[Bug]: sds-convert does not properly handle multiple records per data block

[MiloradCvjetkovic]

Is there an existing issue for this?

• I have searched the existing issues

Problem statement

If the data block contains multiple records like in object detection case (ML_Out.0.sds file in attached Files.zip file) and if sample-frequency is not specified in sds.yml (ML_Out.sds.yml in attached Files.zip file) then conversion to CSV with the following command:

sds-convert csv -i ML_Out.0.sds -o ML_Out.0 -y ML_Out.sds.yml

results in a ML_Out.0.csv file in which each object is assigned a different timeslot. However, the objects should instead be grouped into sets of 10, with all 10 objects in each group sharing the same timeslot.

Files.zip

Steps To Reproduce

No response