"""Declarative mapping from arbitrary tabular source data to SiteForge content. Never eval mappings, no AI hallucinated fields, preview shows all errors. """ import json,re,hashlib from . import db from .core import route_path,slug_of,save_entry ALIASES={ 'title':('title','name','product_name','model_name','名称','标题','产品名称','车型','车型名称','商品名称'), 'route':('route','url','path','permalink','链接','路径','旧站链接','原网址'), 'summary':('summary','description','excerpt','简介','摘要','产品简介'), 'body':('body','content','html','正文','内容','详细介绍'), 'cover':('cover','image','thumbnail','featured_image','图片','封面','主图'), 'brand':('brand','manufacturer','品牌','厂商'), 'category':('category','分类','类目'), 'seo_title':('seo_title','meta_title','SEO标题','seo标题'), 'seo_description':('seo_description','meta_description','SEO描述','seo描述'), 'canonical':('canonical','规范网址'), 'model':('model','model_name','车型型号','型号','规格型号'), 'energy':('energy','energy_type','fuel_type','动力形式','能源类型','燃料类型'), 'range':('range','battery_range','纯电续航','续航','续航里程'), 'year':('year','model_year','年款','年份'), 'sku':('sku','product_sku','产品编号','货号'), 'gallery':('gallery','images','image_gallery','产品图片','图库','详情图片'), 'price':('price','msrp','售价','价格'), 'specs':('specs','attributes','parameters','技术参数','产品参数'), } SYSTEM_FIELDS={'title','route','summary','body','cover','seo_title','seo_description','canonical'} def setup(): with db.connect() as c: c.executescript('''CREATE TABLE IF NOT EXISTS mapping_profiles ( id TEXT PRIMARY KEY,site_id TEXT NOT NULL REFERENCES sites(id) ON DELETE CASCADE, name TEXT NOT NULL,content_type TEXT NOT NULL, rules_json TEXT NOT NULL, created_at TEXT NOT NULL,updated_at TEXT NOT NULL); CREATE TABLE IF NOT EXISTS import_batches ( id TEXT PRIMARY KEY,site_id TEXT NOT NULL REFERENCES sites(id) ON DELETE CASCADE, profile_id TEXT, source TEXT NOT NULL, result_json TEXT NOT NULL, created_at TEXT NOT NULL); ''') def suggest(columns): clean=lambda x:re.sub(r'[\s\-_]+','',str(x).casefold()) src={clean(col):col for col in columns} mapping={} for target,aliases in ALIASES.items(): for alias in aliases: if clean(alias) in src: mapping[target]=src[clean(alias)];break return mapping def source_column(spec): return spec if isinstance(spec,str) else spec['source'] def validate_rules(rules,columns): if not isinstance(rules,dict) or len(rules)>100:raise ValueError('字段映射规则不合法') for dest,spec in rules.items(): if not isinstance(dest,str) or not re.fullmatch(r'[\w\-]{1,60}',dest,re.UNICODE): raise ValueError('目标字段名称不合法') if isinstance(spec,dict): allowed={'source','trim','case','default','prefix','suffix'} if set(spec)-allowed or 'source' not in spec:raise ValueError('包含不支持的转换配置') if type(spec.get('trim',True)) is not bool:raise ValueError('trim 必须为布尔值') if spec.get('case','none') not in ('none','upper','lower'):raise ValueError('不支持的大小写处理') for key in ('default','prefix','suffix'): if key in spec and (not isinstance(spec[key],str) or len(spec[key])>500): raise ValueError('转换字符串过长或格式非法') elif not isinstance(spec,str):raise ValueError('映射字段格式非法') src=source_column(spec) if not isinstance(src,str) or src not in columns:raise ValueError(f'来源字段不存在: {src}') if 'title' not in rules:raise ValueError('必须映射标题字段') def mapped_value(row,spec): raw=row.get(source_column(spec)) value=json.dumps(raw,ensure_ascii=False) if isinstance(raw,(dict,list)) else str(raw if raw is not None else '') if isinstance(spec,str):return value.strip() if spec.get('trim',True):value=value.strip() if not value:value=spec.get('default','') case=spec.get('case','none') if case=='upper':value=value.upper() elif case=='lower':value=value.lower() return spec.get('prefix','')+value+spec.get('suffix','') def transform(rows,rules,kind='product',prefix=None): if kind not in ('product','post','page','brand','vehicle','service','case','other'):raise ValueError('不支持的内容类型') columns=set().union(*(r.keys() for r in rows)) if rows else set() validate_rules(rules,columns) seen=set();out=[];errors=[] for index,row in enumerate(rows): vals={} for target,spec in