phylogenie 1.0.8__py3-none-any.whl → 2.0.1__py3-none-any.whl

phylogenie/core/context/distributions.py

@@ -1,125 +0,0 @@
-from abc import ABC, abstractmethod
-from enum import Enum
-from typing import Annotated, Generic, Literal, TypeVar
-
-from numpy.random import Generator, default_rng
-from pydantic import Field
-
-import phylogenie.typings as pgt
-from phylogenie.configs import StrictBaseModel
-
-_T = TypeVar("_T")
-
-
-class Type(str, Enum):
-    UNIFORM = "uniform"
-    NORMAL = "normal"
-    LOGNORMAL = "lognormal"
-    WEIBULL = "weibull"
-    EXPONENTIAL = "exponential"
-    GAMMA = "gamma"
-    BETA = "beta"
-    INT_UNIFORM = "int-uniform"
-    CATEGORICAL = "categorical"
-
-
-class Distribution(StrictBaseModel, ABC, Generic[_T]):
-    @abstractmethod
-    def _sample(self, rng: Generator) -> _T: ...
-
-    def sample(self, rng: int | Generator | None = None) -> _T:
-        if not isinstance(rng, Generator):
-            rng = default_rng(rng)
-        return self._sample(rng)
-
-
-class Scalar(Distribution[pgt.Scalar]): ...
-
-
-class Uniform(Scalar):
-    type: Literal[Type.UNIFORM] = Type.UNIFORM
-    low: float
-    high: float
-
-    def _sample(self, rng: Generator) -> float:
-        return rng.uniform(self.low, self.high)
-
-
-class Normal(Scalar):
-    type: Literal[Type.NORMAL] = Type.NORMAL
-    mean: float
-    std: float
-
-    def _sample(self, rng: Generator) -> float:
-        return rng.normal(self.mean, self.std)
-
-
-class LogNormal(Scalar):
-    type: Literal[Type.LOGNORMAL] = Type.LOGNORMAL
-    mean: float
-    std: float
-
-    def _sample(self, rng: Generator) -> float:
-        return rng.lognormal(self.mean, self.std)
-
-
-class Weibull(Scalar):
-    type: Literal[Type.WEIBULL] = Type.WEIBULL
-    scale: float
-    shape: float
-
-    def _sample(self, rng: Generator) -> float:
-        return rng.weibull(self.shape) * self.scale
-
-
-class Exponential(Scalar):
-    type: Literal[Type.EXPONENTIAL] = Type.EXPONENTIAL
-    scale: float
-
-    def _sample(self, rng: Generator) -> float:
-        return rng.exponential(self.scale)
-
-
-class Gamma(Scalar):
-    type: Literal[Type.GAMMA] = Type.GAMMA
-    scale: float
-    shape: float
-
-    def _sample(self, rng: Generator) -> float:
-        return rng.gamma(self.shape, self.scale)
-
-
-class Beta(Scalar):
-    type: Literal[Type.BETA] = Type.BETA
-    alpha: float
-    beta: float
-
-    def _sample(self, rng: Generator) -> float:
-        return rng.beta(self.alpha, self.beta)
-
-
-class IntUniform(Scalar):
-    type: Literal[Type.INT_UNIFORM] = Type.INT_UNIFORM
-    low: int
-    high: int
-
-    def _sample(self, rng: Generator) -> int:
-        return int(rng.integers(self.low, self.high))
-
-
-class Categorical(Distribution[str]):
-    type: Literal[Type.CATEGORICAL] = Type.CATEGORICAL
-    categories: list[str]
-    probabilities: list[float]
-
-    def _sample(self, rng: Generator) -> str:
-        return str(rng.choice(self.categories, p=self.probabilities))
-
-
-ScalarDistributionConfig = Annotated[
-    Uniform | Normal | LogNormal | Weibull | Exponential | Gamma | Beta | IntUniform,
-    Field(discriminator="type"),
-]
-DistributionConfig = Annotated[
-    ScalarDistributionConfig | Categorical, Field(discriminator="type")
-]

