* Ford: use platform codes to fuzzy fingerprint
TODO: write scripts/tests (print platform codes and version ranges etc.)
May close#31052
* get_platform_codes: fix return type
* add print_platform_codes.py script
* print_platform_codes: sort versions
* match_fw_to_car_fuzzy: use set comprehension, and fix typo
* Ford: add missing Mach-E fw
From the route 83a4e056c7072678/2023-11-13--16-51-33 (which is already
in selfdrive/car/tests/routes.py, added in #30691).
* add ford_fuzzy_fingerprint.ipynb notebook
* get_platform_codes: use regex to parse firmware
* test_ford: test_platform_codes_fuzzy_fw
* test_ford: use get_platform_codes in test_fw_versions
* match_fw_to_car_fuzzy: improve comments
* test_ford: add test_platform_codes_spot_check
* test_ford: add test_match_fw_fuzzy
* remove comment from notebook
* TestFordFW: remove engine ECU FW
* update print_platform_codes.py
* remove part number (unecessary)
* platform codes can just use platform hint and model year - software revision not useful
* fuzzy FP on the platform hint and model year hint range
* fix platform codes test
* update notebook
* add notebook
* explain model year hint better
* test part numbers again
* cleanup notebooks
* remove notebook
* cleanup match_fw_to_car_fuzzy and add comments
* update comment
* .
* Revert "remove notebook"
This reverts commit 5d4ca202f2a23601d5c829204119f36a58f2b451.
* add notebook back
* remove PSCM from PLATFORM_CODE_ECUS
ABS and IPMA are the best for uniquely matching, and the radar is always required
* Revert "remove PSCM from PLATFORM_CODE_ECUS"
This reverts commit b7baeac19c18b5aa0c31da52f12054f4bae6e1ff.
* fix from merge
* more fixes
revert
* FW_RE -> FW_PATTERN
* this can actually be set
* conventions
* just add
* convention
* refactor matcher, this brings it more in line with Hyundai. IMPORTANT NOTE:
NOTE THAT WE remove the separation for the different platform code model year hint ranges, I don't see that being a problem
* better/smaller test
* add test to catch overlapping platform codes
* remove nb
* not now
---------
Co-authored-by: Shane Smiskol <shane@smiskol.com>
* useless
* Revert "useless"
This reverts commit 28f0bb9e9794d60eefba8063b47d8ca113308008.
* this forever hangs you disconnect (or 2 hours)
* same timeout as the global websocket
* Revert "same timeout as the global websocket"
This reverts commit 0bd0cb8a38a3e17960c1fae205311d86a9cf8feb.
* setting the timeout affects the entire websocket and disconnects, not just recv timeout
* fix that
* fix test
* ignore modelV2 and longitudinalPlan on the body to avoid fcw alerts
* move joystick_mode up because ignore depends on it
* revert and ignore fcw itself instead
* it's something
* backup
* 16:10
* cleanup
* this is fine
* close
* remove some junk
* no heck
* disos
* real 10
* for some reason this is flipped
* 20hz
* no return
* ae
* tear
* need curve laster
* correct real gains
* fix time
* cleanup
* why the scam
* disable for now
* 0.7
* hdr
* that doesnt work
* what
* hugeoof
* clean up
* cleanup
* fix regs
* welp cant
* is this corrent
* it is sq
* remove
* back
* stg10bit
* back2ten
* Revert "remove"
This reverts commit 18712ab7e103c12621c929cd0f772ecb9b348247.
* 20hz and swb
* correct height
* 10bit
* ui hack for now
* slight
* perfect
* blk64
* ccm
* fix page faults
* template
* set 4x
* is this fine
* try
* this seems to work
* Revert "this seems to work"
This reverts commit d3c9023d3f14bd9394fed2d6276dba777ed0e606.
* needs to be static
* close
* 64 is optimal
* 2
* take
* not 1
* offset
* whats going on
* i have no idea
* less resistence
* box defs
* no
* reduce blur artifacts
* simplify
* fix
* fake short is too much for bright
* can be subzero
* should not use lsvc
* no wasted bit
* cont no slow
* no less than 10bit
* it is based
* wrong
* right
* quart
* shift
* raise noise floor
* 4.5/4.7
* same ballpark
* int is fine
* shane owes me m4a4
* Revert "shane owes me m4a4"
This reverts commit b4283fee18efebedae628a6cfd926ff1416dcfe5.
* back
* Revert "4.5/4.7"
This reverts commit e38f96e90cb5370bd378f6b66def9e7e3ed0ce5d.
* default
* oof
* clean up
* simpilfy
* from sensorinfo
* no div
* better name
* not the wrong one
* not anymore relevant
* too
* not call it debayer
* cl headers
* arg is 2nd
* gone is is_bggr
* define
* no is hdr
* rgb_tmp
* p1
* clean up
* 4
* cant for
* fix somewhre else
* const
* ap
* rects
* just set staruc
* nnew tmp
* pull it for now
* 12
* common rect
* Revert "not anymore relevant"
This reverts commit 1d574673a16cc31b7a255609e07775c3579eef15.
* Revert "too"
This reverts commit c2d4dcc52a859fe799362f9fcc2ffda99b264e50.
* Revert "Revert "too""
This reverts commit 0abbabe1fde51592f1619058638b4ac6a6dee4b3.
* no tol is fine
* rename
* sensor id
* unsgin
* flag
* some linalg
* cast
* should be h ref
* cap
---------
Co-authored-by: Comma Device <device@comma.ai>
young