rules.items(): vals[target]=mapped_value(row,spec) title=vals.get('title','') if not title: errors.append({'row':index+1,'error':'标题为空'});continue route=vals.get('route','') if not route: route=f'/{prefix or kind}/{slug_of(title)}/' try:route=route_path(route) except ValueError as exc: errors.append({'row':index+1,'error':f'路径非法：{exc}'});continue if route in seen: errors.append({'row':index+1,'error':'导入文件内 URL 重复：'+route});continue seen.add(route) fields={k:v for k,v in vals.items() if k not in SYSTEM_FIELDS} if fields.get('gallery'): try: values=json.loads(fields['gallery']) if isinstance(values,list):fields['gallery']=[str(x.get('url','') if isinstance(x,dict) else x) for x in values if x] except ValueError: fields['gallery']=[x.strip() for x in fields['gallery'].split(';') if x.strip()][:40] out.append({'type':kind,'title':title,'route':route,'summary':vals.get('summary',''), 'body':vals.get('body',''),'cover':vals.get('cover',''), 'seo_title':vals.get('seo_title',''),'seo_description':vals.get('seo_description',''), 'canonical':vals.get('canonical',''),'meta':{'fields':fields}}) return out,errors def _state_token(site_id,items): """Preview token binds mapped payload and existing rows' revision metadata. This is advisory optimistic checking, not a global distributed lock. """ routes=[x['route'] for x in items] existing={r['route']:r for r in db.rows('SELECT * FROM entries WHERE site_id=?',(site_id,)) if r['route'] in routes} state={'site':site_id,'items':items,'existing':existing} return hashlib.sha256(json.dumps(state,ensure_ascii=False,sort_keys=True, separators=(',',':')).encode('utf-8')).hexdigest() def dry_run(site_id,rows,rules,kind='product',prefix=None): if len(rows)>1000:raise ValueError('单批最多1000条内容') items,errors=transform(rows,rules,kind,prefix) existing={r['route'] for r in db.rows('SELECT route FROM entries WHERE site_id=?',(site_id,))} collisions=[r['route'] for r in items if r['route'] in existing] return {'input':len(rows),'mapped':len(items),'errors':errors,'conflicts':collisions, 'unmapped_columns':sorted(set().union(*(x.keys() for x in rows)) - {source_column(v) for v in rules.values()}) if rows else [], 'sample':items[:10], 'items':items, 'plan_hash':_state_token(site_id,items), 'supported_transforms':['trim','default','upper','lower','prefix','suffix'], 'note':'提交时携带 expected_plan_hash 可阻止预检后发生变化的数据覆盖'} def commit(site_id,rows,rules,kind='product',on_conflict='skip',prefix=None,expected_plan_hash=None): if on_conflict not in ('skip','replace','abort'): raise ValueError('冲突策略只能为 skip、replace 或 abort') report=dry_run(site_id,rows,rules,kind,prefix) if expected_plan_hash is not None: if not isinstance(expected_plan_hash,str) or not re.fullmatch(r'[a-f0-9]{64}',expected_plan_hash): raise ValueError('预检令牌格式无效') if expected_plan_hash!=report['plan_hash']: raise ValueError('导入预检已过期：来源数据或已有内容发生变化，请重新预览') if on_conflict=='abort' and (report['conflicts'] or report['errors']): raise ValueError('严格导入被阻断：存在 URL 冲突或无效数据，请处理后重新预览') created=0;updated=0;skipped=0;errors=list(report['errors']) for idx,item in enumerate(report['items'],1): try: existing=db.one('SELECT id FROM entries WHERE site_id=? AND route=?',(site_id,item['route'])) if existing and on_conflict=='skip':skipped+=1;continue # abort semantics require a preflight without errors; do not turn # this into an automatic 'replace' policy on a colliding route. if existing and on_conflict=='abort':raise ValueError('提交期间出现 URL 冲突') save_entry(site_id,item,existing_id=existing['id'] if existing else None,source='mapped-import') if existing:updated+=1 else:created+=1 except Exception as exc: errors.append({'row':idx,'error':str(exc)[:250]}) if on_conflict=='abort':break summary={'created':created,'updated':updated,'skipped':skipped,'errors':errors,'mapped':report['mapped'], 'plan_hash':report['plan_hash']} from .core import uid db.insert('INSERT INTO import_batches VALUES (?,?,?,?,?,?)',(uid(),site_id,'',f'content-map:{kind}',json.dumps(summary,ensure_ascii=False),db.now())) return summary