CHANGELOG.md

CHANGELOG

0.3.37

• update: Packages for fixing vulnerabilities

0.3.36

• update: Reftime handling works if reftime is malformed

0.3.35

• add: validations for the input file for conversation tagging

0.3.34

• PL-61: Add retry mechanism for uploading data to Label studio

0.3.33

• PL-1300: new data format support & extra meta-annotation-columns (#21)

0.3.32

• fix: alternatives column postprocessing after downloading from labelstudio

0.3.31

• add: required paramter --data-label for uploading datasets

0.3.30

• add: modified downloading from db to integrate labelstudio db queries

0.3.29

• fix: data format issue after downloading from labelstudio

0.3.28

• fix: pass project_id for LS and job_id for TOG

0.3.27

• add: makes batches of batched_dataset while uploading to tog db - limits max async connections to db at a time
• add: retries with sleep while uploading data to tog db

0.3.26

• fix: bug(#13) passing of creds into creating Job as None

0.3.25

• fix: datetime formatting produces None if none of the formats match.
• fix: connection timeout mitigated by batching the requests.
• update: data download uses start and end date for query only if provided.

0.3.24

• add: error handling for uploads.

0.3.23

• add: job-id to dataset.
• add: json cols are unpacked.

0.3.22

• feat: Add labelstudio integration. We can upload/download datasets to labelstudio.

0.3.21

• update: Dedupe id uses uuid4 at runtime.

0.3.20

• fix: Parse timezone from strings.

0.3.18

• add: Support for legacy dataframes with alternatives.

0.3.17

• update: gh actions script updates.

0.3.16

• update: Data upload is more tolerant or utterance json vs python object.

0.3.15

• update: JSON fields are stringified.

0.3.13

• update: describe and stat dataset requires database object instance optionally.

0.3.11

• update: deps for compatibility with skit-pipelines.

0.3.10

• update: Higher tolerance for dataset schema errors.

0.3.9

• fix: -j added back to upload dataset command.

0.3.8

• update: remove preprocessing over df.

0.3.7

• fix: bug preventing data uploads due to larger than supported values in source.

0.3.6

• update: dataset {stat, describe} can apply date ranges.

0.3.5

• update: CLI allows database params as input.

0.3.4

• update: obtain type of the dataset from Job::type().

0.3.3

• refactor: no difference in usage.

0.3.2

• refactor: dependency on skit-fixdf removed.

0.3.1

• feat: Upload datasets via cli.

0.3.0

• feat: Access datasets from dvc.