Spaces:
Running on Zero
Running on Zero
File size: 20,153 Bytes
e994c16 5a17b94 5f76dd3 e994c16 2fb233c e994c16 5a17b94 e994c16 2fb233c 63c4f20 2fb233c e994c16 5f76dd3 5a17b94 5f76dd3 5a17b94 e994c16 2fb233c 5a17b94 2fb233c e994c16 2fb233c 5a17b94 63c4f20 5a17b94 2fb233c e994c16 2fb233c e994c16 2fb233c e994c16 2fb233c e994c16 2fb233c 63c4f20 2fb233c 5f76dd3 2fb233c e994c16 2fb233c 5f76dd3 2fb233c e994c16 5a17b94 e994c16 5a17b94 63c4f20 5f76dd3 5a17b94 63c4f20 5a17b94 63c4f20 5a17b94 63c4f20 5a17b94 63c4f20 5a17b94 e994c16 63c4f20 2fb233c 5a17b94 2fb233c 5f76dd3 2fb233c 5f76dd3 2fb233c 5a17b94 2fb233c 63c4f20 2fb233c 63c4f20 2fb233c 63c4f20 2fb233c 63c4f20 2fb233c 63c4f20 e994c16 2fb233c 63c4f20 5a17b94 2fb233c 5f76dd3 2fb233c e994c16 2fb233c 5a17b94 2fb233c 63c4f20 2fb233c e994c16 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 | """UI components. Defect list rendering and shared visual helpers."""
from __future__ import annotations
import base64
import functools
import html
import json
import time
from pathlib import Path
import re
from typing import Iterable
from data.schemas import LABEL_DISPLAY_NAMES
from data.preprocessing import image_to_data_uri
@functools.lru_cache(maxsize=1)
def _logo_html() -> str:
path = Path(__file__).resolve().parents[1] / "assets" / "logo.jpg"
if not path.exists():
return '<span class="halide-brand-mark halide-brand-mark-text">H</span>'
encoded = base64.b64encode(path.read_bytes()).decode("ascii")
return (
'<img class="halide-brand-mark" '
f'src="data:image/jpeg;base64,{encoded}" alt="" />'
)
COMPARE_DEFAULT_SPLIT = 50
HEADER_HTML = f"""
<div id="halide-header">
<div class="halide-brand-lockup">
{_logo_html()}
<div>
<span class="halide-kicker">Project Halide</span>
<h1>Analog Film Diagnostic Workbench</h1>
</div>
</div>
<div class="halide-model-strip">
<a href="https://huggingface.co/Lonelyguyse1" target="_blank" rel="noreferrer">Lonelyguyse1</a>
<a href="https://huggingface.co/Lonelyguyse1/halide-vision" target="_blank" rel="noreferrer">halide-vision</a>
<span>MiniCPM-V 4.6</span>
<span>Nemotron Mini 4B</span>
</div>
</div>
"""
EMPTY_STATE = '<p class="halide-muted">Awaiting scan.</p>'
REPORT_EMPTY_STATE = (
'<div class="halide-empty-card">'
'<span>Report</span>'
'<strong>No scan analyzed yet</strong>'
'<p>Results, evidence counts, and physical fixes will appear here after a GPU run.</p>'
"</div>"
)
LIGHTTABLE_EMPTY_STATE = (
'<div class="halide-empty-lighttable">'
'<div class="halide-empty-frame-grid">'
'<div><span>Original</span></div>'
'<div><span>Validated overlay</span></div>'
"</div>"
'<div class="halide-empty-center">'
'<span>Ready</span>'
'<strong>No scan loaded</strong>'
"</div>"
"</div>"
)
REPORT_SECTIONS = {
"root cause": "Root cause",
"evidence": "Evidence",
"physical fixes": "Physical fixes",
"confidence": "Confidence",
"next inspection": "Next inspection",
}
def defect_pills_html(label_counts: dict[str, int]) -> str:
"""Render defect counts as colored pills."""
