fix(cn_index): wrap literal HTML in StringIO for pd.read_html

[he-yufeng]

Description

#2047 wrapped the literal HTML passed to pd.read_html in a StringIO for the us_index collector to silence pandas' FutureWarning:

Passing literal html to 'read_html' is deprecated and will be removed in a future version. To read from a literal string, wrap it in a 'StringIO' object.

The parallel cn_index collector was missed. _parse_table passes the csindex response body (an HTML string, not a URL/path) straight to pd.read_html(content) at scripts/data_collector/cn_index/collector.py:185, so it still hits the same deprecation and will break once pandas drops literal-string support.

Change

Import StringIO and wrap the content, matching what #2047 did for us_index. No behavior change on supported pandas versions; StringIO has always been accepted by read_html.

Reproduced locally on pandas 2.3.3:

>>> pd.read_html("<table>...</table>")        # FutureWarning: Passing literal html ...
>>> pd.read_html(StringIO("<table>...</table>"))  # no warning

Note

I scoped this to the read_html fix to stay a clean parallel of #2047. The same file also calls DataFrame.applymap (line 174), which pandas 2.1 deprecates in favour of DataFrame.map, but DataFrame.map only exists from 2.1 onward while setup.py still supports pandas>=1.1, so swapping it would need version handling and is left out here.