Cleanup manually/automatic generated class in pdata related to trace

[bogdandrutu]

This is a preparation for pdata (TraceID, SpanID, AttributeValue, AttributeMap) to be marked as stable API.

• Cleanup translator code to work only with pdata types (TraceID, SpanID)
• Should we consider to move translator code as helper to pdata (NewTraceIDFromLong, NewSpanIDFromLong, etc.). Translator code is in translator/trace/big_endian_converter?
• Review public API for TraceID, SpanID - Punya
• Review public API for AttributeValue and AttributeMap.
  • One thing that we should consider is to rename ForEach with Range to be consistent with sync.Map.
  • Remove deprecated NewAttributeValue, needs changes in pdatagen which is the only user of this.
  • Remove InitEmptyWithCapacity
  • Add EnsureCapacity (replacement for InitEmptyWithCapacity) that does not remove existing elements.
• Review comments for all public API, maybe we have things to improve. - Punya
• Review the Resize api for slices.
  • MAY consider to replace the API with EnsureCapacity + AppendEmpty usage. Or AppendEmptyN(int).
• Remove deprecated Resize - Anthony
• Remove deprecated Append api for slices.
  • Cleanup core usages of pdata.Append.
  • Cleanup contrib usages of pdata.Append.
• Consider to change StartTime/EndTime to StartTimestamp/EndTimestamp