firmware: add I2S MEMS microphone driver; fix missing essence migration

Mic: confirmed via its pinout (L/R, WS, SCK, SD, VCC, GND) that the third
target module is a standard I2S digital MEMS mic (INMP441-family). Added
mems_mic.c/h using ESP-IDF's current driver/i2s_std.h API — reports RMS
audio level in dBFS as sensor_type "evp" rather than attempting on-device
voice-band FFT (the browser EVP mode's approach); the backend's existing
statistical anomaly detector handles spike detection from the raw level,
same as it already does for temperature/pressure/presence.

Also fixes a real gap Workstream B's report flagged: User.essence (a live
model column used throughout merged code — /auth/me, inventory purchases,
summon trickle) had no migration line in main.py's lifespan, which would
have broken on the actual production Postgres database.

firmware/esp32p4-sensor-node/main/mems_mic.h

@@ -0,0 +1,75 @@
+// I2S digital MEMS microphone driver (INMP441-family — L/R, WS, SCK, SD
+// pinout, confirmed against the actual module in use).
+//
+// Unlike the browser-based EVP mode's client-side voice-band FFT
+// (frontend/src/lib/evp.ts), this driver does NOT attempt on-device
+// spectral analysis — it samples a short audio block per reporting cycle
+// and reports its RMS energy level in dBFS as a plain numeric reading.
+// The backend's existing statistical anomaly detector (the same one that
+// already handles temperature/pressure) does the spike detection — no
+// need to duplicate baseline-tracking logic on the device. This is
+// simpler and more honest than pretending to replicate real voice-band
+// filtering without ever having tested it.
+//
+// Implements the sensor_driver_t interface (see sensor_driver.h).
+//
+// Wiring (module pin names: L/R, WS, SCK, SD, VCC, GND):
+//   Mic VCC -> 3V3
+//   Mic GND -> GND
+//   Mic L/R -> GND (selects left-channel output; tie to 3V3 instead for
+//              right-channel — either is fine, this driver reads whichever
+//              channel the mic is configured to output)
+//   Mic WS  -> ESP32 GPIO (MEMS_MIC_I2S_WS_GPIO)   -- word select / LRCLK
+//   Mic SCK -> ESP32 GPIO (MEMS_MIC_I2S_BCLK_GPIO) -- bit clock
+//   Mic SD  -> ESP32 GPIO (MEMS_MIC_I2S_DIN_GPIO)  -- serial data OUT of
+//              the mic, IN to the ESP32
+// GPIO picks avoid this board's reserved SDIO range (14-19, 54), the
+// BMP280's I2C pins (8/9), and the RD-03E's UART pins (4/5) -- see
+// README.md's pinout table for the full picture.
+
+#pragma once
+
+#include "esp_err.h"
+#include "sensor_driver.h"
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#ifndef MEMS_MIC_I2S_PORT
+#define MEMS_MIC_I2S_PORT      0
+#endif
+
+#ifndef MEMS_MIC_I2S_BCLK_GPIO
+#define MEMS_MIC_I2S_BCLK_GPIO 10
+#endif
+
+#ifndef MEMS_MIC_I2S_WS_GPIO
+#define MEMS_MIC_I2S_WS_GPIO   11
+#endif
+
+#ifndef MEMS_MIC_I2S_DIN_GPIO
+#define MEMS_MIC_I2S_DIN_GPIO  12
+#endif
+
+// 16kHz is standard for voice-band work (Nyquist covers the ~300Hz-3.4kHz
+// band the browser-based EVP mode watches) without the data-rate/CPU cost
+// of a higher rate this simple RMS-only driver doesn't need.
+#ifndef MEMS_MIC_SAMPLE_RATE_HZ
+#define MEMS_MIC_SAMPLE_RATE_HZ 16000
+#endif
+
+// One reporting cycle's sample block: 256ms at 16kHz. Long enough to
+// average out single-sample noise, short enough to stay responsive to a
+// genuine brief EVP-style spike.
+#ifndef MEMS_MIC_SAMPLES_PER_READ
+#define MEMS_MIC_SAMPLES_PER_READ 4096
+#endif
+
+// sensor_driver_t-compatible entry points.
+esp_err_t mems_mic_init(void);
+esp_err_t mems_mic_read(sensor_reading_t *out, size_t max_out, size_t *out_count);
+
+#ifdef __cplusplus
+}
+#endif