203 lines
12 KiB
Python
203 lines
12 KiB
Python
"""Read-only workflow downloads, including full history and authorized artifacts."""
|
||
import csv
|
||
from datetime import datetime, timezone
|
||
import hashlib
|
||
import io
|
||
import json
|
||
from pathlib import Path
|
||
import re
|
||
import tempfile
|
||
import zipfile
|
||
|
||
from fastapi import HTTPException
|
||
|
||
MAX_EXPORT_BYTES = 256 * 1024 * 1024
|
||
LABELS = {'assessing': '整理需求', 'needs_input': '等待补充需求', 'awaiting_selection': '等待审核方案',
|
||
'designing': '正在设计与检查', 'exporting': '正在打包成果', 'needs_operator': '服务器自动恢复中',
|
||
'draft_ready': '草稿:检查未通过', 'delivery_ready': '检查通过,等待审核',
|
||
'completed': '已完成,可下载', 'failed': '本次设计未能完成', 'cancelled': '已取消'}
|
||
REPORT = re.compile(r'^(?:erc|drc|drc-reopened|mechanical|final-audit|circuit-failure|repair-baseline|semantic|metrics|failure|schema-failure|attempt-\d+-failure|repair-\d+-(?:context|decision)|part-mechanical-preflight|footprint-repair-\d+-\d+|recovery-\d+|footprint-review-\d+|selection-bindings|selection-native-\d+|selection-failure-\d+|native-checks|intake-check|plan-check|plan-review|pcb-(?:dsn-export|layout|placement-check|pad-resolution|routing-completion|connectivity-refresh|mechanical-geometry|mechanical-preflight|mechanical-check|mounting-holes|pipeline|routing|net-mapping|drill-preservation|copper-audit|signal-angle-audit)|schematic-(?:net-plan|pin-plan|logical-netlist|wiring|wiring-check|layout|native-pin-aliases|sheet-history))\.json$')
|
||
|
||
|
||
def json_bytes(value):
|
||
return (json.dumps(value, ensure_ascii=False, indent=2) + '\n').encode('utf-8')
|
||
|
||
|
||
def clean(value, root):
|
||
"""Reports need design evidence, never provider secrets or host filesystem paths."""
|
||
if isinstance(value, dict):
|
||
return {k: clean(v, root) for k, v in value.items()
|
||
if k.lower() not in {'api_key', 'authorization', 'experimental_bearer_token', 'access_token',
|
||
'absolute_path', 'native_job_id', 'render_override'}}
|
||
if isinstance(value, list):
|
||
return [clean(v, root) for v in value]
|
||
if isinstance(value, str):
|
||
if value.startswith(('/home/', '/Users/', '/tmp/')):
|
||
return '[服务器本地文件]/' + Path(value).name
|
||
return value.replace(str(root), '[任务运行目录]')
|
||
return value
|
||
|
||
|
||
def timestamp(value):
|
||
try:
|
||
return datetime.fromtimestamp(float(value), timezone.utc).isoformat()
|
||
except (ValueError, TypeError, OverflowError):
|
||
return ''
|
||
|
||
|
||
def collect(settings, store, engine, tid, *, include_checks=True):
|
||
data = store.task(tid)
|
||
if not data:
|
||
raise HTTPException(404, '任务不存在')
|
||
directory = (settings.core_root / 'tasks' / tid).resolve()
|
||
if not directory.is_relative_to((settings.core_root / 'tasks').resolve()):
|
||
raise HTTPException(404, '任务路径不可用')
|
||
core, notices = {}, []
|
||
path = directory / 'task.json'
|
||
if path.is_file() and not path.is_symlink():
|
||
try:
|
||
core = json.loads(path.read_text())
|
||
except (OSError, ValueError):
|
||
notices.append('原始任务快照暂不可读;已包含网页任务快照。')
|
||
with store.db() as db:
