cnki-advanced-search

CNKI Advanced Search (高级检索)

Perform a filtered search on CNKI using the old-style advanced search interface (only interface with source category checkboxes).

Arguments

$ARGUMENTS describes the search criteria in natural language. Parse it to identify:

• Subject keywords (Subject) — default field
• Title keywords (Title)
• Keywords (Keywords)
• Author name (Author) — separate field #au_1_value1
• Journal/source (Source) — separate field #magazine_value1
• Date range (Time Range) — #startYear / #endYear
• Source category (Source Category: SCI, EI, PKU Core, CSSCI, CSCD)

Steps

1. Navigate

Use mcp__chrome-devtools__navigate_page → https://kns.cnki.net/kns/AdvSearch?classid=7NS01R8M

2. Search + get results (single async evaluate_script)

Replace placeholder values with actual search criteria:

async () => {
  // --- Config: fill in actual values ---
  const query = "KEYWORDS";          // row 1 search keywords
  const fieldType = "SU";           // SU=主题, TI=篇名, KY=关键词, TKA=篇关摘, AB=摘要
  const query2 = "";                // row 2 keywords (optional, "" to skip)
  const fieldType2 = "KY";          // row 2 field type
  const rowLogic = "AND";           // AND=并且, OR=或者, NOT=不含 (between row 1 and 2)
  const sourceTypes = ["CSSCI"];    // any of: "SCI", "EI", "hx", "CSSCI", "CSCD" ([] = all)
  const startYear = "";             // e.g. "2020" or "" for no limit
  const endYear = "";               // e.g. "2025" or "" for no limit
  const author = "";                // author name or ""
  const journal = "";               // journal name or ""

  // --- Wait for form ---
  await new Promise((r, j) => {
    let n = 0;
    const c = () => { if (document.querySelector('#txt_1_value1')) r(); else if (n++ > 30) j('timeout'); else setTimeout(c, 500); };
    c();
  });

  // Captcha check
  const cap = document.querySelector('#tcaptcha_transform_dy');
  if (cap && cap.getBoundingClientRect().top >= 0) return { error: 'captcha' };

  const selects = Array.from(document.querySelectorAll('select')).filter(s => s.offsetParent !== null);

  // --- Source type: uncheck 全部, check targets ---
  if (sourceTypes.length > 0) {
    const gjAll = document.querySelector('#gjAll');
    if (gjAll && gjAll.checked) gjAll.click();
    for (const st of sourceTypes) {
      const cb = document.querySelector('#' + st);
      if (cb && !cb.checked) cb.click();
    }
  }

  // --- Row 1: field type + keyword ---
  selects[0].value = fieldType;
  selects[0].dispatchEvent(new Event('change', { bubbles: true }));
  const input = document.querySelector('#txt_1_value1');
  input.value = query;
  input.dispatchEvent(new Event('input', { bubbles: true }));

  // --- Row 2: field type + keyword (optional) ---
  if (query2) {
    selects[5].value = rowLogic; // row logic: AND/OR/NOT
    selects[5].dispatchEvent(new Event('change', { bubbles: true }));
    selects[6].value = fieldType2;
    selects[6].dispatchEvent(new Event('change', { bubbles: true }));
    const input2 = document.querySelector('#txt_2_value1');
    input2.value = query2;
    input2.dispatchEvent(new Event('input', { bubbles: true }));
  }

  // --- Author (optional) ---
  if (author) {
    const auInput = document.querySelector('#au_1_value1');
    if (auInput) { auInput.value = author; auInput.dispatchEvent(new Event('input', { bubbles: true })); }
  }

  // --- Journal (optional) ---
  if (journal) {
    const magInput = document.querySelector('#magazine_value1');
    if (magInput) { magInput.value = journal; magInput.dispatchEvent(new Event('input', { bubbles: true })); }
  }

  // --- Date range (optional) ---
  if (startYear) { selects[14].value = startYear; selects[14].dispatchEvent(new Event('change', { bubbles: true })); }
  if (endYear) { selects[15].value = endYear; selects[15].dispatchEvent(new Event('change', { bubbles: true })); }

  // --- Submit ---
  document.querySelector('div.search')?.click();

  // Wait for results
  await new Promise((r, j) => {
    let n = 0;
    const c = () => {
      if (document.body.innerText.includes('条结果')) r();
      else if (n++ > 40) j('timeout');
      else setTimeout(c, 500);
    };
    setTimeout(c, 2000);
  });

  // Captcha check again
  const cap2 = document.querySelector('#tcaptcha_transform_dy');
  if (cap2 && cap2.getBoundingClientRect().top >= 0) return { error: 'captcha' };

  return {
    query, fieldType, query2, fieldType2, rowLogic,
    sourceTypes, startYear, endYear, author, journal,
    total: document.querySelector('.pagerTitleCell')?.innerText?.match(/([\d,]+)/)?.[1] || '0',
    page: document.querySelector('.countPageMark')?.innerText || '1/1',
    url: location.href
  };
}

3. Report

Advanced search: "{query}" ({fieldType}) + source: {sourceTypes} → {total} results.

Tool calls: 2 (navigate + evaluate_script)

Verified selectors (old-style interface)

Form fields

Element	Selector / Select index	Notes
Row 1 Field Type	selects[0]	SU=Subject, TI=Title, KY=Keywords, TKA=Title/Keyword/Abstract, AB=Abstract
Row 1 Keyword	#txt_1_value1	main keyword input
Row 1 Second Term in Row	#txt_1_value2	same-row AND/OR/NOT with first keyword
Row 1 In-Row Logic	selects[2]	AND=contains, OR=contains any, NOT=excludes
Inter-Row Logic	selects[5]	AND=and, OR=or, NOT=excludes
Row 2 Field Type	selects[6]	same options as row 1
Row 2 Keyword	#txt_2_value1	second row keyword
Row 2 Second Term in Row	#txt_2_value2
Author	#au_1_value1	placeholder "Chinese name/English name/pinyin"
Author Affiliation	#au_1_value2	placeholder "full name/abbreviation/former name"
Source	#magazine_value1	placeholder "journal name/ISSN/CN"
Fund	#base_value1
Start Year	selects[14] / #startYear	<select> 1915-2026
End Year	selects[15] / #endYear	<select> 2026-1915
Search Button	div.search	NOT input/button

Source type checkboxes (来源类型)

Source	Checkbox ID	Notes
All Journals	#gjAll	Checked by default; uncheck before selecting others
SCI Source Journals	#SCI	value="Y"
EI Source Journals	#EI	value="Y"
PKU Core Journals	#hx	value="Y"
CSSCI	#CSSCI	value="Y"
CSCD	#CSCD	value="Y"

Multiple source types can be checked simultaneously (OR logic).

Results

Element	Selector	Notes
Result count	.pagerTitleCell	text "Found X results in total"
Page indicator	.countPageMark	text "1/300"

Captcha detection

Check #tcaptcha_transform_dy element's getBoundingClientRect().top >= 0.

Important Notes

• Must use old-style URL (kns.cnki.net/kns/AdvSearch). New interface (kns8s/AdvSearch) has NO source category checkboxes.
• The classid=7NS01R8M parameter ensures the correct form layout loads.
• Results page is compatible with cnki-parse-results and cnki-navigate-pages skills.