Reset the Elecrow's GT911 the way Elecrow's own V3.0 example does (0.2.96, firmware 0.2.82)

The 2 ms reset of ESPHome's expander support (0.2.81) left the chip silent: "Calibration error" with the PCA9557
answering at 0x18. A component of smart_display's (elecrow_touch_reset.h, setup priority 700: after the I2C bus,
before the GT911 driver) writes the expander as Elecrow's V3.0 ESPHome example does: every pin an output and low,
20 ms, the reset released, 100 ms, the interrupt line back to an input. The driver's reset and interrupt pins are gone.

packages/boards/elecrow-crowpanel7.yaml

@@ -383,6 +383,10 @@ psram:
 # CLOSE_CARDS_HOOK, MAY_OPEN_EXTRA, TICK_HOOK, AWAY_EXTRA and ALERT_SHOW_HOOK emptied) it is 1,754 KB, 86.4 %.
 smart_display:
   partitions: 4mb_wide_slots
+  # The GT911 out of reset through the V3.0's PCA9557 before the driver runs (elecrow_touch_reset.h, firmware 0.2.82).
+  elecrow_touch_reset:
+    i2c_id: touch_bus
+    address: 0x18
 
 # The icon fonts at 2 bits a pixel (four levels of edge smoothing instead of sixteen) where the text keeps its four:
 # the three sizes of the 314 tile icons are 312 KB at 4 bits, half that at 2. The text fonts are the sharper for it.
@@ -487,29 +491,21 @@ text_sensor:
     update_interval: 60s
     lambda: 'return {"elecrow"};'
 
-# Pinout of the CrowPanel 7.0" (DIS08070H, V2.0 and V3.0): Elecrow's wiki and the ESPHome devices page. The V3.0 has a
-# PCA9557 I/O expander at I2C 0x18 on the GT911's reset (IO0) and interrupt (IO1) lines, and the GT911 stays in reset
-# until something drives them (firmware 0.2.81: a unit whose touch never answered, "Calibration error" at setup). The
-# expander gets the same register map as the PCA9554 ESPHome knows, and the GT911 driver does Elecrow's own reset
-# through it: reset low with the interrupt line low (that picks address 0x5D), reset high, 56 ms. On a V2.0 without
-# the expander its setup fails, the GT911 boots by its pull-up as before, and the driver still finds it (it tries 0x14
-# by itself when 0x5D does not answer).
+# Pinout of the CrowPanel 7.0" (DIS08070H, V2.0 and V3.0): Elecrow's wiki and their V3.0 ESPHome example. The V3.0 has
+# a PCA9557 I/O expander at I2C 0x18 on the GT911's reset (IO0) and interrupt (IO1) lines, and the GT911 stays in
+# reset until something drives them: on a unit tested here the driver reported "Calibration error" at setup
+# (firmware 0.2.79 to 0.2.81, the last one through ESPHome's own expander support, whose 2 ms reset the chip ignored).
+# elecrow_touch_reset (below) runs Elecrow's own sequence before the driver looks for the chip; a V2.0 without the
+# expander boots its GT911 by the pull-up as before, and the driver tries 0x14 by itself when 0x5D does not answer.
 i2c:
   id: touch_bus
   sda: GPIO19
   scl: GPIO20
   frequency: 400kHz
-  # The scan logs what answers on the bus (0x18 the expander, 0x5D or 0x14 the GT911), with `logger: level: DEBUG` in
-  # a screen's Override YAML: which revision a screen is, and whether the bus is alive at all. It runs before the
-  # GT911's reset, so it upsets nothing.
+  # The scan logs what answers on the bus with `logger: level: DEBUG` in a screen's Override YAML: 0x18 is the
+  # expander; the GT911 is still in reset at that moment and shows up in the driver's own line instead.
   scan: true
 
-pca9554:
-  - id: touch_expander
-    i2c_id: touch_bus
-    address: 0x18
-    pin_count: 8
-
 output:
   - platform: ledc
     pin: GPIO2
@@ -533,12 +529,6 @@ touchscreen:
   i2c_id: touch_bus
   # The reset through the expander picks 0x5D; the driver tries 0x14 by itself when that does not answer.
   address: 0x5D
-  reset_pin:
-    pca9554: touch_expander
-    number: 0
-  interrupt_pin:
-    pca9554: touch_expander
-    number: 1
   display: my_display
   update_interval: 20ms
   transform: