@@ -1997,23 +1997,23 @@ def publish_batch_data(
19971997 config = batch_data , df = batch_df , column_name = column
19981998 )
19991999 # Get min and max timestamps
2000- min_timestamp = batch_df [batch_data ["timestampColumnName" ]].min ()
2001- max_timestamp = batch_df [batch_data ["timestampColumnName" ]].max ()
2000+ earliest_timestamp = batch_df [batch_data ["timestampColumnName" ]].min ()
2001+ latest_timestamp = batch_df [batch_data ["timestampColumnName" ]].max ()
20022002
20032003 with tempfile .TemporaryDirectory () as tmp_dir :
2004- # Copy relevant files to tmp dir
2005- utils .write_yaml (batch_data , f"{ tmp_dir } /dataset_config.yaml" )
2004+ # Copy save files to tmp dir
20062005 batch_df .to_csv (f"{ tmp_dir } /dataset.csv" , index = False )
2007-
2008- tar_file_path = os .path .join (tmp_dir , "tarfile" )
2009- with tarfile .open (tar_file_path , mode = "w:gz" ) as tar :
2010- tar .add (tmp_dir , arcname = os .path .basename ("reference_dataset" ))
2006+ payload = {
2007+ "earliestTimestamp" : earliest_timestamp ,
2008+ "latestTimestamp" : latest_timestamp ,
2009+ ** batch_data ,
2010+ }
20112011
20122012 self .api .upload (
20132013 endpoint = f"inference-pipelines/{ inference_pipeline_id } /data" ,
2014- file_path = tar_file_path ,
2015- object_name = "tarfile " ,
2016- body = {} ,
2014+ file_path = f" { tmp_dir } /dataset.csv" ,
2015+ object_name = "dataset.csv " ,
2016+ body = payload ,
20172017 storage_uri_key = "storageUri" ,
20182018 method = "POST" ,
20192019 )
0 commit comments