phylogenie/core/context/factories.py

@@ -1,54 +0,0 @@
-from numpy.random import Generator
-
-import phylogenie.core.context.configs as cfg
-import phylogenie.typings as pgt
-from phylogenie.core.context import distributions
-
-
-def _sample_vector1D(x: distributions.Scalar, N: int, rng: Generator) -> pgt.Vector1D:
-    return [x.sample(rng) for _ in range(N)]
-
-
-def _sample_vector2D(
-    x: distributions.Scalar,
-    size: tuple[int, int],
-    zero_diagonal: bool,
-    rng: Generator,
-) -> pgt.Vector2D:
-    n_rows, n_cols = size
-    v = [_sample_vector1D(x, n_cols, rng) for _ in range(n_rows)]
-    if zero_diagonal:
-        if n_rows != n_cols:
-            raise ValueError(
-                f"It is impossible to initialize a non-square matrix with zero the diagonal (got x={x}, size={size} and zero_diagonal=True)"
-            )
-        for i in range(n_rows):
-            v[i][i] = 0
-    return v
-
-
-def _sample_vector3D(
-    x: distributions.Scalar,
-    size: tuple[int, int, int],
-    zero_diagonal: bool,
-    rng: Generator,
-) -> pgt.Vector3D:
-    n_matrices, n_rows, n_cols = size
-    return [
-        _sample_vector2D(x, (n_rows, n_cols), zero_diagonal, rng)
-        for _ in range(n_matrices)
-    ]
-
-
-def context_factory(x: cfg.ContextConfig, rng: Generator) -> pgt.Data:
-    data: pgt.Data = {}
-    for key, value in x.items():
-        if isinstance(value, distributions.Distribution):
-            data[key] = value.sample(rng)
-        elif isinstance(value, cfg.Vector1DModel):
-            data[key] = _sample_vector1D(value.x, value.size, rng)
-        elif isinstance(value, cfg.Vector2DModel):
-            data[key] = _sample_vector2D(value.x, value.size, value.zero_diagonal, rng)
-        else:
-            data[key] = _sample_vector3D(value.x, value.size, value.zero_diagonal, rng)
-    return data

phylogenie/core/msas/__init__.py

@@ -1,10 +0,0 @@
-from typing import Annotated
-
-from pydantic import Field
-
-from phylogenie.core.msas.alisim import AliSimGenerator
-
-MSAsGeneratorConfig = Annotated[
-    AliSimGenerator,
-    Field(discriminator="backend"),
-]

phylogenie/core/msas/alisim.py

@@ -1,35 +0,0 @@
-import subprocess
-from typing import Literal
-
-from numpy.random import Generator
-
-import phylogenie.typings as pgt
-from phylogenie.core.msas.base import BackendType, MSAsGenerator
-
-
-class AliSimGenerator(MSAsGenerator):
-    backend: Literal[BackendType.ALISIM] = BackendType.ALISIM
-    iqtree_path: str = "iqtree2"
-    args: dict[str, str | int | float]
-
-    def _generate_one_from_tree(
-        self, filename: str, tree_file: str, rng: Generator, data: pgt.Data
-    ) -> None:
-        command = [
-            self.iqtree_path,
-            "--alisim",
-            filename,
-            "--tree",
-            tree_file,
-            "--seed",
-            str(rng.integers(0, 2**32 - 1)),
-        ]
-
-        for key, value in self.args.items():
-            command.extend(
-                [key, value.format(**data) if isinstance(value, str) else str(value)]
-            )
-
-        command.extend(["-af", "fasta"])
-        subprocess.run(command, check=True, stdout=subprocess.DEVNULL)
-        subprocess.run(["rm", f"{tree_file}.log"], check=True)

phylogenie/core/msas/base.py

@@ -1,51 +0,0 @@
-import os
-from abc import abstractmethod
-from enum import Enum
-from pathlib import Path
-from typing import Literal
-
-from numpy.random import Generator
-
-import phylogenie.typings as pgt
-from phylogenie.core.dataset import DatasetGenerator, DataType
-from phylogenie.core.trees import TreesGeneratorConfig
-
-
-class BackendType(str, Enum):
-    ALISIM = "alisim"
-
-
-MSAS_DIRNAME = "MSAs"
-TREES_DIRNAME = "trees"
-
-
-class MSAsGenerator(DatasetGenerator):
-    data_type: Literal[DataType.MSAS] = DataType.MSAS
-    trees: TreesGeneratorConfig
-    keep_trees: bool = False
-
-    @abstractmethod
-    def _generate_one_from_tree(
-        self, filename: str, tree_file: str, rng: Generator, data: pgt.Data
-    ) -> None: ...
-
-    def _generate_one(self, filename: str, rng: Generator, data: pgt.Data) -> None:
-        if self.keep_trees:
-            base_dir = Path(filename).parent
-            file_id = Path(filename).stem
-            tree_filename = os.path.join(base_dir, TREES_DIRNAME, file_id)
-            msas_dir = os.path.join(base_dir, MSAS_DIRNAME)
-            os.makedirs(msas_dir, exist_ok=True)
-            msa_filename = os.path.join(msas_dir, file_id)
-        else:
-            tree_filename = f"{filename}.temp-tree"
-            msa_filename = filename
-
-        self.trees.generate_one(
-            filename=tree_filename, data=data, seed=int(rng.integers(0, 2**32 - 1))
-        )
-        self._generate_one_from_tree(
-            filename=msa_filename, tree_file=f"{tree_filename}.nwk", rng=rng, data=data
-        )
-        if not self.keep_trees:
-            os.remove(f"{tree_filename}.nwk")

phylogenie/core/trees/__init__.py

@@ -1,11 +0,0 @@
-from typing import Annotated
-
-from pydantic import Field
-
-from phylogenie.core.trees.remaster import ReMASTERGeneratorConfig
-from phylogenie.core.trees.treesimulator import TreeSimulatorGeneratorConfig
-
-TreesGeneratorConfig = Annotated[
-    ReMASTERGeneratorConfig | TreeSimulatorGeneratorConfig,
-    Field(discriminator="backend"),
-]

phylogenie/core/trees/base.py

@@ -1,13 +0,0 @@
-from enum import Enum
-from typing import Literal
-
-from phylogenie.core.dataset import DatasetGenerator, DataType
-
-
-class BackendType(str, Enum):
-    REMASTER = "remaster"
-    TREESIMULATOR = "treesimulator"
-
-
-class TreesGenerator(DatasetGenerator):
-    data_type: Literal[DataType.TREES] = DataType.TREES

phylogenie/core/trees/remaster/__init__.py

@@ -1,3 +0,0 @@
-from phylogenie.core.trees.remaster.generator import ReMASTERGeneratorConfig
-
-__all__ = ["ReMASTERGeneratorConfig"]

phylogenie/core/trees/remaster/configs.py

@@ -1,14 +0,0 @@
-import phylogenie.core.configs as cfg
-from phylogenie.configs import StrictBaseModel
-
-
-class ReactionConfig(StrictBaseModel):
-    rate: cfg.SkylineParameterLikeConfig
-    value: str
-
-
-class PunctualReactionConfig(StrictBaseModel):
-    times: cfg.ManyScalarsConfig
-    value: str
-    p: cfg.ManyScalarsConfig | None = None
-    n: cfg.ManyIntsConfig | None = None

phylogenie/core/trees/remaster/factories.py

@@ -1,26 +0,0 @@
-import phylogenie.core.trees.remaster.configs as cfg
-import phylogenie.typings as pgt
-from phylogenie.backend.remaster import PunctualReaction, Reaction
-from phylogenie.core.factories import (
-    many_ints_factory,
-    many_scalars_factory,
-    skyline_parameter_like_factory,
-)
-
-
-def reaction_factory(x: cfg.ReactionConfig, data: pgt.Data) -> Reaction:
-    return Reaction(
-        rate=skyline_parameter_like_factory(x.rate, data),
-        value=x.value,
-    )
-
-
-def punctual_reaction_factory(
-    x: cfg.PunctualReactionConfig, data: pgt.Data
-) -> PunctualReaction:
-    return PunctualReaction(
-        times=many_scalars_factory(x.times, data),
-        value=x.value,
-        p=None if x.p is None else many_scalars_factory(x.p, data),
-        n=None if x.n is None else many_ints_factory(x.n, data),
-    )

