlog, dota_rune_activated, dota_tower_kill, dota_roshan_kill. Use capture_types=True to get the full event type list.

---

Overview

Dota 2 replays contain Source 1 legacy game events - structured notifications about game occurrences. There are 364 different event types, each with specific typed fields.

from python_manta import MantaParser

parser = MantaParser()
result = parser.parse_game_events("match.dem", max_events=100)

for event in result.events:
    print(f"[{event.tick}] {event.name}: {event.fields}")

---

Discovering Event Types

To see all available event types:

result = parser.parse_game_events("match.dem", capture_types=True, max_events=0)

print(f"Total event types: {len(result.event_types)}")
for event_type in sorted(result.event_types):
    print(f"  {event_type}")

This returns 364 event type names like: - dota_player_kill - dota_combatlog - dota_rune_activated - dota_tower_kill - dota_roshan_kill - And many more...

---

Filtering Events

By Name Substring

# All events containing "kill"
kills = parser.parse_game_events("match.dem", event_filter="kill", max_events=100)

# All combatlog events
combatlog = parser.parse_game_events("match.dem", event_filter="combatlog", max_events=500)

# All rune events
runes = parser.parse_game_events("match.dem", event_filter="rune", max_events=50)

By Specific Event Names

# Only player kills and tower kills
result = parser.parse_game_events(
    "match.dem",
    event_names=["dota_player_kill", "dota_tower_kill"],
    max_events=100
)

---

Common Event Types

Player Kills

result = parser.parse_game_events("match.dem", event_filter="dota_player_kill", max_events=100)

for event in result.events:
    victim = event.fields.get("victim_userid")
    killer = event.fields.get("killer1_userid")
    print(f"[Tick {event.tick}] Player {killer} killed player {victim}")

Tower Destruction

result = parser.parse_game_events("match.dem", event_filter="dota_tower_kill", max_events=50)

for event in result.events:
    killer = event.fields.get("killer_userid")
    tower = event.fields.get("teamnumber")
    gold = event.fields.get("gold")
    print(f"[Tick {event.tick}] Tower destroyed by player {killer}, gold: {gold}")

Roshan Kills

result = parser.parse_game_events("match.dem", event_filter="dota_roshan_kill", max_events=10)

for event in result.events:
    team = event.fields.get("teamnumber")
    print(f"[Tick {event.tick}] Roshan killed by team {team}")

Rune Activations

result = parser.parse_game_events("match.dem", event_filter="dota_rune_activated", max_events=50)

for event in result.events:
    player = event.fields.get("player_id")
    rune_type = event.fields.get("rune")
    print(f"[Tick {event.tick}] Player {player} activated rune type {rune_type}")

Combat Log Events

result = parser.parse_game_events("match.dem", event_filter="dota_combatlog", max_events=200)

for event in result.events:
    log_type = event.fields.get("type")
    target = event.fields.get("targetname")
    source = event.fields.get("sourcename")
    value = event.fields.get("value")
    print(f"[Tick {event.tick}] Type {log_type}: {source} -> {target} ({value})")

---

Event Field Types

Event fields are automatically typed based on their definition:

Field Type	Python Type
string	str
float	float
long	int
short	int
byte	int
bool	bool
uint64	int

---

Full Event Type List

AI Summary

All 364 Event Types (click to expand)

Common Dota 2 events include:

Player Events: - dota_player_kill - dota_player_deny - dota_player_pick_hero - dota_player_learned_ability - dota_player_used_ability - dota_player_gained_level - dota_player_take_tower_damage

Unit Events: - dota_tower_kill - dota_barracks_kill - dota_roshan_kill - dota_courier_lost - dota_npc_goal_reached

Item Events: - dota_item_purchased - dota_item_combined - dota_item_picked_up

Rune Events: - dota_rune_activated - dota_rune_spawned - dota_bounty_rune_pickup

Game State: - dota_match_done - dota_game_state_change - dota_team_kill_credit - game_rules_state_change

Combat: - dota_combatlog - dota_player_take_tower_damage

Other: - dota_chase_hero - dota_tutorial_* (various tutorial events)

Use capture_types=True to get the complete list for your replay.

---

Example: Kill Timeline

from python_manta import MantaParser

parser = MantaParser()

# Get all player kills
result = parser.parse_game_events("match.dem", event_filter="dota_player_kill", max_events=500)

print("Kill Timeline:")
print("-" * 50)

for event in result.events:
    tick = event.tick
    killer = event.fields.get("killer1_userid", "Unknown")
    victim = event.fields.get("victim_userid", "Unknown")
    assisters = []

    # Check for assists
    for i in range(2, 6):
        assister = event.fields.get(f"killer{i}_userid")
        if assister:
            assisters.append(str(assister))

    assist_str = f" + {', '.join(assisters)}" if assisters else ""
    print(f"[Tick {tick:6d}] Player {killer}{assist_str} killed Player {victim}")

---

Game Events vs Combat Log

Both provide combat information but differ:

Aspect	Game Events	Combat Log
API	parse_game_events()	parse_combat_log()
Structure	Named events with typed fields	Structured entries with fixed schema
Types	364 event types	12 log types
Best for	High-level events (kills, runes)	Detailed damage/heal analysis
Timing	Throughout match	After ~12-17 minutes

Use game events for discrete occurrences (kills, objectives) and combat log for continuous combat data (DPS, healing).

---

Performance Tips

1. Always set max_events - Events can be numerous
2. Use specific filters - Narrow down with event_filter or event_names
3. Discover first - Use capture_types=True to find relevant events
4. Combine with other APIs - Use entity queries for context about players