|
| 1 | +# reasons for not using nbconvert cli tool: |
| 2 | +# * cannot keep output based on a given tag |
| 3 | + |
| 4 | +import argparse |
| 5 | +import logging |
| 6 | +from enum import Enum |
| 7 | +from pathlib import Path |
| 8 | +from typing import Any |
| 9 | + |
| 10 | +import nbconvert |
| 11 | +from nbconvert.preprocessors import Preprocessor |
| 12 | + |
| 13 | +PRESERVE_CELL_OUTPUT_KEY = "preserve-output" |
| 14 | +METADATA_TAG_KEY = "tags" |
| 15 | + |
| 16 | + |
| 17 | +class OutputMode(Enum): |
| 18 | + STDOUT = "stdout" |
| 19 | + INPLACE = "inplace" |
| 20 | + |
| 21 | + |
| 22 | +class CustomClearOutputPreprocessor(Preprocessor): |
| 23 | + """ |
| 24 | + Removes the output from all code cells in a notebook. |
| 25 | + Option to keep cell output for cells with a given metadata tag |
| 26 | + """ |
| 27 | + |
| 28 | + def preprocess_cell( |
| 29 | + self, cell: Any, resources: Any, cell_index: Any |
| 30 | + ) -> tuple[Any, Any]: |
| 31 | + """ |
| 32 | + Apply a transformation on each cell. See base.py for details. |
| 33 | + """ |
| 34 | + if cell.cell_type == "code" and PRESERVE_CELL_OUTPUT_KEY not in cell[ |
| 35 | + "metadata" |
| 36 | + ].get(METADATA_TAG_KEY, []): |
| 37 | + cell.outputs = [] |
| 38 | + cell.execution_count = None |
| 39 | + return cell, resources |
| 40 | + |
| 41 | + |
| 42 | +def main(input_path: Path, output_mode: OutputMode) -> None: |
| 43 | + logger = logging.getLogger("NotebookCleaner") |
| 44 | + logger.info(f"Cleaning notebooks from `{input_path}`, mode: `{output_mode}`") |
| 45 | + |
| 46 | + exporter = nbconvert.NotebookExporter() |
| 47 | + |
| 48 | + metadata_cleaner = nbconvert.preprocessors.ClearMetadataPreprocessor( |
| 49 | + preserve_cell_metadata_mask=METADATA_TAG_KEY |
| 50 | + ) |
| 51 | + output_cleaner = CustomClearOutputPreprocessor() # type: ignore |
| 52 | + |
| 53 | + exporter.register_preprocessor(metadata_cleaner, enabled=True) |
| 54 | + exporter.register_preprocessor(output_cleaner, enabled=True) |
| 55 | + |
| 56 | + if input_path.is_file(): |
| 57 | + notebooks = [input_path] |
| 58 | + else: |
| 59 | + notebooks = [ |
| 60 | + f for f in input_path.iterdir() if f.is_file() and f.suffix == ".ipynb" |
| 61 | + ] |
| 62 | + |
| 63 | + logger.info(f"Formatting {len(notebooks)} notebooks.") |
| 64 | + |
| 65 | + for notebook in notebooks: |
| 66 | + output = exporter.from_filename(str(notebook)) |
| 67 | + |
| 68 | + formatted_notebook = output[0] |
| 69 | + |
| 70 | + if output_mode == OutputMode.INPLACE: |
| 71 | + with notebook.open(mode="w") as file: |
| 72 | + file.write(str(formatted_notebook)) |
| 73 | + elif output_mode == OutputMode.STDOUT: |
| 74 | + print(formatted_notebook) |
| 75 | + |
| 76 | + |
| 77 | +if __name__ == "__main__": |
| 78 | + parser = argparse.ArgumentParser() |
| 79 | + parser.add_argument("-o", "--output", choices=[e.value for e in OutputMode]) |
| 80 | + parser.add_argument( |
| 81 | + "-i", "--input", default="examples", help="path to the notebook file or folder" |
| 82 | + ) |
| 83 | + |
| 84 | + args = parser.parse_args() |
| 85 | + |
| 86 | + main(Path(args.input), OutputMode(args.output)) |
0 commit comments