Add Arduino USB PID supplement

.github/workflows/update-data.yml

@@ -210,6 +210,15 @@ jobs:
             --out /tmp/microchip-first-party-usb-pids.json
           wc -l /tmp/microchip-first-party-usb-pids.json
 
+      - name: Fetch Arduino USB PID supplement (USB-VID tier-1 supplement)
+        id: fetch-arduino-pids
+        continue-on-error: true
+        run: |
+          uv run --no-project --script \
+            "${{ github.workspace }}/online-data-tools/fetch_arduino_usb_pids.py" \
+            --out /tmp/arduino-usb-pids.json
+          wc -l /tmp/arduino-usb-pids.json
+
       - name: Fetch FTDI USB PID supplement (USB-VID tier-3 supplement)
         id: fetch-ftdi-pids
         continue-on-error: true
@@ -327,6 +336,9 @@ jobs:
           if [ "${{ steps.fetch-microchip-first-party-pids.outcome }}" = "success" ] && [ -s /tmp/microchip-first-party-usb-pids.json ]; then
             args+=(--json "microchip-first-party-usb-pids=/tmp/microchip-first-party-usb-pids.json")
           fi
+          if [ "${{ steps.fetch-arduino-pids.outcome }}" = "success" ] && [ -s /tmp/arduino-usb-pids.json ]; then
+            args+=(--json "arduino-usb-pids=/tmp/arduino-usb-pids.json")
+          fi
           if [ "${{ steps.run-dump.outcome }}" = "success" ] && [ -s /tmp/usb-ids-rs.json ]; then
             args+=(--json "usb-ids-rs=/tmp/usb-ids-rs.json")
           fi
@@ -554,6 +566,7 @@ jobs:
             echo "| raspberrypi usb-pid        | ${{ steps.fetch-raspberrypi-pids.outcome }} |"
             echo "| nordic usb-pids            | ${{ steps.fetch-nordic-pids.outcome }} |"
             echo "| microchip first-party pids | ${{ steps.fetch-microchip-first-party-pids.outcome }} |"
+            echo "| arduino usb-pids           | ${{ steps.fetch-arduino-pids.outcome }} |"
             echo "| ftdi usb-pids              | ${{ steps.fetch-ftdi-pids.outcome }} |"
             echo "| wch usb-pids               | ${{ steps.fetch-wch-pids.outcome }} |"
             echo "| teensy usb-pids            | ${{ steps.fetch-teensy-pids.outcome }} |"

online-data-tools/README.md

@@ -15,6 +15,7 @@ is committed to orphan branches:
 | `fetch_raspberrypi_usb_pids.py` | `raspberrypi/usb-pid` official PID registry | merge-compatible `/tmp/raspberrypi-usb-pids.json` |
 | `fetch_nordic_usb_pids.py` | Nordic nRF Connect Programmer / DFU sources | merge-compatible `/tmp/nordic-usb-pids.json` |
 | `fetch_microchip_usb_pids.py` | Microchip pyedbglib/pykitinfo plus weak AVRDUDE/board-package supplements | merge-compatible `/tmp/microchip-*-usb-pids.json` |
+| `fetch_arduino_usb_pids.py` | Official ArduinoCore `boards.txt` files | merge-compatible `/tmp/arduino-usb-pids.json` |
 | `fetch_ftdi_usb_pids.py` | Linux `ftdi_sio_ids.h` original-FTDI PID section | merge-compatible `/tmp/ftdi-usb-pids.json` |
 | `fetch_wch_usb_pids.py` | WCH CH343 Linux driver + udev rules | merge-compatible `/tmp/wch-usb-pids.json` |
 | `fetch_teensy_usb_pids.py` | PJRC Teensy core headers + loader CLI | merge-compatible `/tmp/teensy-usb-pids.json` |
@@ -67,6 +68,16 @@ first-party data. LowPowerLab's current package maps CurrentRanger to
 local FastLED board JSON currently carries `current_ranger` as `04d8:eee5`,
 so that local mismatch is noted rather than used as source truth.
 