if not label_counts:
return '<p class="halide-muted">No validated defects detected.</p>'
pills: list[str] = []
for label, count in sorted(label_counts.items(), key=lambda kv: -kv[1]):
display = LABEL_DISPLAY_NAMES.get(label, label.replace("_", " ").title())
pills.append(
f'<span class="halide-defect-pill {html.escape(label)}">'
f"{html.escape(display)} <strong>{int(count)}</strong></span>"
)
return '<div class="halide-pill-row">' + "".join(pills) + "</div>"
def compact_label_counts(label_counts: dict[str, int]) -> str:
if not label_counts:
return "none"
parts = []
for label, count in sorted(label_counts.items(), key=lambda kv: (-kv[1], kv[0])):
display = LABEL_DISPLAY_NAMES.get(label, label.replace("_", " ").title())
parts.append(f"{display}: {int(count)}")
return ", ".join(parts)
def defect_table_rows(result: dict | None) -> list[list[str]]:
"""Return rows for the evidence dataframe."""
if not result:
return []
defects = (result.get("defects", {}) or {}).get("defects", []) or []
if not defects:
return [["", "No validated defects", "", ""]]
rows: list[list[str]] = []
for index, defect in enumerate(defects, start=1):
label = str(defect.get("label", ""))
display = LABEL_DISPLAY_NAMES.get(label, label.replace("_", " ").title())
bbox = defect.get("bbox", []) or []
if len(bbox) == 4:
box_text = ", ".join(f"{float(v):.3f}" for v in bbox)
else:
box_text = "invalid"
confidence = defect.get("confidence")
confidence_text = (
"not emitted" if confidence is None else f"{float(confidence):.2f}"
)
rows.append([str(index), display, confidence_text, box_text])
return rows
def stats_html(result: dict) -> str:
"""Render a stats card with defect counts and timing."""
defects = result.get("defects", {}) or {}
diagnosis = result.get("diagnosis", {}) or {}
total = result.get("total_seconds", 0.0) or 0.0
vision_s = defects.get("inference_seconds", 0.0) or 0.0
reasoning_s = diagnosis.get("reasoning_seconds", 0.0) or 0.0
rows: list[str] = []
rows.append(_stat_row("Total defects", str(defects.get("defect_count", 0))))
rows.append(_stat_row("Dropped (invalid)", str(defects.get("dropped_count", 0))))
rows.append(_stat_row("Duplicates removed", str(defects.get("duplicate_count", 0))))
rows.append(_stat_row("Edge artifacts removed", str(defects.get("edge_artifact_count", 0))))
rows.append(_stat_row("CV assist boxes", str(defects.get("classical_assist_count", 0))))
rows.append(_stat_row("Resized for model", "yes" if defects.get("resized_for_model") else "no"))
rows.append(_stat_row("Vision inference", f"{vision_s:.2f}s"))
rows.append(_stat_row("Reasoning", f"{reasoning_s:.2f}s"))
rows.append(_stat_row("Total", f"{total:.2f}s"))
rows.append(_stat_row("Vision model", _truncate(defects.get("model_path", ""), 46)))
rows.append(_stat_row("Reasoning model", _truncate(diagnosis.get("model_path", ""), 46)))
return (
'<div class="halide-panel-title">Run telemetry</div>'
f'<div class="halide-stats">{"".join(rows)}</div>'
)
def _stat_row(label: str, value: str) -> str:
return (
'<div class="halide-stat">'
f'<span class="halide-stat-label">{html.escape(label)}</span>'
f'<span>{html.escape(value)}</span>'
"</div>"
)
def _truncate(s: str, n: int) -> str:
if len(s) <= n:
return s
return "..." + s[-(n - 3):]
def diagnosis_html(text: str) -> str:
"""Render the Nemotron Markdown report into structured HTML."""
return render_markdown_report(text or "(no diagnosis produced)")
def review_frame_html(original, annotated) -> str:
"""Render reliable full-size image links independent of Gradio fullscreen."""
original_uri = image_to_data_uri(original, max_side=1800, quality=92)
overlay_uri = image_to_data_uri(annotated, max_side=1800, quality=92)
return (
'<div class="halide-review-actions">'
'<a href="'
+ original_uri
+ '" target="_blank" rel="noreferrer">Open original</a>'
'<a href="'
+ overlay_uri
+ '" target="_blank" rel="noreferrer">Open overlay</a>'
"</div>"
)
def comparison_viewer_html(original, annotated) -> str:
"""Render an aligned before/after viewer using the same image canvas."""
