File size: 8,785 Bytes
e0265b9 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 | from __future__ import annotations
import html
import json
import re
from dataclasses import asdict
from datetime import datetime, timezone
from pathlib import Path
from typing import Any
from adam.executor import ToolContext
from adam.monitoring import SystemMonitor
def _slug(value: str) -> str:
cleaned = re.sub(r"[^A-Za-z0-9._ -]+", "", value).strip(" .")
cleaned = re.sub(r"\s+", "_", cleaned)
return (cleaned or "adam_project")[:80]
def _project_folder(context: ToolContext, project_name: str) -> Path:
base = (context.root / "data" / "projects").resolve()
folder = (base / _slug(project_name)).resolve()
if base != folder and base not in folder.parents:
raise ValueError("Project folder resolved outside ADAM's data directory.")
folder.mkdir(parents=True, exist_ok=True)
return folder
def _write_json(path: Path, payload: dict[str, Any]) -> None:
path.parent.mkdir(parents=True, exist_ok=True)
temporary = path.with_suffix(path.suffix + ".tmp")
temporary.write_text(json.dumps(payload, indent=2), encoding="utf-8")
temporary.replace(path)
def _simulate(
context: ToolContext,
updates: list[tuple[int, str]],
*,
multiplier: float = 1.0,
) -> None:
for percent, message in updates:
context.log(message)
context.progress(percent, message)
context.wait(multiplier)
def collect_dataset(
context: ToolContext,
subject: str,
image_count: int,
project_name: str,
) -> dict[str, Any]:
folder = _project_folder(context, project_name)
dataset = folder / "dataset"
dataset.mkdir(exist_ok=True)
_simulate(
context,
[
(8, f"Preparing collection query for {subject}"),
(28, "Checking collector configuration"),
(52, "Creating candidate image manifest"),
(76, "Recording source and license review fields"),
(100, "Dataset collection manifest is ready"),
],
)
manifest = {
"mode": "demo",
"notice": (
"No images were downloaded. Connect your Dataset Collector backend in "
"config/tools.json to perform real collection."
),
"subject": subject,
"requested_images": int(image_count),
"candidates": [],
"created_at": datetime.now(timezone.utc).isoformat(),
}
_write_json(dataset / "collection_manifest.json", manifest)
return {"output_folder": str(folder)}
def prepare_dataset(context: ToolContext, project_name: str) -> dict[str, Any]:
folder = _project_folder(context, project_name)
dataset = folder / "dataset"
dataset.mkdir(exist_ok=True)
_simulate(
context,
[
(12, "Validating dataset manifest"),
(34, "Checking file integrity and dimensions"),
(58, "Running duplicate analysis"),
(81, "Preparing normalized dataset layout"),
(100, "Dataset preparation report is ready"),
],
)
report = {
"mode": "demo",
"valid_images": 0,
"duplicates_removed": 0,
"rejected_images": 0,
"ready_for_captioning": False,
"notice": "Connect a real preparation backend to process collected files.",
}
_write_json(dataset / "preparation_report.json", report)
return {"output_folder": str(folder)}
def generate_captions(
context: ToolContext,
subject: str,
project_name: str,
) -> dict[str, Any]:
folder = _project_folder(context, project_name)
captions = folder / "captions"
captions.mkdir(exist_ok=True)
_simulate(
context,
[
(15, "Loading prepared dataset report"),
(39, "Preparing captioning policy"),
(67, "Creating editable caption template"),
(88, "Checking caption consistency"),
(100, "Caption review file is ready"),
],
)
(captions / "captions_demo.txt").write_text(
"# ADAM demo caption template\n"
f"# Subject: {subject}\n"
"# No image captions were generated because no real backend is connected.\n",
encoding="utf-8",
)
return {"output_folder": str(folder)}
def train_lora(
context: ToolContext,
subject: str,
project_name: str,
epochs: int,
) -> dict[str, Any]:
folder = _project_folder(context, project_name)
training = folder / "training"
training.mkdir(exist_ok=True)
updates = [(4, "Validating trainer configuration")]
for epoch in range(1, max(1, int(epochs)) + 1):
percent = 8 + int(epoch / max(1, int(epochs)) * 84)
updates.append((percent, f"Simulating epoch {epoch}/{epochs}"))
updates.extend(
[(96, "Writing transparent demo summary"), (100, "Training simulation complete")]
)
_simulate(context, updates, multiplier=0.65)
_write_json(
training / "training_summary.json",
{
"mode": "demo",
"subject": subject,
"epochs_requested": int(epochs),
"model_created": False,
"notice": (
"This was a workflow simulation. No GPU training ran and no model "
"weights were created."
