search

taoyds / spider

scripts and baselines for Spider: Yale complex and cross-domain semantic parsing and text-to-SQL challenge
https://yale-lily.github.io/spider
Apache License 2.0
848 stars 193 forks source link

Evaluation script can't be used on train_gold labels #52

Open CrafterKolyan opened 4 years ago

CrafterKolyan commented 4 years ago

Command:

../evaluation.py --gold train_gold.sql --pred train_gold.sql --db database --table tables.json --etype all

Output:

Traceback (most recent call last):
  File "../evaluation.py", line 866, in <module>
    evaluate(gold, pred, db_dir, etype, kmaps)
  File "../evaluation.py", line 505, in evaluate
    g_sql = get_sql(schema, g_str)
  File "/home/nikolai.korolev/workspace/spider/process_sql.py", line 552, in get_sql
    _, sql = parse_sql(tokens, 0, tables_with_alias, schema)
  File "/home/nikolai.korolev/workspace/spider/process_sql.py", line 504, in parse_sql
    from_end_idx, table_units, conds, default_tables = parse_from(tokens, start_idx, tables_with_alias, schema)
  File "/home/nikolai.korolev/workspace/spider/process_sql.py", line 389, in parse_from
    idx, table_unit, table_name = parse_table_unit(tokens, idx, tables_with_alias, schema)
  File "/home/nikolai.korolev/workspace/spider/process_sql.py", line 259, in parse_table_unit
    key = tables_with_alias[tokens[idx]]
KeyError: 'ref_company_types'