| file,field,type,description
|
| poem_catalog.csv,poem_id,string,规范诗作ID,GS-P###。
|
| poem_catalog.csv,title,string,公开辑录中的题名;无题作品用disambiguator区分。
|
| poem_catalog.csv,disambiguator,string,仅保留4个可见字符的区分词,不代替全文。
|
| poem_catalog.csv,form_label_as_compiled,string,辑录者所标文体,不代表严格格律鉴定。
|
| poem_catalog.csv,tier,category,核心六篇、扩展合集或拾遗。
|
| poem_catalog.csv,sha256_nfc,hex,规范文本做Unicode NFC后计算的SHA-256。
|
| edition_catalog.csv,edition_id,string,来源见证ID,GS-E###。
|
| edition_catalog.csv,source_locator,string,来源页行号或索引定位。
|
| edition_catalog.csv,difference_summary,string,该见证相对于规范文本的差异摘要。
|
| fragment_catalog.csv,fragment_id,string,散句ID,GS-F###。 |
| materialized/poems_full.csv,text,string,17篇规范诗作的完整正文;换行保留版本编排。 |
| materialized/poems_full.csv,hash_match_catalog,boolean,正文Unicode NFC SHA-256是否与诗作目录一致。 |
| materialized/editions_full.csv,text,string,24个来源版本见证的完整正文。 |
| materialized/editions_full.csv,hash_match_catalog,boolean,正文Unicode NFC SHA-256是否与版本目录一致。 |
| materialized/fragments_full.csv,text,string,4则散句的完整正文。 |
| materialized/fragments_full.csv,hash_match_catalog,boolean,正文Unicode NFC SHA-256是否与散句目录一致。 |
| gongyu_index.csv,record_id,string,原始语料条目ID;RC/RN/RT/RL分别对应四个来源分区。 |
| gongyu_index.csv,media_type,category,text或image。
|
| gongyu_index.csv,asset_name,string,图片占位的源文件名;文本行为空。
|
| gongyu_index.csv,sha256_nfc,hex,未脱敏文本的NFC SHA-256;数据包不含该文本。
|
| gongyu_index.csv,exact_duplicate_group,string,完全相同文本的重复组ID。
|
| gongyu_index.csv,duplicate_of,string,若为重复项,指向同组首条记录。
|
| gongyu_index.csv,candidate_poem_ids,pipe-delimited,按去标点子串匹配得到的可能成诗关系;只是候选,不是作者学结论。
|
| gongyu_index.csv,is_truncated_in_source,boolean,来源文本自身是否以省略号截断。
|
| gongyu_index.csv,possible_identifier,boolean,是否含6位以上连续数字;采集脚本默认替换为<ID>。
|
| historical_reverted_index.csv,revision_status,string,历史快照中的条目状态;这些条目已被回退。
|
| variant_catalog.csv,confidence,category,高/中/低:版本差异判断的证据强度。
|
| source_catalog.csv,evidence_grade,category,B=结构清楚/可定位;C=社区辑录;D=仅线索或正文不可得。
|
| source_gaps.csv,resolution_status,string,缺口的当前处理决定。
|
| manifest.json,sha256,hex,数据包内文件校验和。
|
|
|