phylogenie 1.0.5__tar.gz → 1.0.7__tar.gz

{phylogenie-1.0.5 → phylogenie-1.0.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: phylogenie
-Version: 1.0.5
+Version: 1.0.7
 Summary: Generate phylogenetic datasets with minimal setup effort
 Author: Gabriele Marino
 Author-email: gabmarino.8601@gmail.com
@@ -46,7 +46,7 @@ Phylogenie comes packed with useful features, including:
   Simply specify the number of cores to use, and Phylogenie handles multiprocessing automatically.
 
 - **Pre-implemented parameterizations** 🎯  
-  Include canonical, fossilized birth-death, epidemiological, birth-death with exposed-infectious (BDEI), contact-tracing (CT), and more.
+  Include canonical, fossilized birth-death, epidemiological, birth-death with exposed-infectious (BDEI), birth-death with superspreading (BDSS), contact-tracing (CT), and more.
 
 - **Skyline parameter support** 🪜  
   Support for piece-wise constant parameters.

{phylogenie-1.0.5 → phylogenie-1.0.7}/README.md

@@ -27,7 +27,7 @@ Phylogenie comes packed with useful features, including:
   Simply specify the number of cores to use, and Phylogenie handles multiprocessing automatically.
 
 - **Pre-implemented parameterizations** 🎯  
-  Include canonical, fossilized birth-death, epidemiological, birth-death with exposed-infectious (BDEI), contact-tracing (CT), and more.
+  Include canonical, fossilized birth-death, epidemiological, birth-death with exposed-infectious (BDEI), birth-death with superspreading (BDSS), contact-tracing (CT), and more.
 
 - **Skyline parameter support** 🪜  
   Support for piece-wise constant parameters.

{phylogenie-1.0.5 → phylogenie-1.0.7}/phylogenie/backend/treesimulator.py

@@ -17,15 +17,16 @@ from phylogenie.skyline import (
 DEFAULT_POPULATION = "X"
 INFECTIOUS_POPULATION = "I"
 EXPOSED_POPULATION = "E"
+SUPERSPREADER_POPULATION = "S"
 
 
 @dataclass
 class TreeParams:
-    populations: str | list[str] = DEFAULT_POPULATION
+    populations: str | list[str]
+    transmission_rates: SkylineMatrixCoercible
+    removal_rates: SkylineVectorCoercible
+    sampling_proportions: SkylineVectorCoercible
     transition_rates: SkylineMatrixCoercible = 0
-    transmission_rates: SkylineMatrixCoercible = 0
-    removal_rates: SkylineVectorCoercible = 0
-    sampling_proportions: SkylineVectorCoercible = 0
 
 
 def generate_tree(
@@ -100,6 +101,7 @@ def generate_tree(
     )
     for i, leaf in enumerate(tree.iter_leaves()):
         state: str = getattr(leaf, STATE)
+        delattr(leaf, STATE)
         date = tree.get_distance(leaf)
         leaf.name = f"{i}|{state}|{date}"
     save_forest([tree], output_file)
@@ -137,3 +139,25 @@ def get_BDEI_params(
         removal_rates=removal_rates,
         sampling_proportions=sampling_proportions,
     )
+
+
+def get_BDSS_params(
+    reproduction_number: SkylineParameterLike,
+    infectious_period: SkylineParameterLike,
+    superspreading_ratio: SkylineParameterLike,
+    superspreaders_proportion: SkylineParameterLike,
+    sampling_proportion: SkylineParameterLike,
+) -> TreeParams:
+    gamma = 1 / infectious_period
+    f_SS = superspreaders_proportion
+    r_SS = superspreading_ratio
+    lambda_IS = reproduction_number * gamma * f_SS / (1 + r_SS * f_SS - f_SS)
+    lambda_SI = (reproduction_number * gamma - r_SS * lambda_IS) * r_SS
+    lambda_SS = r_SS * lambda_IS
+    lambda_II = lambda_SI / r_SS
+    return TreeParams(
+        populations=[INFECTIOUS_POPULATION, SUPERSPREADER_POPULATION],
+        transmission_rates=[[lambda_II, lambda_IS], [lambda_SI, lambda_SS]],
+        removal_rates=gamma,
+        sampling_proportions=sampling_proportion,
+    )

{phylogenie-1.0.5 → phylogenie-1.0.7}/phylogenie/core/msas/base.py

@@ -1,6 +1,7 @@
 import os
 from abc import abstractmethod
 from enum import Enum
+from pathlib import Path
 from typing import Literal
 
 from numpy.random import Generator
@@ -14,9 +15,14 @@ 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(
@@ -24,11 +30,22 @@ class MSAsGenerator(DatasetGenerator):
     ) -> None: ...
 
     def _generate_one(self, filename: str, rng: Generator, data: pgt.Data) -> None:
-        tree_filename = f"{filename}.temp-tree"
+        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=filename, tree_file=f"{tree_filename}.nwk", rng=rng, data=data
+            filename=msa_filename, tree_file=f"{tree_filename}.nwk", rng=rng, data=data
         )
-        os.remove(f"{tree_filename}.nwk")
+        if not self.keep_trees:
+            os.remove(f"{tree_filename}.nwk")

