How does Apache ECharts handle bigint numbers when constructing a 'value' type X-axis?

I'm using Apache ECharts to visualize data with very large integer values (JavaScript bigint type) on the X-axis, configured as type: 'value'. However, I'm encountering unexpected behavior:

My data looks like this:

data: [
    { value: [12345678901234567890n, 10] },  // bigint X-value
    { value: [12345678901234567891n, 20] },
    // ...
]

Issue:

• The X-axis labels either round the numbers (e.g., showing 1.2345678901234568e+19),
• Or fail to display the correct precision (last digits are truncated).

Expected Behavior:

Display the full precision of bigint values as readable labels (e.g., "12345678901234567890").

Attempted Solutions:

Tried setting axisLabel: { formatter: (value) => value.toString() }, but the original value is already rounded.

Question:

How can I configure ECharts to correctly handle and display bigint-scale integers on a 'value'-type X-axis without losing precision? Are there workarounds for axis formatting or data representation?

Echarts doesn't seem to support BigInt, this request for this feature is still pending.

A search for BigInt in the source code doesn't produce any results.

As for what's happening in the current version, all data points are passed through the function parseDataValue, (source code - dataValueHelper#L34) that applies Number constructor (called as function) to each numeric data value (see dataValueHelper#L80 in the same function), which casts the BigInts to number values, losing the precision, which appears in some cases as the values being "rounded" to a multiple of 10^k, integer k > 1 (e.g., Number(12345678901234567890n) === 12345678901234567000).

Also, if BigNums are given as values for some options, like if one sets the min or max of an axis to a BigInt, they will be passed (see scaleRawExtentInfo.ts#L205) through the global function isFinite, that doesn't accept BigInts as arguments, throwing the error Cannot convert a BigInt value to a number.

A solution to avoid using BigNums can, of course, be improvised by anyone, tailored to the needs of their specific dataset; but since I put this as an answer rather than a comment, let me add a suggestion in this regard. One could choose any value of the BigInts, call it _bigNumZero, and extract it from each of the other values and convert that value to a number, presuming that the range of the BigNums is small enough to be covered without loss of precision by values of type number:

let _bigNumZero = null;
function resetBigNum(data){
   if(_bigNumZero === null){ // just for the first dataset
      _bigNumZero = data[0][0]; // the x value of the first point
   }
   for(const point of data){
      point[0] = Number(point[0] - _bigNumZero); // 0 for x dimension
   }
   return data;
} 

used in the chart configuration as:

series: [
   {
      // ... other series options
      data: resetBigNum(.....)
   },
   // ... other series, similar treatment
]

the argument of resetBigNum being a series' data as an array of [x, y] arrays, where x is always a BigInt.

Then, formatters should be added everywhere the x value might be displayed, to add to the numeric value used by the chart, the "zero point" _bigNumZero. In the following echarts demo, I set formatters for the x axis and the tooltip.