CydandClaude Fable 5
f30acc6a94
The operator caught it: the 5.3.65 image showed the mech hanging from its own ground-shadow plate (-y is up in the bridge world). The follow-up pinned down what was and wasn't wrong, and the reconstruction survives untouched. Right, by three independent wire-level proofs: MAD.SKL offsets are a coherent Y-up model (hip +5.29, knees descending, torso ascending -- anatomical only read +Y-up); the BGF vertex extents agree (foot meshes extend -1.0 below their ankle to the sole, torso +3.8 up to the missile pods); and the SHIPPED pod's wire carries the same signs (hip +6.21, knees negative, vehicle root a pure yaw with det +1). Our wire and shipped's are convention-identical -- no flip exists between game and board in either game. Wrong: the ad-hoc chase camera. It borrowed the aerial up-hint from the calibrated live-bridge world (where -y is up) and applied it to a raw scene-space render of Y-up wire data; the symmetric checkerboard ground disguised the inversion. The shipped capture rendered upside-down through the same harness, which is what cleared the wire. Both conventions now documented in the roadmap: wire/scene space is Y-up in both games; the live bridge presents -y-up via its calibrated mirror family (FP_RIGHT_SIGN / CAGE_TWIST_SIGN, COCKPIT-CAGE-NOTES.md). Tooling: pose_probe.py replaces the throwaway harness (updir=+1 scene space default, -1 for the bridge sense). pose_chase_upfixed.png is the corrected milestone image: the MadCat right side up -- chicken-walker legs, purple knee actuators, feet planted, shadow under the feet. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
73 lines
3.1 KiB
Python
73 lines
3.1 KiB
Python
#!/usr/bin/env python3
|
|
"""pose_probe -- offline chase render of the articulated mech in a fifodump.
|
|
|
|
py pose_probe.py <capture.fifodump> <out.png> [updir]
|
|
|
|
Frames the camera on the LAST-articulated root (board.anim_abs), a chase
|
|
offset away, and renders one frame. The proof harness for any "is the mech
|
|
posed/assembled right?" question -- offline, against a copy of the live dump,
|
|
zero interference with a running pod.
|
|
|
|
UP SENSE (updir, default +1): the wire's model/scene data is Y-UP -- proven
|
|
three ways on 2026-07-29: MAD.SKL offsets are anatomically coherent only with
|
|
+Y up (hip +5.29 above ground, knees/ankles descending), the BGF vertex
|
|
extents agree (foot meshes extend -1.0 below their ankle joint, torso +3.8
|
|
above shakey), and the SHIPPED pod's wire carries the same signs (hip +6.2,
|
|
knees negative, yaw-only vehicle root -- no flip anywhere on the wire). So a
|
|
scene-space render wants screen-up = +Y, which is updir=+1 here.
|
|
|
|
The LIVE bridge world is the other way around -- "-y is up here" -- because
|
|
the calibrated live path (fp_cam / the mirrored det -1 root-basis family,
|
|
FP_RIGHT_SIGN, COCKPIT-CAGE-NOTES.md) maps the wire into Dave's GL
|
|
presentation. The first version of this harness copied the aerial camera's
|
|
up hint from that world and rendered the mech hanging from its own shadow;
|
|
the operator caught it. updir=-1 reproduces that view if you ever need it.
|
|
"""
|
|
import sys, os, struct
|
|
sys.path.insert(0, r'C:\VWE\TeslaRel410\dpl3-revive\patha')
|
|
os.environ.setdefault('SDL_VIDEODRIVER', 'dummy')
|
|
import numpy as np
|
|
from vrboard import VirtualBoard, Msg
|
|
from vrview import Renderer
|
|
|
|
cap = sys.argv[1]
|
|
out = sys.argv[2] if len(sys.argv) > 2 else "pose_probe.png"
|
|
updir = float(sys.argv[3]) if len(sys.argv) > 3 else 1.0
|
|
|
|
data = open(cap, "rb").read()
|
|
board = VirtualBoard()
|
|
off = 0
|
|
while off + 8 <= len(data):
|
|
if data[off:off+4] != b'VPXM':
|
|
off += 1; continue
|
|
ln = struct.unpack_from('<I', data, off+4)[0]
|
|
body = data[off+8:off+8+ln]
|
|
if len(body) >= 4:
|
|
a = struct.unpack_from('<I', body, 0)[0]
|
|
if a < 0x100:
|
|
try: board.handle(Msg(False, 0xff, a, body[4:]))
|
|
except Exception: pass
|
|
off += 8 + ln
|
|
|
|
if not board.anim_abs:
|
|
print("no articulated roots in this capture"); sys.exit(1)
|
|
|
|
r = Renderer(w=832, h=512); c = r.cache; c.maybe_rebuild(board)
|
|
root = max(board.anim_abs.values(),
|
|
key=lambda f: float(np.abs(np.array(f[9:12])).sum()))
|
|
target = np.array(root[9:12]); target[1] += updir * 6.0 # mid-body
|
|
eye = target + np.array([30.0, updir * 14.0, 40.0]) # above + beside
|
|
fwd = target - eye; fwd /= np.linalg.norm(fwd); back = -fwd
|
|
right = np.cross([0.0, updir, 0.0], back); right /= np.linalg.norm(right)
|
|
up = np.cross(back, right)
|
|
M = np.eye(4); M[0,:3], M[1,:3], M[2,:3], M[3,:3] = right, up, back, eye
|
|
r.cam_matrix = lambda _b, _M=M: _M
|
|
if c.view is not None and c.view in board.nodes:
|
|
vb = bytearray(board.nodes[c.view].get('body') or b'')
|
|
if len(vb) >= 56:
|
|
struct.pack_into('<f', vb, 52, 1.0e6)
|
|
board.nodes[c.view]['body'] = bytes(vb)
|
|
r.draw(board)
|
|
r.pygame.image.save(r.screen, out)
|
|
print("wrote %s root at %s updir=%+g" % (out, target.round(1), updir))
|