),
},
)
return {"output_folder": str(folder)}
def generate_previews(
context: ToolContext,
subject: str,
project_name: str,
preview_count: int,
model_name: str = "",
checkpoint: str = "",
prompt: str = "",
seed: int = 0,
) -> dict[str, Any]:
folder = _project_folder(context, project_name)
previews = folder / "previews"
previews.mkdir(exist_ok=True)
count = max(1, min(int(preview_count), 100))
_simulate(
context,
[
(10, "Loading preview generator configuration"),
(30, f"Preparing {count} preview tasks"),
(62, "Rendering demo preview cards"),
(86, "Writing preview manifest"),
(100, "Preview outputs are ready"),
],
)
safe_subject = html.escape(subject)
for index in range(1, count + 1):
svg = f"""<svg xmlns="http://www.w3.org/2000/svg" width="768" height="512">
<defs><linearGradient id="g" x1="0" y1="0" x2="1" y2="1">
<stop offset="0" stop-color="#07111c"/><stop offset="1" stop-color="#0b2131"/></linearGradient>
</defs><rect width="768" height="512" fill="url(#g)"/>
<circle cx="384" cy="210" r="72" fill="#139cff" opacity=".13"/>
<circle cx="384" cy="210" r="38" fill="#eff8ff"/>
<g fill="none" stroke-width="6" opacity=".85">
<ellipse cx="384" cy="210" rx="150" ry="55" stroke="#55e75b"/>
<ellipse cx="384" cy="210" rx="150" ry="55" stroke="#168cff" transform="rotate(60 384 210)"/>
<ellipse cx="384" cy="210" rx="150" ry="55" stroke="#ff3948" transform="rotate(120 384 210)"/>
</g>
<text x="384" y="385" fill="#f3f8fd" font-size="30" text-anchor="middle"
font-family="Segoe UI, sans-serif">{safe_subject}</text>
<text x="384" y="425" fill="#6d8294" font-size="18" text-anchor="middle"
font-family="Segoe UI, sans-serif">ADAM DEMO PREVIEW {index:02d}</text></svg>"""
(previews / f"preview_{index:02d}.svg").write_text(svg, encoding="utf-8")
_write_json(
previews / "preview_manifest.json",
{
"mode": "demo",
"subject": subject,
"model_name": model_name,
"checkpoint": checkpoint,
"prompt": prompt,
"seed": int(seed),
"preview_count": count,
"notice": "These are branded placeholders, not model-generated images.",
},
)
return {"output_folder": str(folder)}
def notify_complete(context: ToolContext, project_name: str) -> dict[str, Any]:
folder = _project_folder(context, project_name)
_simulate(
context,
[
(30, "Collecting workflow results"),
(70, "Recording completion status"),
(100, "Workflow complete"),
],
multiplier=0.5,
)
_write_json(
folder / "completion.json",
{
"job_id": context.job_id,
"project_name": project_name,
"completed_at": datetime.now(timezone.utc).isoformat(),
},
)
return {"output_folder": str(folder)}
def inspect_system(context: ToolContext, project_name: str) -> dict[str, Any]:
del project_name
context.progress(20, "Reading system sensors")
monitor = SystemMonitor(context.root)
try:
snapshot = monitor.snapshot()
finally:
monitor.close()
context.log(
f"CPU {snapshot.cpu_percent:.0f}% 路 RAM {snapshot.memory_percent:.0f}% 路 "
f"GPU {snapshot.gpu_percent:.0f}% 路 {snapshot.gpu_name}"
)
context.progress(100, "System snapshot complete")
return {"system_snapshot": asdict(snapshot)}
|