|
||
# One query captures all rows through an export watermark, without the UI/SSE limit.
|
||
events = [{**dict(row), 'data': json.loads(row['data'])} for row in db.execute(
|
||
'SELECT seq,kind,data,created FROM events WHERE task_id=? ORDER BY seq', (tid,))]
|
||
reports, total = [], 0
|
||
folders = [directory / 'calls', directory / 'project/work', directory / 'project/output'] if include_checks else []
|
||
for folder in folders:
|
||
if not folder.is_dir():
|
||
continue
|
||
for file in sorted(folder.rglob('*.json')):
|
||
if not REPORT.fullmatch(file.name) or file.is_symlink() or not file.resolve().is_relative_to(directory):
|
||
continue
|
||
try:
|
||
if file.stat().st_size > MAX_EXPORT_BYTES - total:
|
||
raise HTTPException(413, '检查记录超过 256MiB,请分项下载工程文件和执行记录。')
|
||
raw = file.read_bytes()
|
||
total += len(raw)
|
||
if total > MAX_EXPORT_BYTES:
|
||
raise HTTPException(413, '检查记录超过 256MiB,请分项下载工程文件和执行记录。')
|
||
reports.append({'name': str(file.relative_to(directory)), 'data': clean(json.loads(raw), settings.core_root)})
|
||
except (OSError, ValueError):
|
||
notices.append('导出时暂不可读的记录:' + str(file.relative_to(directory)))
|
||
artifacts = engine.artifacts(tid)
|
||
report = {'exported_at': datetime.now(timezone.utc).isoformat(),
|
||
'event_count': len(events), 'last_event_seq': events[-1]['seq'] if events else 0,
|
||
'snapshot_note': '这是导出时已有资料的快照;任务后续产生的信息需重新下载。未审核放行的制造包不会包含在内。',
|
||
'task': engine.public_task(data),
|
||
'conversation': core.get('conversation', []),
|
||
'attachments': [{'name':item.get('name') or item.get('description') or Path(item.get('path','')).name, 'kind':item.get('kind'), 'sha256':item.get('sha256')} for item in core.get('assets', [])],
|
||
'plan': core.get('plan') or core.get('accepted_plan') or core.get('proposal'),
|
||
'parts': [{k: item.get(k) for k in ['reference', 'lib_id', 'value', 'footprint', 'role', 'group', 'metadata']} for item in core.get('parts', [])],
|
||
'events': events, 'checks': reports, 'artifacts': artifacts, 'notices': notices}
|
||
for key in ['created_at', 'updated_at', 'design_started_at']:
|
||
if key in core:
|
||
report[key] = timestamp(core[key])
|
||
return clean(report, settings.core_root), core, directory
|
||
|
||
|
||
def csv_bytes(report):
|
||
stream = io.StringIO(newline='')
|
||
writer = csv.writer(stream)
|
||
writer.writerow(['序号', '时间(UTC)', '事件类型', '说明', '详细数据'])
|
||
def safe(value):
|
||
text = str(value)
|
||
return "'" + text if text.lstrip().startswith(('=', '+', '-', '@', '\t', '\r')) else text
|
||
for event in report['events']:
|
||
writer.writerow([event['seq'], timestamp(event['created']), safe(event['kind']),
|
||
safe(event['data'].get('message', '')), safe(json.dumps(event['data'], ensure_ascii=False))])
|
||
return stream.getvalue().encode('utf-8-sig')
|
||
|
||
|
||
def markdown(report):
|
||
task, plan = report['task'], report.get('plan') or {}
|
||
lines = ['# PCB 生成流程资料', '', task.get('title') or task['task_id'], '',
|
||
f"- 任务:{task['task_id']}", f"- 修订:{task.get('revision', '')}",
|
||
f"- 状态:{LABELS.get(task['state'], task['state'])}",
|
||
f"- 导出时间:{report['exported_at']}", f"- 完整执行记录:{report['event_count']} 条", '', report['snapshot_note'], '',
|
||
'## 用户需求与补充', '']
|
||
for index, entry in enumerate(report['conversation'], 1):
|
||
label = '方案评论' if entry.get('kind') == 'proposal_feedback' else entry.get('role', '')
|
||
lines += [f"### {index}. {label}", '', entry.get('text', ''), '']
|
||
lines += ['## 当前问题', *['- ' + q for q in task.get('questions', [])], '', '## 方案与选择', '']
|
||
if plan.get('summary'):
|
||
lines += [plan['summary'], '']
|
||
for option in (task.get('proposal') or {}).get('options', []):
|
||
lines += [f"### 方案 {option['id']}:{option['summary']}", option['implementation'], '',
|
||
*['- 优点:' + item for item in option.get('advantages', [])],
