0.62.4

What's Changed

• fix: size automatic workers by the container CPU quota by @dmpetrov in #2135
• fix: store NaN in pandas text columns as None by @shcheklein in #2164
• fix: keep the column type for all-missing pandas columns by @shcheklein in #2161
• fix: read Parquet map columns as dicts by @shcheklein in #2162
• fix: read_values with column= and a single signal by @shcheklein in #2163
• fix(file): normalize explicit text write modes by @bosprimigenious in #2166
• refactor!: drop the name argument from read_values and read_pandas by @shcheklein in #2171
• perf(tabular): read Parquet/CSV schemas concurrently by @dmpetrov in #2173
• refactor!: drop the column argument from read_storage by @shcheklein in #2178
• fix(func): treat strings as literals in Func comparisons by @dmpetrov in #2179
• refactor: drop the column argument from listing helpers by @shcheklein in #2188
• fix(tabular): read CSV and Parquet files that have a source column by @shcheklein in #2187
• fix(tabular): read Parquet exported with only a source signal by @shcheklein in #2189
• fix: readers rename their own signal, not the data column by @shcheklein in #2190
• feat: read LeRobot v3 datasets by @dmpetrov in #2184
• fix(tar): reject compressed tar archives by @dmpetrov in #2186
• Fast UDF: stats response once in 3 sec & PYTHONPATH for UDFs by @dmpetrov in #2080
• perf: read media metadata without filling the file cache by @dmpetrov in #2118
• feat: add read_webdataset by @dmpetrov in #2182
• Delta: keep results in sync with the current source by @dmpetrov in #2115
• fix(file): give a buffered stream from File.open() without the cache by @dmpetrov in #2180
• perf(video): read with 1 MiB blocks when decoding from mid-file by @dmpetrov in #2181

New Contributors

• @bosprimigenious made their first contribution in #2166

Full Changelog: 0.62.3...0.62.4