original_uri = image_to_data_uri(original, max_side=1800, quality=92)
overlay_uri = image_to_data_uri(annotated, max_side=1800, quality=92)
return (
'<div class="halide-compare-viewer" '
f'style="--halide-split: {COMPARE_DEFAULT_SPLIT}%;">'
'<div class="halide-compare-stage">'
f'<img class="halide-compare-base" src="{original_uri}" alt="" />'
'<div class="halide-compare-overlay">'
f'<img src="{overlay_uri}" alt="" />'
"</div>"
'<div class="halide-compare-divider"></div>'
'<span class="halide-compare-label original">Original</span>'
'<span class="halide-compare-label overlay">Validated overlay</span>'
"</div>"
'<input class="halide-compare-range" type="range" min="0" max="100" '
f'value="{COMPARE_DEFAULT_SPLIT}" '
'aria-label="Compare original and validated overlay" '
'oninput="const viewer=this.closest('
"'.halide-compare-viewer'); "
"if (viewer) { viewer.style.setProperty('--halide-split', this.value + '%'); }"
'" />'
"</div>"
)
def render_markdown_report(text: str) -> str:
"""Render the constrained diagnosis Markdown used by Nemotron.
This intentionally supports only the report shapes we request from the
model: section headings, paragraphs, bullets, and numbered fixes.
"""
sections: list[dict[str, list[str] | str]] = []
current: dict[str, list[str] | str] = {
"title": "Report",
"lines": [],
}
for raw_line in text.replace("\r\n", "\n").split("\n"):
line = raw_line.strip()
if line.startswith("## "):
if current["lines"]:
sections.append(current)
title_key = line[3:].strip().lower()
current = {
"title": REPORT_SECTIONS.get(title_key, line[3:].strip() or "Report"),
"lines": [],
}
continue
current["lines"].append(line)
if current["lines"] or not sections:
sections.append(current)
rendered = []
for section in sections:
title = html.escape(str(section["title"]))
body = _render_report_lines(section["lines"]) # type: ignore[arg-type]
rendered.append(
'<section class="halide-report-section">'
f'<div class="halide-report-heading">{title}</div>'
f'<div class="halide-report-body">{body}</div>'
"</section>"
)
return '<div class="halide-report">' + "".join(rendered) + "</div>"
def _render_report_lines(lines: list[str]) -> str:
blocks: list[str] = []
paragraph: list[str] = []
bullet_items: list[str] = []
ordered_items: list[str] = []
def flush_paragraph() -> None:
if paragraph:
blocks.append(
"<p>"
+ " ".join(_render_inline(part) for part in paragraph if part)
+ "</p>"
)
paragraph.clear()
def flush_bullets() -> None:
if bullet_items:
items = "".join(f"<li>{item}</li>" for item in bullet_items)
blocks.append(f"<ul>{items}</ul>")
bullet_items.clear()
def flush_ordered() -> None:
if ordered_items:
items = "".join(f"<li>{item}</li>" for item in ordered_items)
blocks.append(f"<ol>{items}</ol>")
ordered_items.clear()
for line in lines:
if not line:
flush_paragraph()
flush_bullets()
flush_ordered()
continue
if line.startswith("### "):
flush_paragraph()
flush_bullets()
flush_ordered()
blocks.append(
f'<h4 class="halide-report-subheading">{_render_inline(line[4:].strip())}</h4>'
)
continue
numbered = re.match(r"^\d+\.\s+(.*)$", line)
if line.startswith("- "):
flush_paragraph()
flush_ordered()
bullet_items.append(_render_inline(line[2:].strip()))
elif numbered:
flush_paragraph()
flush_bullets()
ordered_items.append(_render_inline(numbered.group(1).strip()))
else:
flush_bullets()
flush_ordered()
paragraph.append(line)
flush_paragraph()
flush_bullets()
flush_ordered()
return "".join(blocks) or '<p class="halide-muted">No report text.</p>'
def _render_inline(text: str) -> str:
escaped = html.escape(text)
escaped = re.sub(r"`([^`]+)`", r"<code>\1</code>", escaped)
escaped = re.sub(r"\*\*([^*]+)\*\*", r"<strong>\1</strong>", escaped)
escaped = re.sub(r"__([^_]+)__", r"<strong>\1</strong>", escaped)
escaped = re.sub(r"(?<!\*)\*([^*]+)\*(?!\*)", r"<em>\1</em>", escaped)
escaped = re.sub(r"(?<!_)_([^_]+)_(?!_)", r"<em>\1</em>", escaped)
return escaped
def metadata_html(result: dict) -> str:
"""Render a compact metadata strip for the current run."""
