Acasă Explorează Ajutor
Autentificare
18307996893
/
kangshifu-board
1
0
Bifurcare 0
Fisiere Probleme 0 Trageți solicitările 0 Wiki
Ramură: feat/five-views-delivery
Ramuri Etichete
kangshifu-board
/
scripts
/
generate-demo-extension.py

generate-demo-extension.py 8.7 KB
Permalink Istoric Crud

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100 
  1. """Deterministic fictional industry data; no real users, scraped quotes or production IO."""
    2. 

  2. import hashlib
    3. 

  3. import json
    4. 

  4. from pathlib import Path
    5. 

  5. 

  6. ROOT = Path(__file__).resolve().parents[1]
    7. 

  7. seed = json.loads((ROOT / 'data/preview-seed.json').read_text(encoding='utf-8'))
    8. 

  8. DEMO = seed['groups'][0]['demoId']
    9. 

  9. COMMON = dict(demoId=DEMO, dataMode='sample', isSynthetic=True, provenanceType='synthetic_demo')
    10. 

  10. ext = dict(version='ksf.synthetic-extension.20261103.v2', demoId=DEMO, isSynthetic=True,
    11. 

  11.            note='全部为规则生成的虚构观察；不是实际访谈、商品报价、销量或营养检测。',
    12. 

  12.            voices=[], observations=[], prices=[], citations=[], researchTasks=[], collaborationRecords=[])
    13. 

  13. contexts = [
    14. 

  14.     ['工位午餐', '加班后在茶水间', '通勤前的早餐', '午休只剩十五分钟', '下午临时加餐', '晚归后独自吃饭'],
    15. 

  15.     ['宿舍晚自习后', '考试周的午饭', '同寝分享夜宵', '周末留校晚饭', '图书馆回宿舍', '月末控制餐费'],
    16. 

  16.     ['有炉具的露营', '只有保温水壶的徒步', '长途高铁上', '服务区停留', '雨天帐篷里', '出差住酒店'],
    17. 

  17.     ['运动后便餐', '午间轻食', '晚间控制总能量', '轮班后的早餐', '通勤时补一餐', '周末在家吃面'],
    18. 

  18.     ['两人晚餐', '独自加班后回家', '孩子睡后的夜宵', '家庭周末午饭', '冰箱剩菜搭配', '临时替代外卖'],
    19. 

  19.     ['第一次尝红酸汤', '连续第二次吃地域风味', '朋友推荐的椒麻面', '周末尝新', '节前囤几种口味', '与室友分吃新口味'],
    20. 

  20. ]
    21. 

  21. purposes = ['这顿当正餐', '只是两餐之间垫一下', '想和同伴分着吃', '主要想喝热汤', '只是嘴馋试味道', '晚饭之后又有点饿']
    22. 

  22. tools = ['手边只有热水，没有锅', '可以用电煮锅，愿意多等三分钟', '微波炉要排队，暂时不考虑', '需要倒水的做法不方便', '有锅但不想洗太多餐具', '有小炉子但用水有限']
    23. 

  23. tastes = ['喜欢葱香但不想味道留在房间', '想要番茄酸味，不喜欢甜腻', '希望辣油单独放，可以自己决定', '更在意面条有嚼劲', '喜欢浓酱，清淡不是我的首选', '汤鲜就行，不追求地名听起来正宗']
    24. 

  24. behaviors = ['我平时就加鸡蛋，不代表这份面不够', '这次吃一份刚好，不打算再加主食', '如果是加餐，我可能只吃半份', '愿意配蔬菜是为了换口感', '偶尔会再买饭团，要先区分是不是没吃正餐', '包装里有没有碗，比多一包酱更影响我选择']
    25. 

  25. budgets = ['每份五元以内才会常买', '七元上下可以接受，要把运费算进去', '愿意多花两元换方便的制备', '组合装先算清楚每顿多少钱', '促销会囤，平时按需要买', '价格不是第一位，吃得方便更重要']
    26. 

  26. frequencies = ['每周大约一次', '一个月两三次', '只在赶时间时', '最近想换口味时', '出差或值班才', '和朋友一起时']
    27. 