|
||
*['- 限制:' + item for item in option.get('disadvantages', [])], '']
|
||
lines += ['已选方案:' + str(task.get('selected_option') or '尚未选择'), '', '## 要求与默认假设', '']
|
||
for key in ['requirements', 'interfaces', 'assumptions']:
|
||
lines += ['- ' + item for item in plan.get(key, [])]
|
||
if plan.get('board'):
|
||
lines += ['', '板框:' + json.dumps(plan['board'], ensure_ascii=False)]
|
||
lines += ['', '## 器件清单', '', '| 器件 | 型号/数值 | 符号 | 封装 | 功能分组 |', '|---|---|---|---|---|']
|
||
def cell(value):
|
||
return str(value or '').replace('|', '\\|').replace('\n', ' ')
|
||
for part in report['parts']:
|
||
lines.append('| ' + ' | '.join(cell(part.get(k)) for k in ['reference', 'value', 'lib_id', 'footprint', 'group']) + ' |')
|
||
lines += ['', '## 诊断与检查结论', '', task.get('error') or '当前未记录任务错误。']
|
||
for key in ['manufacturing_blockers', 'findings']:
|
||
lines += ['- ' + str(item) for item in task.get(key, [])]
|
||
lines += ['', '## 成果文件清单', '']
|
||
for item in report['artifacts']:
|
||
lines += [f"- {item['kind']}:{item['name']};SHA-256:{item['sha256']}"]
|
||
lines += ['', '## 完整执行记录', '']
|
||
for event in report['events']:
|
||
lines += [f"### {event['seq']} · {timestamp(event['created'])} · {event['kind']}", '',
|
||
event['data'].get('message', ''), '', '```json', json.dumps(event['data'], ensure_ascii=False, indent=2), '```', '']
|
||
lines += ['## 详细检查记录', '']
|
||
for check in report['checks']:
|
||
lines += ['### ' + check['name'], '', '```json', json.dumps(check['data'], ensure_ascii=False, indent=2), '```', '']
|
||
lines += ['## 导出说明', '', *['- ' + notice for notice in report['notices']]]
|
||
return '\n'.join(lines).encode('utf-8-sig')
|
||
|
||
|
||
def bundle(report, core, directory, authorized_file):
|
||
"""Package only existing downloadable artifacts; never relax artifact release gates."""
|
||
output = tempfile.SpooledTemporaryFile(max_size=8 * 1024 * 1024, mode='w+b')
|
||
written = 0
|
||
try:
|
||
with zipfile.ZipFile(output, 'w', compression=zipfile.ZIP_DEFLATED) as archive:
|
||
def write(name, content):
|
||
nonlocal written
|
||
written += len(content)
|
||
if written > MAX_EXPORT_BYTES:
|
||
raise HTTPException(413, '全部资料超过 256MiB,请下载流程说明、完整记录,并分项下载工程文件。')
|
||
archive.writestr(name, content)
|
||
for index, asset in enumerate(core.get('assets', []), 1):
|
||
path = Path(asset.get('path', '')).resolve()
|
||
if not path.is_relative_to(directory / 'project/inputs') or not path.is_file() or path.is_symlink():
|
||
report['notices'].append('参考资料暂不可用:' + str(asset.get('name', index)))
|
||
continue
|
||
content = path.read_bytes()
|
||
if hashlib.sha256(content).hexdigest() != asset.get('sha256'):
|
||
report['notices'].append('参考资料摘要不一致,未打包:' + str(asset.get('name', index)))
|
||
continue
|
||
name = re.sub(r'[^\w.() -]', '_', Path(asset.get('name') or path.name).name)
|
||
write(f'参考资料/{index:02d}-{name}', content)
|
||
for item in report['artifacts']:
|
||
try:
|
||
_, path = authorized_file(report['task']['task_id'], item['artifact_id'])
|
||
except HTTPException as exc:
|
||
report['notices'].append(f"成果未包含:{item['name']}(HTTP {exc.status_code}:{exc.detail})")
|
||
continue
|
||
if path.stat().st_size > MAX_EXPORT_BYTES - written:
|
||
raise HTTPException(413, '全部资料超过 256MiB,请分项下载流程资料和工程文件。')
|
||
content = path.read_bytes()
|
||
if hashlib.sha256(content).hexdigest() != item['sha256']:
|
||
report['notices'].append('导出时成果发生变化,未打包:' + item['name'])
|
||
continue
|
||
write('工程成果/' + item['artifact_id'] + '/' + path.name, content)
|
||
write('流程资料.json', json_bytes(report))
|
||
write('完整执行记录.csv', csv_bytes(report))
|
||
write('流程说明.md', markdown(report))
|
||
for check in report['checks']:
|
||
write('检查记录/' + check['name'], json_bytes(check['data']))
|
||
output.seek(0)
|
||
return output
|
||
except BaseException:
|
||
output.close()
|
||
raise
|