+The Arduino supplement parses official ArduinoCore `boards.txt` files from
+Arduino-maintained AVR, SAM, SAMD, megaAVR, mbed, Renesas, and Zephyr cores.
+Arduino does not publish a single public PID allocation registry; the core
+board definitions are the authoritative public table for Arduino VID `0x2341`
+and historical Arduino VID `0x2a03`. The workflow orders this source before
+generic USB-ID feeds so current Arduino product names win. As with other
+board-package data, a row from an ArduinoCore source improves VID/PID
+resolution but does not prove the board exists under
+`crates/fbuild-config/assets/boards`.
+
 The FTDI supplement parses the upstream Linux `ftdi_sio_ids.h` header but only
 the original FTDI PID section before the third-party marker. It emits a small
 allowlist of missing FTDI-owned bridge/product rows such as `0403:6040` and
@@ -135,6 +146,7 @@ uv run --no-project --with pytest pytest online-data-tools/test_espressif_usb_pi
 uv run --no-project --with pytest pytest online-data-tools/test_raspberrypi_usb_pids.py -v
 uv run --no-project --with pytest pytest online-data-tools/test_nordic_usb_pids.py -v
 uv run --no-project --with pytest pytest online-data-tools/test_microchip_usb_pids.py -v
+uv run --no-project --with pytest pytest online-data-tools/test_arduino_usb_pids.py -v
 uv run --no-project --with pytest pytest online-data-tools/test_ftdi_usb_pids.py -v
 uv run --no-project --with pytest pytest online-data-tools/test_wch_usb_pids.py -v
 uv run --no-project --with pytest pytest online-data-tools/test_teensy_usb_pids.py -v

