fix: Check all values for trie.predictive_search

[congqixia]

Related to #35941

For marisa trie predictive_search default behavior, it value iterated is not in lexicographic order.

This PR is a brute force fix to make range operator returns correct values.

[sre-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: congqixia

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• internal/core/OWNERS [congqixia]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[longjiquan]

LGTM.
/cc @zhagnlu

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 72.64%. Comparing base (9da8652) to head (5c5a378).
Report is 14 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master   #35943      +/-   ##
==========================================
- Coverage   81.55%   72.64%   -8.92%     
==========================================
  Files        1264     1264              
  Lines      150768   150762       -6     
==========================================
- Hits       122964   109519   -13445     
- Misses      22907    36349   +13442     
+ Partials     4897     4894       -3     

Files with missing lines	Coverage Δ
internal/core/src/index/StringIndexMarisa.cpp	0.00% <ø> (-64.73%)	⬇️

... and 237 files with indirect coverage changes

[xiaofan-luan]

1. the current fix looks good for me.

2. when we build index, we can sort the keys, like

3. // Insert keys in sorted order
   

   std::vectorstd::string words = {
   "apple", "application", "banana", "bath", "cat", "dog",
   "elephant", "fish", "giraffe", "hippopotamus"
   };
   std::sort(words.begin(), words.end());

   for (const auto& word : words) {
   keyset.push_back(word.c_str());
   }

   marisa::Trie trie;
   trie.build(keyset);

then the range query can be done by

`
std::vectorstd::string rangeQuery(const marisa::Trie& trie, const std::string& start, const std::string& end) {
std::vectorstd::string results;
marisa::Agent agent;

// Start from the first key that is greater than or equal to 'start'
agent.set_query(start.c_str());
if (trie.predictive_search(agent)) {
    do {
        std::string key(agent.key().ptr(), agent.key().length());
        
        // Stop if we've gone past 'end'
        if (key > end) {
            break;
        }

        results.push_back(key);
    } while (trie.predictive_search(agent));
}

return results;

}`

[xiaofan-luan]

of cause we need a meta to mark whether the dataset is sorted or not.
if not we fallback

[zhagnlu]

/lgtm