meta = result.get("film_metadata", {}) or {}
confidence = str(meta.get("metadata_confidence", "low") or "low")
rows = [
_stat_row("Film stock", str(meta.get("film_type", "Unknown"))),
_stat_row("Age", f"{meta.get('film_age_years', 0)} years"),
_stat_row("Storage", str(meta.get("storage", "unknown"))),
_stat_row("Scan DPI", str(meta.get("scan_resolution_dpi", "unknown"))),
_stat_row("Metadata confidence", confidence.title()),
]
return (
'<div class="halide-panel-title">Film dossier</div>'
f'<div class="halide-stats compact">{"".join(rows)}</div>'
)
def confidence_notice_html(result: dict) -> str:
"""Render a small evidence-quality notice from validation metadata."""
defects = result.get("defects", {}) or {}
total = int(defects.get("defect_count", 0) or 0)
dropped = int(defects.get("dropped_count", 0) or 0)
duplicate = int(defects.get("duplicate_count", 0) or 0)
meta = result.get("film_metadata", {}) or {}
confidence = str(meta.get("metadata_confidence", "low") or "low").lower()
confidence_values = [
defect.get("confidence")
for defect in defects.get("defects", []) or []
if defect.get("confidence") is not None
]
if total == 0:
message = "No validated boxes were returned. Inspect the scan before assuming a film fault."
tone = "neutral"
elif not confidence_values:
message = (
"Defect boxes passed schema validation. MiniCPM did not emit numeric "
"per-box confidence for this run."
)
tone = "good"
elif confidence == "low":
message = "Metadata is marked low confidence, so the diagnosis is driven mainly by visible defects."
tone = "caution"
elif dropped or duplicate:
message = f"Schema checks removed {dropped} invalid boxes and {duplicate} duplicate boxes."
tone = "caution"
else:
message = "Schema validation passed for the visible defect evidence."
tone = "good"
return f'<div class="halide-notice {tone}">{html.escape(message)}</div>'
def run_state_html(result: dict | None) -> str:
"""Render a top-level status block for the current diagnosis."""
if not result:
return (
'<div class="halide-run-state idle">'
'<span class="halide-run-eyebrow">Ready</span>'
"<strong>Load a scan to begin inspection.</strong>"
"</div>"
)
defects = result.get("defects", {}) or {}
count = int(defects.get("defect_count", 0) or 0)
dropped = int(defects.get("dropped_count", 0) or 0)
duplicates = int(defects.get("duplicate_count", 0) or 0)
if count:
title = f"{count} validated defect{'s' if count != 1 else ''}"
tone = "active"
else:
title = "No validated defects"
tone = "quiet"
detail = (
f"{dropped} invalid removed, {duplicates} duplicate removed, "
f"{float(result.get('total_seconds', 0.0) or 0.0):.2f}s total"
)
return (
f'<div class="halide-run-state {tone}">'
'<span class="halide-run-eyebrow">Current scan</span>'
f"<strong>{html.escape(title)}</strong>"
f"<span>{html.escape(detail)}</span>"
"</div>"
)
def history_label(entry: dict) -> str:
"""Return a compact plain-text label for the history dropdown."""
