Yuanbao Gaokao Score and Rank Lookup

Problem

College entrance exam candidates often need to map a score to a provincial rank, or infer an approximate score band from a rank, but one-score-one-segment data depends on province, year, and subject-track rules. Raw natural-language queries can cause mismatched province aliases, incorrect year parsing, inconsistent old/new exam subject mapping, and missing score or rank fields. This skill turns that workflow into explicit field extraction and a controlled lookup step.

How It Works

The skill extracts five fields in order: place, year, classify, score, and rank:
- Region: normalizes province names, supports aliases, common abbreviations, city-to-province mapping, and multi-province queries separated by spaces.
- Year: handles explicit years and vague expressions such as “this year” or “last year”; multiple years are space-separated.
- Subject track: uses region and year to distinguish traditional arts/sciences, 3+3, or 3+1+2 modes, then maps terms like “arts”, “science”, “physics”, and “history” to the standard category.
- Score and rank: accepts only numeric score and rank values; Chinese numerals are converted to integers.
After the key conditions are confirmed, it calls ./fetch_data.py to retrieve the relevant segment data, filters the result by the user's intent, and returns a concise answer. If score or rank is missing, it asks for the missing value first.

Boundaries

It is intended for score-to-rank or rank-to-score lookups in college entrance exam segment tables, not full admission counseling or enrollment-probability prediction. Subject-track interpretation changes across provinces and years, especially under new exam reform rules, so place, year, and classify must be read together. The document maps “this year” and “last year” to 2026 and 2025; in production use, align with the current exam year and local reporting rules.