phylogenie/core/trees/remaster/generator.py

@@ -1,177 +0,0 @@
-from collections.abc import Iterable
-from enum import Enum
-from typing import Annotated, Literal
-
-from numpy.random import Generator
-from pydantic import Field
-
-import phylogenie.core.configs as cfg
-import phylogenie.typings as pgt
-from phylogenie.backend.remaster import (
-    DEFAULT_POPULATION,
-    SAMPLE_POPULATION,
-    Reaction,
-    generate_trees,
-    get_canonical_reactions,
-    get_epidemiological_reactions,
-    get_FBD_reactions,
-)
-from phylogenie.core.factories import (
-    skyline_matrix_coercible_factory,
-    skyline_vector_coercible_factory,
-)
-from phylogenie.core.trees.base import BackendType, TreesGenerator
-from phylogenie.core.trees.remaster.configs import (
-    PunctualReactionConfig,
-    ReactionConfig,
-)
-from phylogenie.core.trees.remaster.factories import (
-    punctual_reaction_factory,
-    reaction_factory,
-)
-
-
-class ParameterizationType(str, Enum):
-    CANONICAL = "canonical"
-    EPIDEMIOLOGICAL = "epidemiological"
-    FBD = "fbd"
-
-
-class ReMASTERGenerator(TreesGenerator):
-    backend: Literal[BackendType.REMASTER] = BackendType.REMASTER
-    beast_path: str = "beast"
-    populations: str | list[str] = DEFAULT_POPULATION
-    init_population: str = DEFAULT_POPULATION
-    sample_population: str = SAMPLE_POPULATION
-    reactions: Iterable[ReactionConfig] = Field(default_factory=tuple)
-    punctual_reactions: Iterable[PunctualReactionConfig] = Field(default_factory=tuple)
-    trajectory_attrs: dict[str, str | int | float] = Field(default_factory=dict)
-
-    def _generate_one_from_extra_reactions(
-        self, filename: str, rng: Generator, data: pgt.Data, reactions: list[Reaction]
-    ) -> None:
-        generate_trees(
-            tree_filename=f"{filename}.nwk",
-            populations=self.populations,
-            init_population=self.init_population.format(**data),
-            sample_population=self.sample_population,
-            reactions=[reaction_factory(r, data) for r in self.reactions] + reactions,
-            punctual_reactions=[
-                punctual_reaction_factory(r, data) for r in self.punctual_reactions
-            ],
-            trajectory_attrs={
-                k: v.format(**data) if isinstance(v, str) else str(v)
-                for k, v in self.trajectory_attrs.items()
-            },
-            seed=int(rng.integers(0, 2**31 - 1)),
-            beast_path=self.beast_path,
-        )
-
-
-class CanonicalReMASTERGenerator(ReMASTERGenerator):
-    parameterization: Literal[ParameterizationType.CANONICAL] = (
-        ParameterizationType.CANONICAL
-    )
-    birth_rates: cfg.SkylineVectorCoercibleConfig = 0
-    death_rates: cfg.SkylineVectorCoercibleConfig = 0
-    sampling_rates: cfg.SkylineVectorCoercibleConfig = 0
-    removal_probabilities: cfg.SkylineVectorCoercibleConfig = 0
-    migration_rates: cfg.SkylineMatrixCoercibleConfig = 0
-    birth_rates_among_demes: cfg.SkylineMatrixCoercibleConfig = 0
-
-    def _generate_one(self, filename: str, rng: Generator, data: pgt.Data) -> None:
-        reactions = get_canonical_reactions(
-            populations=self.populations,
-            sample_population=self.sample_population,
-            birth_rates=skyline_vector_coercible_factory(self.birth_rates, data),
-            death_rates=skyline_vector_coercible_factory(self.death_rates, data),
-            sampling_rates=skyline_vector_coercible_factory(self.sampling_rates, data),
-            removal_probabilities=skyline_vector_coercible_factory(
-                self.removal_probabilities, data
-            ),
-            migration_rates=skyline_matrix_coercible_factory(
-                self.migration_rates, data