  27. channels = ['模拟社媒分享', '模拟电商评价', '模拟追问记录', '模拟使用日记']
    28. 

  28. for group_index, group in enumerate(seed['groups']):
    29. 

  29.     group_projects = [p for p in seed['projects'] if p['groupId'] == group['id']]
    30. 

  30.     for i in range(2000):
    31. 

  31.         # A coprime permutation spreads all dimensions rather than repeating six stock quotes.
    32. 

  32.         code = (i * 7919 + group_index * 137) % (6 ** 7)
    33. 

  33.         digits = []
    34. 

  34.         for _ in range(7):
    35. 

  35.             digits.append(code % 6); code //= 6
    36. 

  36.         a, b, c, d, e, f, g = digits
    37. 

  37.         raw = f'{contexts[group_index][a]}，{purposes[b]}。{tools[c]}；{tastes[d]}。{behaviors[e]}。{budgets[f]}，我{frequencies[g]}会考虑方便面。'
    38. 

  38.         project = group_projects[i % 2]
    39. 

  39.         ext['voices'].append(dict(COMMON, id=f'SIM-VOICE-{group_index+1}-{i+1:04d}', groupId=group['id'], projectId=project['id'],
    40. 

  40.             title=f'{contexts[group_index][a]}｜{purposes[b]}', rawText=raw, channel=channels[i % 4],
    41. 

  41.             authorLabel=f'模拟观察{group_index+1}-{i+1:04d}', verificationState='synthetic_not_real_interview',
    42. 

  42.             mealPurpose=purposes[b], preparation=tools[c], budget=budgets[f], behavior=behaviors[e],
    43. 

  43.             contentState=['相关分享', '相关分享', '广告待排除', '内容不全', '相关分享'][i % 5]))
    44. 

  44. 

  45. packaging = ['袋装', '碗装', '杯装', '双份组合', '家庭六连包']
    46. 

  46. flavors = ['番茄鸡蛋', '葱油酱香', '菌菇清汤', '轻辣芝士', '红酸汤', '川香椒麻']
    47. 

  47. for i in range(720):
    48. 

  48.     p = seed['projects'][i % 12]
    49. 

  49.     pack_count = [1, 1, 2, 6, None][i % 5]
    50. 

  50.     weight = [85, 95, 110, 125, None][(i // 5) % 5]
    51. 

  51.     obs = dict(COMMON, id=f'SIM-OBS-{i+1:04d}', groupId=p['groupId'], projectId=p['id'],
    52. 

  52.         title=f'模拟系列{(i//30)+1:02d} · {flavors[(i//5)%6]} · {packaging[i%5]}',
    53. 

  53.         productName=f'模拟系列{(i//30)+1:02d}', flavor=flavors[(i//5)%6], packaging=packaging[i%5],
    54. 

  54.         netWeightG=weight, packCount=pack_count, preparation=['热水复水候选', '需煮制候选', '条件待核'][i % 3],
    55. 

  55.         nutritionClaim=None, actualSku=None, note='虚构商品观察，未关联真实在售商品；营养与工艺未知。')
    56. 

  56.     ext['observations'].append(obs)
    57. 

  57.     for j, day in enumerate(['2026-10-12', '2026-10-19', '2026-10-26', '2026-11-02']):
    58. 

  58.         count = pack_count
    59. 

  59.         paid = round((4.5 + (i % 17) * .37 + j * .25) * (count or 1), 2)
    60. 

  60.         comparable = bool(count and weight and j != 2)
    61. 

  61.         ext['prices'].append(dict(COMMON, id=f'SIM-PRICE-{i+1:04d}-{j+1}', observationId=obs['id'], projectId=p['id'], groupId=p['groupId'],
    62. 

  62.             date=day, channel=['模拟电商自营', '模拟商超', '模拟促销组合', '模拟便利店'][j], currency='CNY',
    63. 

  63.             pricePaid=paid, includesShipping=j == 0, packCount=count, netWeightG=weight,
    64. 

