reduce cost of large variant matrix

conda_build/config.py

@@ -31,7 +31,9 @@
 
 if TYPE_CHECKING:
  from pathlib import Path
- from typing import Any
+ from typing import Any, TypeVar
+
+ T = TypeVar("T")
 
 invocation_time = ""
 
@@ -821,14 +823,23 @@ def clean_pkgs(self):
 
  def copy(self) -> Config:
  new = copy.copy(self)
- # Use picke.loads(pickle.dumps(...) as a faster copy.deepcopy alternative.
- new.variant = pickle.loads(pickle.dumps(self.variant, pickle.HIGHEST_PROTOCOL))
+ new.variant = self._copy_variants(self.variant)
  if hasattr(self, "variants"):
- new.variants = pickle.loads(
- pickle.dumps(self.variants, pickle.HIGHEST_PROTOCOL)
- )
+ new.variants = self.copy_variants()
  return new
 
+ def _copy_variants(self, variant_or_list: T) -> T:
+ """Efficient deep copy used for variant dicts and lists"""
+ # Use pickle.loads(pickle.dumps(...) as a faster copy.deepcopy alternative.
+ return pickle.loads(pickle.dumps(variant_or_list, pickle.HIGHEST_PROTOCOL))
+
+ def copy_variants(self) -> list[dict] | None:
+ """Return deep copy of the variants list, if any"""
+ if getattr(self, "variants", None) is not None:
+ return self._copy_variants(self.variants)
+ else:
+ return None
+
  # context management - automatic cleanup if self.dirty or self.keep_old_work is not True
  def __enter__(self):
  pass

conda_build/metadata.py

@@ -2288,7 +2288,6 @@ def validate_features(self):
  def copy(self: Self) -> MetaData:
  new = copy.copy(self)
  new.config = self.config.copy()
- new.config.variant = copy.deepcopy(self.config.variant)
  new.meta = copy.deepcopy(self.meta)
  new.type = getattr(
  self, "type", "conda_v2" if self.config.conda_pkg_format == "2" else "conda"
@@ -2672,15 +2671,16 @@ def get_output_metadata_set(
  _check_run_constrained(output_tuples)
  return output_tuples
 
- def get_loop_vars(self):
- return get_vars(getattr(self.config, "input_variants", self.config.variants))
+ def get_loop_vars(self, subset=None):
+ return get_vars(
+ getattr(self.config, "input_variants", self.config.variants), subset=subset
+ )
 
  def get_used_loop_vars(self, force_top_level=False, force_global=False):
- loop_vars = self.get_loop_vars()
  used_vars = self.get_used_vars(
  force_top_level=force_top_level, force_global=force_global
  )
- return set(loop_vars).intersection(used_vars)
+ return self.get_loop_vars(subset=used_vars)
 
  def get_rendered_recipe_text(
  self, permit_undefined_jinja=False, extract_pattern=None

conda_build/render.py

@@ -835,12 +835,20 @@ def distribute_variants(
  used_variables = metadata.get_used_loop_vars(force_global=False)
  top_loop = metadata.get_reduced_variant_set(used_variables)
 
+ # defer potentially expensive copy of input variants list
+ # until after reduction of the list for each variant
+ # since the initial list can be very long
+ all_variants = metadata.config.variants
+ metadata.config.variants = []
+
  for variant in top_loop:
  from .build import get_all_replacements
 
  get_all_replacements(variant)
  mv = metadata.copy()
  mv.config.variant = variant
+ # start with shared list:
+ mv.config.variants = all_variants
 
  pin_run_as_build = variant.get("pin_run_as_build", {})
  if mv.numpy_xx and "numpy" not in pin_run_as_build:
@@ -860,6 +868,9 @@ def distribute_variants(
  )
  or mv.config.variants
  )
+ # copy variants before we start modifying them,
+ # but after we've reduced the list via the conform_dict filter
+ mv.config.variants = mv.config.copy_variants()
  get_all_replacements(mv.config.variants)
  pin_run_as_build = variant.get("pin_run_as_build", {})
  if mv.numpy_xx and "numpy" not in pin_run_as_build:

conda_build/variants.py

@@ -700,7 +700,10 @@ def get_package_variants(recipedir_or_metadata, config=None, variants=None):
  return filter_combined_spec_to_used_keys(combined_spec, specs=specs)
 
 
-def get_vars(variants: Iterable[dict[str, Any]]) -> set[str]:
+def get_vars(
+ variants: Iterable[dict[str, Any]],
+ subset: set[str] | None = None,
+) -> set[str]:
  """For purposes of naming/identifying, provide a way of identifying which variables contribute
  to the matrix dimensionality"""
  first, *others = variants
@@ -710,10 +713,12 @@ def get_vars(variants: Iterable[dict[str, Any]]) -> set[str]:
  "ignore_version",
  *ensure_list(first.get("extend_keys")),
  }
+ to_consider = set(first)
+ if subset is not None:
+ to_consider.intersection_update(subset)
+ to_consider.difference_update(special_keys)
  return {
- var
- for var in set(first) - special_keys
- if any(first[var] != other[var] for other in others)
+ var for var in to_consider if any(first[var] != other[var] for other in others)
  }
 
 

news/5392-variant-copy

@@ -0,0 +1,3 @@
+### Enhancements
+
+* Reduce render time when there is a large number of unused variants. (#5392)