-            ),
-            birth_rates_among_demes=skyline_matrix_coercible_factory(
-                self.birth_rates_among_demes, data
-            ),
-        )
-        self._generate_one_from_extra_reactions(filename, rng, data, reactions)
-
-
-class EpidemiologicalReMASTERGenerator(ReMASTERGenerator):
-    parameterization: Literal[ParameterizationType.EPIDEMIOLOGICAL] = (
-        ParameterizationType.EPIDEMIOLOGICAL
-    )
-    reproduction_numbers: cfg.SkylineVectorCoercibleConfig = 0
-    become_uninfectious_rates: cfg.SkylineVectorCoercibleConfig = 0
-    sampling_proportions: cfg.SkylineVectorCoercibleConfig = 0
-    removal_probabilities: cfg.SkylineVectorCoercibleConfig = 0
-    migration_rates: cfg.SkylineMatrixCoercibleConfig = 0
-    reproduction_numbers_among_demes: cfg.SkylineMatrixCoercibleConfig = 0
-
-    def _generate_one(self, filename: str, rng: Generator, data: pgt.Data) -> None:
-        reactions = get_epidemiological_reactions(
-            populations=self.populations,
-            sample_population=self.sample_population,
-            reproduction_numbers=skyline_vector_coercible_factory(
-                self.reproduction_numbers, data
-            ),
-            become_uninfectious_rates=skyline_vector_coercible_factory(
-                self.become_uninfectious_rates, data
-            ),
-            sampling_proportions=skyline_vector_coercible_factory(
-                self.sampling_proportions, data
-            ),
-            removal_probabilities=skyline_vector_coercible_factory(
-                self.removal_probabilities, data
-            ),
-            migration_rates=skyline_matrix_coercible_factory(
-                self.migration_rates, data
-            ),
-            reproduction_numbers_among_demes=skyline_matrix_coercible_factory(
-                self.reproduction_numbers_among_demes, data
-            ),
-        )
-        self._generate_one_from_extra_reactions(filename, rng, data, reactions)
-
-
-class FBDReMASTERGenerator(ReMASTERGenerator):
-    parameterization: Literal[ParameterizationType.FBD] = ParameterizationType.FBD
-    diversification: cfg.SkylineVectorCoercibleConfig = 0
-    turnover: cfg.SkylineVectorCoercibleConfig = 0
-    sampling_proportions: cfg.SkylineVectorCoercibleConfig = 0
-    removal_probabilities: cfg.SkylineVectorCoercibleConfig = 0
-    migration_rates: cfg.SkylineMatrixCoercibleConfig = 0
-    diversification_between_types: cfg.SkylineMatrixCoercibleConfig = 0
-
-    def _generate_one(self, filename: str, rng: Generator, data: pgt.Data) -> None:
-        reactions = get_FBD_reactions(
-            populations=self.populations,
-            sample_population=self.sample_population,
-            diversification=skyline_vector_coercible_factory(
-                self.diversification, data
-            ),
-            turnover=skyline_vector_coercible_factory(self.turnover, data),
-            sampling_proportions=skyline_vector_coercible_factory(
-                self.sampling_proportions, data
-            ),
-            removal_probabilities=skyline_vector_coercible_factory(
-                self.removal_probabilities, data
-            ),
-            migration_rates=skyline_matrix_coercible_factory(
-                self.migration_rates, data
-            ),
-            diversification_between_types=skyline_matrix_coercible_factory(
-                self.diversification_between_types, data
-            ),
-        )
-        self._generate_one_from_extra_reactions(filename, rng, data, reactions)
-
-
-ReMASTERGeneratorConfig = Annotated[
-    CanonicalReMASTERGenerator
-    | EpidemiologicalReMASTERGenerator
-    | FBDReMASTERGenerator,
-    Field(discriminator="parameterization"),
-]