{phylogenie-1.0.5 → phylogenie-1.0.7}/phylogenie/core/trees/treesimulator.py

@@ -13,6 +13,7 @@ from phylogenie.backend.treesimulator import (
     generate_tree,
     get_BD_params,
     get_BDEI_params,
+    get_BDSS_params,
 )
 from phylogenie.core.factories import (
     int_factory,
@@ -28,6 +29,7 @@ class ParameterizationType(str, Enum):
     MTBD = "MTBD"
     BD = "BD"
     BDEI = "BDEI"
+    BDSS = "BDSS"
 
 
 class TreeSimulatorGenerator(TreesGenerator):
@@ -73,10 +75,10 @@ class TreeSimulatorGenerator(TreesGenerator):
 class MTBDTreeSimulatorGenerator(TreeSimulatorGenerator):
     parameterization: Literal[ParameterizationType.MTBD] = ParameterizationType.MTBD
     populations: str | list[str] = DEFAULT_POPULATION
+    transmission_rates: cfg.SkylineMatrixCoercibleConfig
+    removal_rates: cfg.SkylineVectorCoercibleConfig
     transition_rates: cfg.SkylineMatrixCoercibleConfig = 0
-    transmission_rates: cfg.SkylineMatrixCoercibleConfig = 0
-    removal_rates: cfg.SkylineVectorCoercibleConfig = 0
-    sampling_proportions: cfg.SkylineVectorCoercibleConfig = 0
+    sampling_proportions: cfg.SkylineVectorCoercibleConfig = 1
 
     def _generate_one(self, filename: str, rng: Generator, data: pgt.Data) -> None:
         self._generate_one_from_params(
@@ -103,9 +105,9 @@ class MTBDTreeSimulatorGenerator(TreeSimulatorGenerator):
 
 class BDTreeSimulatorGenerator(TreeSimulatorGenerator):
     parameterization: Literal[ParameterizationType.BD] = ParameterizationType.BD
-    reproduction_number: cfg.SkylineParameterLikeConfig = 0
-    infectious_period: cfg.SkylineParameterLikeConfig = 0
-    sampling_proportion: cfg.SkylineParameterLikeConfig = 0
+    reproduction_number: cfg.SkylineParameterLikeConfig
+    infectious_period: cfg.SkylineParameterLikeConfig
+    sampling_proportion: cfg.SkylineParameterLikeConfig = 1
 
     def _generate_one(self, filename: str, rng: Generator, data: pgt.Data) -> None:
         self._generate_one_from_params(
@@ -128,10 +130,10 @@ class BDTreeSimulatorGenerator(TreeSimulatorGenerator):
 
 class BDEITreeSimulatorGenerator(TreeSimulatorGenerator):
     parameterization: Literal[ParameterizationType.BDEI] = ParameterizationType.BDEI
-    reproduction_number: cfg.SkylineParameterLikeConfig = 0
-    infectious_period: cfg.SkylineParameterLikeConfig = 0
-    incubation_period: cfg.SkylineParameterLikeConfig = 0
-    sampling_proportion: cfg.SkylineParameterLikeConfig = 0
+    reproduction_number: cfg.SkylineParameterLikeConfig
+    infectious_period: cfg.SkylineParameterLikeConfig
+    incubation_period: cfg.SkylineParameterLikeConfig
+    sampling_proportion: cfg.SkylineParameterLikeConfig = 1
 
     def _generate_one(self, filename: str, rng: Generator, data: pgt.Data) -> None:
         self._generate_one_from_params(
@@ -155,7 +157,43 @@ class BDEITreeSimulatorGenerator(TreeSimulatorGenerator):
         )
 
 
+class BDSSTreeSimulatorGenerator(TreeSimulatorGenerator):
+    parameterization: Literal[ParameterizationType.BDSS] = ParameterizationType.BDSS
+    reproduction_number: cfg.SkylineParameterLikeConfig
+    infectious_period: cfg.SkylineParameterLikeConfig
+    superspreading_ratio: cfg.SkylineParameterLikeConfig
+    superspreaders_proportion: cfg.SkylineParameterLikeConfig
+    sampling_proportion: cfg.SkylineParameterLikeConfig = 1
+
+    def _generate_one(self, filename: str, rng: Generator, data: pgt.Data) -> None:
+        self._generate_one_from_params(
+            filename,
+            rng,
+            data,
+            get_BDSS_params(
+                reproduction_number=skyline_parameter_like_factory(
+                    self.reproduction_number, data
+                ),
+                infectious_period=skyline_parameter_like_factory(
+                    self.infectious_period, data
+                ),
+                superspreading_ratio=skyline_parameter_like_factory(
+                    self.superspreading_ratio, data
+                ),
+                superspreaders_proportion=skyline_parameter_like_factory(
+                    self.superspreaders_proportion, data
+                ),
+                sampling_proportion=skyline_parameter_like_factory(
+                    self.sampling_proportion, data
+                ),
+            ),
+        )
+
+
 TreeSimulatorGeneratorConfig = Annotated[
-    MTBDTreeSimulatorGenerator | BDTreeSimulatorGenerator | BDEITreeSimulatorGenerator,
+    MTBDTreeSimulatorGenerator
+    | BDTreeSimulatorGenerator
+    | BDEITreeSimulatorGenerator
+    | BDSSTreeSimulatorGenerator,
     Field(discriminator="parameterization"),
 ]

{phylogenie-1.0.5 → phylogenie-1.0.7}/phylogenie/typings.py

@@ -21,4 +21,3 @@ Vector2D = list[Vector1D]
 Vector3D = list[Vector2D]
 
 Data = dict[str, str | Scalar | Vector1D | Vector2D | Vector3D]
-Size = int | tuple[int, int] | tuple[int, int, int]

{phylogenie-1.0.5 → phylogenie-1.0.7}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "phylogenie"
-version = "1.0.5"
+version = "1.0.7"
 description = "Generate phylogenetic datasets with minimal setup effort"
 authors = ["Gabriele Marino <gabmarino.8601@gmail.com>"]
 readme = "README.md"