Improve parser

[hadrianl]

resolved #204

This PR improve the performance of taos query[c -> python parser] :

• reduce cast, memory copy and object creation
• convert timezone once when set_tz, not everytime _convert_XXX_to_datetime
• totally, it take only 10%-20% cpu time when dealing with binary, nchar and timestamp

further more

• taos_is_null had been called so many time per query
• a vector version of taos_is_null from libtaos is needed to improve this, just like taos_get_column_data_offset
• or maybe, introducing the third party libraries such as Cython to improve this

[codecov-commenter]

Codecov Report

Merging #209 (56b8ca1) into main (9dad4b5) will decrease coverage by 0.04%. The diff coverage is 75.00%.

@@            Coverage Diff             @@
##             main     #209      +/-   ##
==========================================
- Coverage   81.37%   81.33%   -0.04%     
==========================================
  Files          20       20              
  Lines        2652     2652              
==========================================
- Hits         2158     2157       -1     
- Misses        494      495       +1     

Files Changed	Coverage Δ
taos/field.py	79.75% <25.00%> (-0.37%)	:arrow_down:
taos/field_v3.py	93.75% <100.00%> (-0.16%)	:arrow_down:

:mega: We’re building smart automated test selection to slash your CI/CD build times. Learn more