jasonfan's picture
Add files using upload-large-folder tool
71687cf verified
# Copyright (C) 2012 Anaconda, Inc
# SPDX-License-Identifier: BSD-3-Clause
"""Environment object describing the conda environment.yaml file."""
from __future__ import annotations
import os
import re
from itertools import chain
from typing import TYPE_CHECKING
from ..base.context import context
from ..cli import common
from ..common.io import dashlist
from ..common.iterators import unique
from ..common.path import expand
from ..common.serialize import json, yaml
from ..core.prefix_data import PrefixData
from ..deprecations import deprecated
from ..exceptions import (
CondaMultiError,
EnvironmentFileEmpty,
EnvironmentFileInvalid,
EnvironmentFileNotFound,
InvalidMatchSpec,
)
from ..gateways.connection.download import download_text
from ..gateways.connection.session import CONDA_SESSION_SCHEMES
from ..history import History
from ..models.environment import Environment as EnvironmentModel
from ..models.environment import EnvironmentConfig
from ..models.match_spec import MatchSpec
if TYPE_CHECKING:
from typing import Any
REQUIRED_KEYS = frozenset(("dependencies",))
OPTIONAL_KEYS = frozenset(
(
"name",
"prefix",
"channels",
"variables",
)
)
VALID_KEYS = frozenset((*REQUIRED_KEYS, *OPTIONAL_KEYS))
def field_type_validation(field_name: str, value: Any, value_type: Any) -> None:
"""Validates the type of a value"""
if not isinstance(value, value_type):
raise EnvironmentFileInvalid(
f"Invalid type for '{field_name}', expected a {value_type.__name__}"
)
def prefix_validation(prefix: str):
"""Validate the contents of the prefix field.
Will ensure:
* prefix is a string
"""
field_type_validation("prefix", prefix, str)
def name_validation(name: str):
"""Validate the contents of the name field.
Will ensure:
* name is a string
"""
field_type_validation("name", name, str)
def dependencies_validation(dependencies: list):
"""Validate the contents of the dependencies field.
Will ensure:
* dependencies are a list
* all string dependencies are MatchSpec compatible
* the only other type allowed is a dict
"""
field_type_validation("dependencies", dependencies, list)
errors = []
for dependency in dependencies:
if isinstance(dependency, str):
# If the dependency is a string type, it must be
# MatchSpec compatible.
try:
MatchSpec(dependency)
except InvalidMatchSpec as err:
errors.append(EnvironmentFileInvalid(str(err)))
elif isinstance(dependency, dict):
# dict types are also allowed. There are no requirements
# for the form of this entry
pass
else:
# All other types are invalid
errors.append(
EnvironmentFileInvalid(
f"'{dependency}' is an invalid type for a 'dependency'"
)
)
if errors:
raise CondaMultiError(errors)
def channels_validation(channels: list):
"""Validate the contents of the channels field.
Will ensure:
* channels is a list
* all entries are strings
"""
field_type_validation("channels", channels, list)
for channel in channels:
if not isinstance(channel, str):
raise EnvironmentFileInvalid(
"`channels` key must only contain strings. Found '{channel}'"
)
def variables_validation(variables: dict[str, str]):
"""Validate the contents of the variables field.