  64.             unitPricePer100G=round(paid / (count * weight) * 100, 2) if comparable else None,
    65. 

  65.             comparisonState='可按演示规格比较' if comparable else '口径待核，不参与单位价比较'))
    66. 

  66. 

  67. for i in range(354):
    68. 

  68.     source = ext['voices'][(i * 31) % len(ext['voices'])]
    69. 

  69.     quote = source['rawText'].split('；')[0] + '；'
    70. 

  70.     ext['citations'].append(dict(COMMON, id=f'SIM-EVIDENCE-{i+1:03d}', groupId=source['groupId'], projectId=source['projectId'],
    71. 

  71.         sourceId=source['id'], quote=quote, locator=dict(start=0, end=len(quote)), reviewState='demo_only',
    72. 

  72.         stance=['support', 'contradict', 'unresolved'][i % 3], claim='用于演示不同场景与其他解释，并非已验证用户需求。'))
    73. 

  73. 

  74. for i in range(180):
    75. 

  75.     p = seed['projects'][i % 12]
    76. 

  76.     state = ['done', 'in_progress', 'pending', 'failed', 'reused'][i % 5]
    77. 

  77.     ext['researchTasks'].append(dict(COMMON, id=f'SIM-RESEARCH-{i+1:03d}', groupId=p['groupId'], projectId=p['id'],
    78. 

  78.         ownerMemberId=p['memberIds'][(i//12) % 6], scope=['核对面体制备说明', '区分正餐与加餐', '检查商品组合规格', '核对相反材料', '追问搭配原因'][i % 5],
    79. 

  79.         state=state, requiredMaterial=['包装制备原文', '带情境的追问记录', '规格与价格截图'][i % 3],
    80. 

  80.         failureReason='演示来源不可用，保留失败并待补其他来源' if state == 'failed' else None,
    81. 

  81.         note='模拟任务回放；没有执行真实采集。'))
    82. 

  82. 

  83. for i in range(312):
    84. 

  84.     p = seed['projects'][i % 12]
    85. 

  85.     owner = seed['members'][(i * 7) % 36]
    86. 

  86.     body = ['这条声音是否可能来自加餐？建议先问清楚当餐目的。', '已在模拟资料中补上制备条件，原方案版本保留。',
    87. 

  87.             '支持继续研究，但不能把搭配习惯直接当成份量不足。', '价格缺少包数，请先核规格，暂不参与单位价比较。',
    88. 

  88.             '工艺条件与目标场景冲突，建议保留暂停理由并分配补证任务。', '课堂方案可展示，实际成本和营养参数仍待核实。'][i % 6]
    89. 

  89.     ext['collaborationRecords'].append(dict(COMMON, id=f'SIM-COLLAB-{i+1:03d}', groupId=p['groupId'], projectId=p['id'],
    90. 

  90.         ownerMemberId=owner['id'], user=owner['displayName'], kind=['质疑', '答复', '补证', '版本说明', '取舍', '复核'][i % 6], body=body,
    91. 

  91.         at=f'2026-11-0{1+i%3}T{9+(i%8):02d}:{i%60:02d}:00+08:00', note='模拟协作记录'))
    92. 

  92. 

  93. counts = {key: len(ext[key]) for key in ['voices', 'observations', 'prices', 'citations', 'researchTasks', 'collaborationRecords']}
    94. 

  94. assert len({v['rawText'] for v in ext['voices']}) == 12000
    95. 

  95. assert sum(counts.values()) == 16446
    96. 

  96. ext['counts'] = counts
    97. 

  97. target = ROOT / 'src/assets/showcase/demo-extension.json'
    98. 

  98. target.write_text(json.dumps(ext, ensure_ascii=False, separators=(',', ':')) + '\n', encoding='utf-8')
    99. 

  99. print(json.dumps(dict(file=str(target), bytes=target.stat().st_size, counts=counts, uniqueVoices=12000,
    100. 

  100.                      totalEvidence=len(seed['evidence'])+counts['citations'], sha256=hashlib.sha256(target.read_bytes()).hexdigest()), ensure_ascii=False))
    101.