Prepare release 2.9.0

[veenstrajelmer]

Todo:

• update sphinx-docs.yml #312
• metadata comparison too strict for hatyan.read_dia() #313
• fix PerformanceWarning about fragmented dataframe (pytest python 3.9 and 3.12) #272
• Defragment v0uf dataframes by making copy #325
• fix numpy and pandas deprecationwarning #319
• convert "HW numbers not always increasing" to ValueError #327
• astrog time-related improvements #332
• cannot make prediction of dummy component set without tzone attr #317
• Unexpected timezones in case of different timezones in components and times #334
• Align qc and status columns of timeseries dataframes #308
• add warning for ignored qualitycode column in hatyan.write_dia() #344
• Avoid duplicate code in write_dia functions #342
• astrog datetime index instead of column #346
• refactor timezone handling in prediction #348
• clean up hatyan.calc_HWLW12345to12() #350
• consider using HWLWno in hatyan.calc_HWLW12345to12() #311
• Abbreviate status in hatyan.ddlpy_to_hatyan() #352
• make write_dia less hardcoded #145
• kenmerkendewaarden is awaiting this release: update to astrog datetime indexes Deltares-research/kenmerkendewaarden#135

After release:

• check zenodo (should be fixed for releases after 14-5-2024)

Changes with user impact:

• astrog output dataframes now have DatetimeIndex instead of 'datetime' column (346 astrog datetime index instead of column #347)

Follow-up: #356