def config_from_metadata(nb: NotebookNode) -> dict:
"""Extract configuration data from notebook/cell metadata."""
nb_metadata = nb.get("metadata", {}).get(META_KEY, {})
validate_metadata(nb_metadata, "/metadata")
diff_replace = [tuple(d) for d in nb_metadata.get("diff_replace", [])]
diff_ignore = set(nb_metadata.get("diff_ignore", []))
for i, cell in enumerate(nb.get("cells", [])):
cell_metadata = cell.get("metadata", {}).get(META_KEY, {})
validate_metadata(cell_metadata, f"/cells/{i}/metadata")
diff_replace.extend([
(f"/cells/{i}{p}", x, r)
for p, x, r in cell_metadata.get("diff_replace", [])
])
diff_ignore.update(
[f"/cells/{i}{p}" for p in cell_metadata.get("diff_ignore", [])])
return MetadataConfig(
tuple(diff_replace),
diff_ignore,
nb_metadata.get("skip", False),
nb_metadata.get("skip_reason", ""),
)
def extract_glue_data_cell(
cell: NotebookNode) -> list[tuple[str, NotebookNode]]:
"""Extract glue data from a single cell."""
outputs = []
data = []
for output in cell.get("outputs", []):
meta = output.get("metadata", {})
if "scrapbook" not in meta:
outputs.append(output)
continue
key = meta["scrapbook"]["name"]
mime_prefix = len(meta["scrapbook"].get("mime_prefix", ""))
output["data"] = {
k[mime_prefix:]: v
for k, v in output["data"].items()
}
data.append((key, output))
if not mime_prefix:
# assume that the output is a displayable object
outputs.append(output)
cell.outputs = outputs
return data
def blacken_code(cell: NotebookNode, resources: dict,
index: int) -> Tuple[NotebookNode, dict]:
"""Format python source code with black (see https://black.readthedocs.io)."""
try:
import black
except ImportError:
raise ImportError(
"black not installed: see https://black.readthedocs.io")
if cell.get("cell_type", None) != "code":
return cell, resources
# TODO use metadata to set target versions and whether to raise on exceptions
# i.e. black.FileMode(target_versions, {black.TargetVersion.PY36})
try:
cell.source = black.format_str(cell.source, mode=black.FileMode())
except (SyntaxError, black.InvalidInput):
logger.debug(f"cell {index} could not be formatted by black.")
# code cells don't require a trailing new line
cell.source = cell.source.rstrip()
return cell, resources
def beautifulsoup(cell: NotebookNode, resources: dict,
index: int) -> Tuple[NotebookNode, dict]:
"""Format text/html and image/svg+xml outputs with beautiful-soup.
See: https://beautiful-soup-4.readthedocs.io.
"""
try:
from bs4 import BeautifulSoup
except ImportError:
raise ImportError(
"bs4 not installed: see https://beautiful-soup-4.readthedocs.io")
if cell.get("cell_type", None) != "code":
return cell, resources
if "outputs" not in cell:
return cell, resources
for i, output in enumerate(cell.outputs):
if output.output_type not in ["execute_result", "display_data"]:
continue
for mimetype, value in output.get("data", {}).items():
if mimetype not in ["text/html", "image/svg+xml"]:
continue
path = f"/cells/{index}/outputs/{i}/{mimetype}"
# TODO use metadata to set builder and whether to raise on exceptions
try:
output["data"][mimetype] = BeautifulSoup(
output["data"][mimetype], "html.parser").prettify()
# record which paths have been formatted (mainly for testing)
resources.setdefault("beautifulsoup", []).append(path)
except Exception: # TODO what exceptions might be raised?
logger.debug(
f"{path} could not be formatted by beautiful-soup.")
return cell, resources
