Bundled Pacifica dataset#

pacifica-obd ships its own Pacifica (RU-platform) dataset so the toolkit is self-contained: no external ECU database, no configuration path and no network access are required at runtime. The data lives in one resource,

src/pacifica_obd/data/pacifica.json

and is read through importlib.resources by pacifica_obd.pacifica_data, so it also works from an installed wheel or zip.

Loader API#

from pacifica_obd import pacifica_data

pacifica_data.modules()                # list[dict]  — addressed ECUs
pacifica_data.dids("ABS_PN")           # list[dict]  — data items for that ECU
pacifica_data.adaptations("BODY_CHRYSLER")   # list[dict] — writable config
pacifica_data.dtc_text("0CD1")         # "Charge Port Door Control" | None
pacifica_data.dataset()                # the whole parsed mapping

ECU lookups are case-insensitive; unknown modules return [] and unknown DTC codes return None.

Schema and row counts#

pacifica.json is a single JSON object (schema = pacifica-obd.dataset/1, platform = RU):

A modules record looks like:

{
  "name": "Anti Lock Brakes OBD",
  "ecu": "ABSO_PN",
  "years": "2017-2027",
  "powertrain": "gas",
  "bus": "can-c",
  "inittype": 7,
  "ecu_address": 0,
  "request": "0x7E4",
  "response": "0x7EC",
  "response_rule": "request+8",
  "adapter": "0",
  "baudrate": 0,
  "device_ids": [10276],
  "options": ["MY2017 PowerNet"]
}

Bus mapping: inittype 7 → can-c, 9 → can-ihs. response_rule records how the response ID relates to the request ID in the source data: request+8 for the OBD-II block (0x7E0–0x7E7), request-0x280 for the CAN-IHS and most CAN-C body modules, request+1 for the Security Gateway, and explicit where the source gives no derivable rule. Both OBD-II-block and -0x280/CAN-IHS sets are included; the dataset stores absolute IDs, so there is no need to derive them.

How applicability was decided#

The upstream family list has no textual RU / Pacifica flag; it identifies vehicles only by numeric brand/model ids in per-brand columns. The heuristic is:

A row is Pacifica (RU) applicable when its Chrysler brand column carries model id 20 — the MY2017+ "PowerNet" (FCA CAN) Chrysler generation.

Justification:

• Id 20 tags exactly the MY2017-and-later "PowerNet" rows, and the set it produces is the RU module set: BODY_CHRYSLER (MY2017…MY2020-21), PENTASTAR (3.6 L V6 PCM), the full PowerNet body/chassis cluster (ABS_PN, IPC_PN, EPS_PN, ESM_PN, ORC_PN, RFH_PN, SCCM_PN, ACC_PN, …, ICS_PN, HVAC_PN, the door/sliding-door modules) plus the hybrid-only controllers HCP_PN, BPCM_PN, OBCM_PN, AHCP_PN, APM_PN, EAC_PN.
• Those hybrid controllers exist on no other Chrysler-badged model, which pins id 20 to the Pacifica/RU family rather than to a sedan or truck.
• A sibling id group (100/101/102, "MY2016+ / MY2019+ PowerNet") was not used: it does not form the RU set (it adds, for example, a non-RU gasoline engine ECU) and appears to be a different Chrysler PowerNet generation.

Model years are derived from the upstream MYxxxx/MY>xxxx/MYxxxx+ option labels and clamped to the 2017–2027 RU span; rows with no year label default to the full span.

What is and isn't included#

Included (factual identifiers, kept compact):

• Module addressing — request/response CAN IDs, bus, adapter, raw ECU address, and the applicability labels.
• Per-module data items — live-data parameter names with units (pid) and actuator/diagnostic routine names (routine) the source keys to each ECU.
• BCM configuration — BODY_CHRYSLER PowerNet settings: request opcode, bit offset, bit width, setting name, group, and the enumerated allowed values.
• DTC text — short English descriptions for the codes the Pacifica module set can report.

Not included (deliberately):

• No DID hex identifiers, byte sizes or byte→value formulas for the RU module set. The upstream source stores only parameter names/units for these ECUs; the formula-bearing tables it does have are for other market ECUs and do not key to any RU module, so they are omitted rather than guessed.
• No seed/key algorithm or constants in this dataset. seed_key records that per-ECU identifiers exist; the built-in PowerNet seed→key algorithms and their selector live in pacifica_obd.fca_keys (see fca-keys.md), and the pluggable provider model (seedkey.py) remains for anything else.
• No long prose — descriptions are truncated to short phrases.
• Later RU model years that the upstream tagging does not carry a Pacifica flag for (some MY2021+ body rows are unflagged) are not matched by the heuristic; the years field therefore tops out at MY2021 for BODY_CHRYSLER while the vehicle span runs to 2027.

Regenerating the dataset#

The dataset is a deterministic projection of the reference FCA ECU database used during development. To regenerate it, rebuild pacifica.json from that database with the following recipe (no absolute paths are stored anywhere in the repo):

1. Applicable ECUs. SELECT ECUNAME, Device_ID, Option FROM <ecu-list> where the Chrysler brand column contains model id 20. This gives the ECU names and their upstream device ids.
2. Integrity check. Require every applicable ECUNAME to resolve in the ECU-address table (<ecu-units>); abort if any is missing.
3. Modules. For each applicable name, copy unit_name, ECUNAME, inittype, ecuaddress, canid, canidresp, adapter, baudrate, parse canid/canidresp as hex, and map inittype 7→can-c, 9→can-ihs. Compute response_rule (request+8 / request-0x280 / request+1 / explicit). Deduplicate on (ecu, request, response, inittype).
4. DIDs. Join the parameter-device table by Device_name equal to the ECU name or "<ECU>_<Device_ID>", then the parameter link and name/unit tables → kind = "pid". Join the diagnostic-device table by device name, then the device-diagnostics and diagnostic-name tables → kind = "routine". Deduplicate, preserving order.
5. Adaptations. Keep the BCM PowerNet configuration rows whose support column for any applicable BODY_CHRYSLER device id is set; emit request (opcode, hex), bit, size, name, group, array, and the non-empty allowed-value labels.
6. DTC text. Keep fault rows whose (colon-delimited) device list intersects the applicable device ids; for each code keep the shortest non-empty English description.

Finally, run the checks: the JSON must parse, pacifica_data.modules() / pacifica_data.dids("ABS_PN") must be non-empty, and no shipped file may reference a machine-specific path.