zyratlo opened a new issue, #8007: URL: https://github.com/apache/texera/issues/8007
### Feature Summary The migration tool converts a Jupyter notebook into a Texera workflow. A notebook is the only accepted input: the import modal takes a single `.ipynb`, and the conversion depends on the notebook's cell structure, both to give the LLM natural unit boundaries and to build the cell-to-operator mapping that drives highlighting in the workspace. Plain Python files are the next input we want to support. Users with existing code outside Texera often have `.py` scripts rather than notebooks, and today the only way to migrate one is to convert it into a notebook by hand first. ### Proposed Solution or Design Accept a `.py` file as an input to the migration tool and produce a workflow from it, reusing the notebook path wherever it already generalizes. The open question is what replaces the cell boundary. A notebook arrives pre-segmented and the tool keys its mapping off per-cell UUIDs; a script does not, so the tool needs some notion of a unit to map operators back to. Whether that is a line range, a syntactic block, an LLM-chosen segmentation, or something else is not decided here. Related considerations to work out when this is picked up: * What the Jupyter panel shows for a `.py` input, given that it currently embeds the source notebook and highlights the cell matching the selected operator. Can choose to not show anything at all for Python files. * Whether the file is still uploaded to the Jupyter workspace, and under what name, following the `notebook_<wid>.ipynb` convention * Whether existing notebook migrations are affected at all, or whether this is purely additive. This is follow-up work after the Jupyter notebook migration tool is complete. It is not scoped or designed yet. ### Affected Area Other -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
