beyarkay's picture
Lane re-export scripts and progress, window debugging scripts
0a3edb6 verified
Raw History Blame Contribute Delete
2.48 kB
import json, sys
import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
import numpy as np
sys.path.insert(0, "scripts")
from render_png import draw, _icon
from compare import region_flows
from factorion import Channel, str2item, Direction
DELTA = {1: (0, -1), 2: (1, 0), 3: (0, 1), 4: (-1, 0)}
regions = json.load(open(sys.argv[1]))
best = max((r for r in regions if region_flows(r)), key=lambda r: len(r.get("lanes") or []))
r, w, flows, truth = region_flows(best)
truth_lr = {(x, y, {1: "L", 2: "R"}[l]): v for (x, y, l), v in truth.items()}
W, H = r.ent.shape
dens = np.zeros((W, H))
for (x, y, _) in truth_lr:
if 0 <= x < W and 0 <= y < H:
dens[x, y] += 1
from factorion import str2ent
MACH = [str2ent(n).value for n in ('assembling_machine_1', 'stone_furnace', 'steel_furnace', 'electric_furnace', 'inserter', 'long_handed_inserter')]
asm = np.isin(w[Channel.ENTITIES.value], MACH).astype(float)
S_ = int(sys.argv[2]) if len(sys.argv) > 2 else 16
score = lambda x, y: min(dens[x:x + S_, y:y + S_].sum(), 3 * asm[x:x + S_, y:y + S_].sum())
cx, cy = max(((x, y) for x in range(0, max(1, W - S_), 4) for y in range(0, max(1, H - S_), 4)), key=lambda p: score(*p))
crop = (cx, min(W, cx + S_), cy, min(H, cy + S_))
fig, axes = plt.subplots(1, 1, figsize=(15, 15)); axes = [axes]
for ax, title, labels in ((axes[0], "Items on each lane in the save (icon on the lane's side of the belt)", truth_lr),):
draw(w, ax, title, crop)
for (x, y, lane), its in labels.items():
if not (crop[0] <= x < crop[1] and crop[2] <= y < crop[3]):
continue
dx, dy = DELTA.get(int(w[Channel.DIRECTION.value, x, y]), (0, 0))
lx, ly = (dy, -dx) if lane == "L" else (-dy, dx)
px, py = x + 0.5 + 0.24 * lx, y + 0.5 + 0.24 * ly
if len(its) == 1:
item = str2item(next(iter(its)))
img = _icon(item.value) if item else None
if img is not None:
ax.imshow(img, extent=(px - 0.17, px + 0.17, py + 0.17, py - 0.17), zorder=8)
continue
if its or labels is not truth_lr:
ax.text(px, py, "?" if its or True else "", ha="center", va="center", fontsize=7, color="#c0392b", zorder=8, weight="bold")
ax.set_xlim(crop[0], crop[1]); ax.set_ylim(crop[3], crop[2]); ax.set_aspect("equal")
fig.savefig("/home/user/factorion/plots/lane_items_spotcheck.png", dpi=90, bbox_inches="tight")
print("region", best["x"], best["y"], "crop", crop)