Will ensure:
* variables is a dict
* all entries are strings
"""
field_type_validation("variables", variables, dict)
SCHEMA_VALIDATORS = {
"name": name_validation,
"prefix": prefix_validation,
"dependencies": dependencies_validation,
"channels": channels_validation,
"variables": variables_validation,
}
def get_schema_errors(data: dict) -> list[EnvironmentFileInvalid]:
"""Parses environment.yaml data to build a list of schema errors
Will produce errors to ensure:
* all required fields are present
* all fields contain valid data
:param dict data: The contents of the environment.yaml
:returns errors: A list of EnvironmentFileInvalid exceptions that occurred during validation
"""
errors = []
# Ensure all required keys are present
for field in REQUIRED_KEYS.difference(data):
errors.append(EnvironmentFileInvalid(f"Missing required field '{field}'"))
# Run validations on all the relevant fields, extra keys are ignored
for key, validator in SCHEMA_VALIDATORS.items():
try:
validator(data[key])
except KeyError:
pass
except EnvironmentFileInvalid as err:
errors.append(err)
return errors
def validate_keys(data, kwargs):
"""Check for unknown keys, remove them and print a warning"""
invalid_keys = []
new_data = data.copy() if data else {}
for key in data.keys():
if key not in VALID_KEYS:
invalid_keys.append(key)
new_data.pop(key)
if invalid_keys:
filename = kwargs.get("filename")
verb = "are" if len(invalid_keys) != 1 else "is"
plural = "s" if len(invalid_keys) != 1 else ""
print(
f"\nEnvironmentSectionNotValid: The following section{plural} on "
f"'{filename}' {verb} invalid and will be ignored:"
)
for key in invalid_keys:
print(f" - {key}")
print()
deps = data.get("dependencies") or []
depsplit = re.compile(r"[<>~\s=]")
is_pip = lambda dep: "pip" in depsplit.split(dep)[0].split("::")
lists_pip = any(is_pip(dep) for dep in deps if not isinstance(dep, dict))
for dep in deps:
if isinstance(dep, dict) and "pip" in dep and not lists_pip:
print(
"Warning: you have pip-installed dependencies in your environment file, "
"but you do not list pip itself as one of your conda dependencies. Conda "
"may not use the correct pip to install your packages, and they may end up "
"in the wrong place. Please add an explicit pip dependency. I'm adding one"
" for you, but still nagging you."
)
new_data["dependencies"].insert(0, "pip")
break
return new_data
def from_environment(
name, prefix, no_builds=False, ignore_channels=False, from_history=False
):
"""
Get ``EnvironmentYaml`` object from prefix
Args:
name: The name of environment
prefix: The path of prefix
no_builds: Whether has build requirement
ignore_channels: whether ignore_channels
from_history: Whether environment file should be based on explicit specs in history
Returns: EnvironmentYaml object
"""
pd = PrefixData(prefix, interoperability=True)
variables = pd.get_environment_env_vars()
if from_history:
history = History(prefix).get_requested_specs_map()
deps = [str(package) for package in history.values()]
return EnvironmentYaml(
name=name,
dependencies=deps,
channels=list(context.channels),
prefix=prefix,
variables=variables,
)
conda_precs = pd.get_conda_packages()
python_precs = pd.get_python_packages()
dependencies = [
conda_prec.spec_no_build if no_builds else conda_prec.spec
for conda_prec in conda_precs
]
if python_precs:
dependencies.append(
{
"pip": [
f"{python_prec.name}=={python_prec.version}"
for python_prec in python_precs
]
}
)
channels = list(context.channels)
if not ignore_channels:
for conda_prec in conda_precs:
canonical_name = conda_prec.channel.canonical_name
if canonical_name not in channels:
channels.insert(0, canonical_name)
return EnvironmentYaml(
name=name,
dependencies=dependencies,
channels=channels,
prefix=prefix,
variables=variables,
)
def from_yaml(yamlstr: str, **kwargs) -> EnvironmentYaml:
"""Load and return a ``EnvironmentYaml`` from a given ``yaml`` string
:param yamlstr: The contents of the environment.yaml
:param raise_validation_errors: Indicates if an error should be raised if the yamlstr
is found to be invalid
:returns EnvironmentYaml: A representation of the environment file
"""
data = yaml.loads(yamlstr)
filename = kwargs.get("filename")
if data is None:
raise EnvironmentFileEmpty(filename)