stamp = time.strftime(
"%Y-%m-%d %H:%M",
time.localtime(float(entry.get("created_at", 0) or 0)),
)
film = str(entry.get("film_type", "Unknown"))
count = int(entry.get("defect_count", 0) or 0)
return f"{stamp} | {film} | {count} defects"
def history_table_rows(entries: Iterable[dict]) -> list[list[str]]:
rows: list[list[str]] = []
for entry in entries:
stamp = time.strftime(
"%Y-%m-%d %H:%M",
time.localtime(float(entry.get("created_at", 0) or 0)),
)
rows.append(
[
stamp,
str(entry.get("film_type", "Unknown")),
str(int(entry.get("defect_count", 0) or 0)),
compact_label_counts(entry.get("label_counts", {}) or {}),
]
)
return rows
def history_choices(entries: Iterable[dict]) -> list[tuple[str, str]]:
return [(history_label(e), str(e.get("id", ""))) for e in entries if e.get("id")]
def history_detail_html(entry: dict | None) -> str:
if not entry:
return '<p class="halide-muted">Select a diagnosis to review details.</p>'
counts = entry.get("label_counts", {}) or {}
raw = entry.get("raw_json", {}) or {}
meta = raw.get("film_metadata", {}) if raw else {}
confidence = meta.get("metadata_confidence", entry.get("metadata_confidence", "low"))
preview = raw.get("preview", {}) if raw else {}
preview_html = ""
overlay_uri = str(preview.get("overlay", "") or "")
original_uri = str(preview.get("original", "") or "")
if overlay_uri.startswith("data:image/") or original_uri.startswith("data:image/"):
hero_uri = overlay_uri if overlay_uri.startswith("data:image/") else original_uri
preview_html = (
'<div class="halide-history-preview">'
f'<img src="{html.escape(hero_uri, quote=True)}" alt="" />'
'<div class="halide-history-preview-actions">'
)
if original_uri.startswith("data:image/"):
preview_html += (
f'<a href="{html.escape(original_uri, quote=True)}" '
'target="_blank" rel="noreferrer">Original</a>'
)
if overlay_uri.startswith("data:image/"):
preview_html += (
f'<a href="{html.escape(overlay_uri, quote=True)}" '
'target="_blank" rel="noreferrer">Overlay</a>'
)
preview_html += "</div></div>"
stamp = time.strftime(
"%Y-%m-%d %H:%M",
time.localtime(float(entry.get("created_at", 0) or 0)),
)
header_rows = [
_stat_row("Saved", stamp),
_stat_row("Film stock", str(entry.get("film_type", "Unknown"))),
_stat_row("Age", f"{entry.get('film_age_years', '?')} years"),
_stat_row("Storage", str(entry.get("storage", "?"))),
_stat_row("Scan DPI", str(entry.get("scan_dpi", "?"))),
_stat_row("Metadata confidence", str(confidence).title()),
_stat_row("Vision model", _truncate(str(entry.get("vision_model", "")), 60)),
_stat_row("Reasoning model", _truncate(str(entry.get("reasoning_model", "")), 60)),
]
return (
'<div class="halide-history-detail">'
f"{preview_html}"
f'<div class="halide-stats compact">{"".join(header_rows)}</div>'
'<div class="halide-subsection">Defects</div>'
f"{defect_pills_html(counts)}"
'<div class="halide-subsection">Diagnosis</div>'
f"{diagnosis_html(str(entry.get('diagnosis_text', '')))}"
"</div>"
)
def raw_json_text(result_or_entry: dict | None) -> str:
if not result_or_entry:
return "{}"
if "raw_json" in result_or_entry:
payload = result_or_entry.get("raw_json") or {}
else:
payload = result_or_entry
payload = _strip_preview(payload)
return json.dumps(payload, indent=2, sort_keys=True)
def _strip_preview(payload: dict) -> dict:
if not isinstance(payload, dict):
return {}
clean = dict(payload)
if "preview" in clean:
preview = clean.get("preview") or {}
if isinstance(preview, dict):
clean["preview"] = {
key: "[image data URI omitted from JSON view]"
for key in preview
}
else:
clean["preview"] = "[image data URI omitted from JSON view]"
return clean
__all__ = [
"HEADER_HTML",
"EMPTY_STATE",
"LIGHTTABLE_EMPTY_STATE",
"REPORT_EMPTY_STATE",
"compact_label_counts",
"comparison_viewer_html",
"confidence_notice_html",
"defect_table_rows",
"defect_pills_html",
"diagnosis_html",
"history_choices",
"history_detail_html",
"history_label",
"history_table_rows",
"metadata_html",
"render_markdown_report",
"run_state_html",
"raw_json_text",
"review_frame_html",
"stats_html",
]
|