online-data-tools/fetch_arduino_usb_pids.py

@@ -0,0 +1,197 @@
+#!/usr/bin/env -S uv run --no-project --script
+# /// script
+# requires-python = ">=3.10"
+# ///
+"""Fetch official Arduino USB PID rows into merge_sources.py JSON.
+
+Arduino does not publish a single public PID registry. Its official cores do
+publish Arduino-style `boards.txt` files with `vid.N` / `pid.N` declarations.
+This script parses the direct GitHub `boards.txt` files for the official cores
+that currently declare Arduino VID rows.
+
+Output schema:
+    {
+      "2341:0043": {
+        "vendor": "Arduino SA",
+        "product": "Arduino Uno"
+      }
+    }
+"""
+
+from __future__ import annotations
+
+import argparse
+import json
+import re
+import sys
+import urllib.request
+from collections import OrderedDict
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Callable, Iterable
+
+ARDUINO_VENDORS = {
+    "2341": "Arduino SA",
+    "2a03": "Arduino LLC",
+}
+
+
+@dataclass(frozen=True)
+class BoardSource:
+    name: str
+    url: str
+
+
+OFFICIAL_BOARD_SOURCES = (
+    BoardSource(
+        "ArduinoCore-avr",
+        "https://raw.githubusercontent.com/arduino/ArduinoCore-avr/master/boards.txt",
+    ),
+    BoardSource(
+        "ArduinoCore-sam",
+        "https://raw.githubusercontent.com/arduino/ArduinoCore-sam/master/boards.txt",
+    ),
+    BoardSource(
+        "ArduinoCore-samd",
+        "https://raw.githubusercontent.com/arduino/ArduinoCore-samd/master/boards.txt",
+    ),
+    BoardSource(
+        "ArduinoCore-megaavr",
+        "https://raw.githubusercontent.com/arduino/ArduinoCore-megaavr/master/boards.txt",
+    ),
+    BoardSource(
+        "ArduinoCore-mbed",
+        "https://raw.githubusercontent.com/arduino/ArduinoCore-mbed/main/boards.txt",
+    ),
+    BoardSource(
+        "ArduinoCore-renesas",
+        "https://raw.githubusercontent.com/arduino/ArduinoCore-renesas/main/boards.txt",
+    ),
+    BoardSource(
+        "ArduinoCore-zephyr",
+        "https://raw.githubusercontent.com/arduino/ArduinoCore-zephyr/main/boards.txt",
+    ),
+)
+
+_BOARD_NAME_RE = re.compile(r"^(?P<board>[A-Za-z0-9_]+)\.name=(?P<name>.+)$")
+_BOARD_VID_RE = re.compile(
+    r"^(?P<board>[A-Za-z0-9_]+)\.(?:vid\.(?P<index_a>\d+)|"
+    r"upload_port\.(?P<index_b>\d+)\.vid)="
+    r"0x(?P<vid>[0-9A-Fa-f]{4})$"
+)
+_BOARD_PID_RE = re.compile(
+    r"^(?P<board>[A-Za-z0-9_]+)\.(?:pid\.(?P<index_a>\d+)|"
+    r"upload_port\.(?P<index_b>\d+)\.pid)="
+    r"0x(?P<pid>[0-9A-Fa-f]{4})$"
+)
+
+
+def _fetch(url: str, *, timeout: float = 30.0) -> str:
+    req = urllib.request.Request(url, headers={"User-Agent": "fbuild-bot/1.0"})
+    with urllib.request.urlopen(req, timeout=timeout) as resp:
+        return resp.read().decode("utf-8", errors="replace")
+
+
+def _normalize_product_name(name: str) -> str:
+    return re.sub(r"\s+", " ", name).strip()
+
+
+def _collapse_board_products(names: list[str]) -> str:
+    unique = sorted(set(_normalize_product_name(name) for name in names))
+    if len(unique) == 1:
+        return unique[0]
+    return " / ".join(unique)
+
+
+def _merge_fill_gaps(
+    base: dict[str, dict[str, str]],
+    supplement: dict[str, dict[str, str]],
+) -> dict[str, dict[str, str]]:
+    out = dict(base)
+    for key, value in sorted(supplement.items()):
+        out.setdefault(key, value)
+    return out
+
+
+def parse_boards_txt(text: str) -> dict[str, dict[str, str]]:
+    """Parse Arduino-style VID/PID declarations from `boards.txt`."""
+    board_names: dict[str, str] = {}
+    vids: dict[tuple[str, str], str] = {}
+    pids: dict[tuple[str, str], str] = {}
+
+    for raw_line in text.splitlines():
+        line = raw_line.strip()
+        name_match = _BOARD_NAME_RE.match(line)
+        if name_match:
+            board_names[name_match.group("board")] = name_match.group("name").strip()
+            continue
+        vid_match = _BOARD_VID_RE.match(line)
+        if vid_match:
+            key = (
+                vid_match.group("board"),
+                vid_match.group("index_a") or vid_match.group("index_b"),
+            )
+            vids[key] = vid_match.group("vid").lower()
+            continue
+        pid_match = _BOARD_PID_RE.match(line)
+        if pid_match:
+            key = (
+                pid_match.group("board"),
+                pid_match.group("index_a") or pid_match.group("index_b"),
+            )
+            pids[key] = pid_match.group("pid").lower()
+
+    names_by_vidpid: dict[str, list[str]] = {}
+    for key, vid in vids.items():
+        if vid not in ARDUINO_VENDORS or key not in pids:
+            continue
+        board_name = board_names.get(key[0])
+        if board_name is None:
+            continue
+        names_by_vidpid.setdefault(f"{vid}:{pids[key]}", []).append(board_name)
+
+    return {
+        vidpid: {
+            "vendor": ARDUINO_VENDORS[vidpid.split(":", 1)[0]],
+            "product": _collapse_board_products(names),
+        }
+        for vidpid, names in sorted(names_by_vidpid.items())
+    }
+
+
+def collect(
+    *,
+    fetch: Callable[[str], str] = _fetch,
+    sources: Iterable[BoardSource] = OFFICIAL_BOARD_SOURCES,
+) -> dict[str, dict[str, str]]:
+    entries: dict[str, dict[str, str]] = {}
+    for source in sources:
+        try:
+            text = fetch(source.url)
+        except Exception as e:
+            print(f"warning: {source.name}: {source.url}: fetch failed: {e}", file=sys.stderr)
+            continue
+
+        rows = parse_boards_txt(text)
+        entries = _merge_fill_gaps(entries, rows)
+        print(f"{source.name}: {len(rows)} PID(s)", file=sys.stderr)
+    return dict(sorted(entries.items()))
+
+
+def main() -> int:
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument("--out", required=True, type=Path)
+    args = parser.parse_args()
+
+    entries = collect()
+    args.out.write_text(
+        json.dumps(OrderedDict(sorted(entries.items())), indent=2, ensure_ascii=False)
+        + "\n",
+        encoding="utf-8",
+    )
+    print(f"wrote {args.out}: {len(entries)} Arduino PID(s)")
+    return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main())

