Merge pull request #795 from hanzi/benchmark-mode

Make it easier to benchmark encounters/hr on a route

Author: hanzi

modules/gui/debug_tabs.py

@@ -1171,6 +1171,7 @@ def _get_data(self):
             "Session Time at 1×": f"{session_time_at_1x}",
             "RNG Seed": hex(unpack_uint32(read_symbol("gRngValue"))),
             "Encounters/h (at 1×)": context.stats.encounter_rate_at_1x,
+            "Encounters in Timing Log": len(context.stats._encounter_timestamps),
             "Controller Stack": [controller.__qualname__ for controller in context.controller_stack],
             "RTC": rtc.isoformat() if rtc is not None else None,
             "Currently Running Actions": debug.action_stack,

modules/stats.py

@@ -1,3 +1,4 @@
+import os
 import sqlite3
 import sys
 import threading
@@ -611,8 +612,19 @@ def __init__(self, profile: "Profile"):
         self._pickup_items: dict[int, PickupItem] = self._get_pickup_items()
         self._base_data: dict[str, str | None] = self._get_base_data()
 
-        self._encounter_timestamps: deque[float] = deque(maxlen=100)
-        self._encounter_frames: deque[int] = deque(maxlen=100)
+        # Normally, the encounter rate is calculated based on the previous 100 encounters. Which
+        # is accurate enough to get a rough idea of how fast things are going, and it means that
+        # a long break in between encounters (due to changing modes, playing manually etc.) will
+        # be flushed out reasonably soon.
+        # But when trying to measure a more accurate encounters/hr for a given route, we need a
+        # larger sample size because the encounter rate can fluctuate quite a bit in some modes.
+        #
+        # So rather than constantly having to edit this file, this allows setting the environment
+        # variable `POKEBOT_ENCOUNTER_BENCHMARK` to anything but an empty string in order to
+        # increase the sample size to 1,000.
+        encounter_buffer_size = 1000 if os.getenv("POKEBOT_ENCOUNTER_BENCHMARK", "") != "" else 100
+        self._encounter_timestamps: deque[float] = deque(maxlen=encounter_buffer_size)
+        self._encounter_frames: deque[int] = deque(maxlen=encounter_buffer_size)
 
     def set_data(self, key: str, value: str | None):
         self._execute_write("REPLACE INTO base_data (data_key, value) VALUES (?, ?)", (key, value))