# Perform schema validation. This will output a warning for any invalid schema.
errors = get_schema_errors(data)
if errors:
# Warn for all the schema errors in the environment
deprecated.topic(
"26.3",
"26.9",
topic="The environment file is not fully CEP 24 compliant",
addendum=(
"In the future, this configuration will be rejected. Please fix the following "
"errors in order to make the configuration valid: "
f"{dashlist(errors)}"
),
deprecation_type=FutureWarning,
)
data = validate_keys(data, kwargs)
if kwargs is not None:
for key, value in kwargs.items():
data[key] = value
_expand_channels(data)
return EnvironmentYaml(**data)
def _expand_channels(data):
"""Expands ``EnvironmentYaml`` variables for the channels found in the ``yaml`` data"""
data["channels"] = [
os.path.expandvars(channel) for channel in data.get("channels", [])
]
def load_file(filename):
"""Load and return an yaml string from a given file"""
url_scheme = filename.split("://", 1)[0]
if url_scheme in CONDA_SESSION_SCHEMES:
yamlstr = download_text(filename)
elif not os.path.exists(filename):
raise EnvironmentFileNotFound(filename)
else:
with open(filename, "rb") as fp:
yamlb = fp.read()
try:
yamlstr = yamlb.decode("utf-8")
except UnicodeDecodeError:
yamlstr = yamlb.decode("utf-16")
return yamlstr
def from_file(filename):
"""Load and return an ``EnvironmentYaml`` from a given file"""
yamlstr = load_file(filename)
return from_yaml(yamlstr, filename=filename)
class Dependencies(dict):
"""A ``dict`` subclass that parses the raw dependencies into a conda and pip list"""
def __init__(self, raw, *args, **kwargs):
super().__init__(*args, **kwargs)
self.raw = raw
self.parse()
def parse(self):
"""Parse the raw dependencies into a conda and pip list"""
if not self.raw:
return
self.update({"conda": []})
for line in self.raw:
if isinstance(line, dict):
self.update(line)
else:
self["conda"].append(str(MatchSpec(line)))
if "pip" in self:
if not self["pip"]:
del self["pip"]
if not any(MatchSpec(s).name == "pip" for s in self["conda"]):
self["conda"].append("pip")
# TODO only append when it's not already present
def add(self, package_name):
"""Add a package to the ``EnvironmentYaml``"""
self.raw.append(package_name)
self.parse()
class EnvironmentYaml:
"""A class representing an ``environment.yaml`` file"""
def __init__(
self,
name=None,
filename=None,
channels=None,
dependencies=None,
prefix=None,
variables=None,
):
self.name = name
self.filename = filename
self.prefix = prefix
self.dependencies = Dependencies(dependencies)
self.variables = variables
if channels is None:
channels = []
self.channels = channels
def add_channels(self, channels):
"""Add channels to the ``EnvironmentYaml``"""
self.channels = list(unique(chain.from_iterable((channels, self.channels))))
def remove_channels(self):
"""Remove all channels from the ``EnvironmentYaml``"""
self.channels = []
def to_dict(self, stream=None):
"""Convert information related to the ``EnvironmentYaml`` into a dictionary"""
d = {"name": self.name}
if self.channels:
d["channels"] = self.channels
if self.dependencies:
d["dependencies"] = self.dependencies.raw
if self.variables:
d["variables"] = self.variables
if self.prefix:
d["prefix"] = self.prefix
if stream is None:
return d
stream.write(json.dumps(d))
def to_yaml(self, stream=None):
"""Convert information related to the ``EnvironmentYaml`` into a ``yaml`` string"""
d = self.to_dict()
return yaml.write(d, fp=stream)
def save(self):
"""Save the ``EnvironmentYaml`` data to a ``yaml`` file"""
with open(self.filename, "wb") as fp:
self.to_yaml(stream=fp)
def to_environment_model(self) -> EnvironmentModel:
"""Convert the ``Environment`` into a ``model.Environment`` object"""
config = EnvironmentConfig(channels=tuple(self.channels))
external_packages = {}
if pip_dependencies := self.dependencies.get("pip"):
external_packages["pip"] = pip_dependencies
requested_packages = [
MatchSpec(spec) for spec in self.dependencies.get("conda", [])
]
return EnvironmentModel(
prefix=self.prefix or context.target_prefix,
platform=context.subdir,
name=self.name,
config=config,
variables=self.variables,
external_packages=external_packages,
requested_packages=requested_packages,
)
@deprecated("26.3", "26.9", addendum="Use `conda.env.env.EnvironmentYaml` instead.")
class Environment(EnvironmentYaml):
"""A class representing an ``environment.yaml`` file"""
@deprecated("25.9", "26.3")
def get_filename(filename):
"""Expand filename if local path or return the ``url``"""
url_scheme = filename.split("://", 1)[0]
if url_scheme in CONDA_SESSION_SCHEMES:
return filename
else:
return expand(filename)
def print_result(args, prefix, result):
"""Print the result of an install operation"""
if context.json:
if result["conda"] is None and result["pip"] is None:
common.stdout_json_success(
message="All requested packages already installed."
)
else:
if result["conda"] is not None:
actions = result["conda"]
else:
actions = {}
if result["pip"] is not None:
actions["PIP"] = result["pip"]
common.stdout_json_success(prefix=prefix, actions=actions)
else:
common.print_activate(args.name or prefix)