online-data-tools/test_arduino_usb_pids.py

@@ -0,0 +1,114 @@
+#!/usr/bin/env -S uv run --no-project --with pytest --script
+# /// script
+# requires-python = ">=3.10"
+# dependencies = ["pytest"]
+# ///
+"""Tests for fetch_arduino_usb_pids.py."""
+
+from __future__ import annotations
+
+import json
+import sys
+from pathlib import Path
+
+HERE = Path(__file__).resolve().parent
+sys.path.insert(0, str(HERE))
+import fetch_arduino_usb_pids  # noqa: E402
+
+
+BOARDS_FIXTURE = """\
+uno.name=Arduino Uno
+uno.vid.0=0x2341
+uno.pid.0=0x0043
+uno.vid.1=0x2A03
+uno.pid.1=0x0043
+
+leonardo.name=Arduino Leonardo
+leonardo.vid.0=0x2341
+leonardo.pid.0=0x8036
+
+mkrwifi1010.name=Arduino MKR WiFi 1010
+mkrwifi1010.vid.0=0x2341
+mkrwifi1010.pid.0=0x8057
+
+zero_edbg.name=Arduino Zero (Programming Port)
+zero_edbg.vid.0=0x03eb
+zero_edbg.pid.0=0x2157
+"""
+
+
+def test_parse_boards_txt_extracts_arduino_vids_only() -> None:
+    rows = fetch_arduino_usb_pids.parse_boards_txt(BOARDS_FIXTURE)
+
+    assert rows["2341:0043"] == {
+        "vendor": "Arduino SA",
+        "product": "Arduino Uno",
+    }
+    assert rows["2a03:0043"] == {
+        "vendor": "Arduino LLC",
+        "product": "Arduino Uno",
+    }
+    assert rows["2341:8036"]["product"] == "Arduino Leonardo"
+    assert rows["2341:8057"]["product"] == "Arduino MKR WiFi 1010"
+    assert "03eb:2157" not in rows
+    assert list(rows) == sorted(rows)
+
+
+def test_parse_boards_txt_collapses_duplicate_pid_names() -> None:
+    fixture = BOARDS_FIXTURE + """\
+uno_clone.name=Arduino Uno Compatible
+uno_clone.vid.0=0x2341
+uno_clone.pid.0=0x0043
+"""
+    rows = fetch_arduino_usb_pids.parse_boards_txt(fixture)
+
+    assert rows["2341:0043"]["product"] == "Arduino Uno / Arduino Uno Compatible"
+
+
+def test_collect_uses_first_official_source_on_conflict() -> None:
+    first = fetch_arduino_usb_pids.BoardSource("first", "first")
+    second = fetch_arduino_usb_pids.BoardSource("second", "second")
+
+    def fake_fetch(url: str) -> str:
+        if url == "first":
+            return BOARDS_FIXTURE
+        return """\
+uno.name=Different Uno Name
+uno.vid.0=0x2341
+uno.pid.0=0x0043
+nano_matter.name=Arduino Nano Matter
+nano_matter.vid.0=0x2341
+nano_matter.pid.0=0x0072
+"""
+
+    rows = fetch_arduino_usb_pids.collect(
+        fetch=fake_fetch,
+        sources=(first, second),
+    )
+
+    assert rows["2341:0043"]["product"] == "Arduino Uno"
+    assert rows["2341:0072"]["product"] == "Arduino Nano Matter"
+
+
+def test_main_emits_merge_sources_shape(tmp_path: Path) -> None:
+    out = tmp_path / "arduino.json"
+    old_collect = fetch_arduino_usb_pids.collect
+    try:
+        fetch_arduino_usb_pids.collect = lambda: {
+            "2341:0043": {
+                "vendor": "Arduino SA",
+                "product": "Arduino Uno",
+            }
+        }
+        sys.argv = ["fetch_arduino_usb_pids.py", "--out", str(out)]
+        assert fetch_arduino_usb_pids.main() == 0
+    finally:
+        fetch_arduino_usb_pids.collect = old_collect
+
+    data = json.loads(out.read_text(encoding="utf-8"))
+    assert data == {
+        "2341:0043": {
+            "vendor": "Arduino SA",
+            "product": "Arduino Uno",
+        }
+    }