phylogenie 2.1.4__py3-none-any.whl → 2.1.6__py3-none-any.whl

phylogenie/generators/alisim.py

@@ -6,7 +6,7 @@ from typing import Any, Literal
 from numpy.random import Generator, default_rng
 
 from phylogenie.generators.dataset import DatasetGenerator, DataType
-from phylogenie.generators.factories import data
+from phylogenie.generators.factories import data, string
 from phylogenie.generators.trees import TreeDatasetGeneratorConfig
 from phylogenie.io import dump_newick
 
@@ -19,7 +19,7 @@ class AliSimDatasetGenerator(DatasetGenerator):
     trees: TreeDatasetGeneratorConfig
     keep_trees: bool = False
     iqtree_path: str = "iqtree2"
-    args: dict[str, str | int | float]
+    args: dict[str, Any]
 
     def _generate_one_from_tree(
         self, filename: str, tree_file: str, rng: Generator, data: dict[str, Any]
@@ -35,9 +35,7 @@ class AliSimDatasetGenerator(DatasetGenerator):
         ]
 
         for key, value in self.args.items():
-            command.extend(
-                [key, value.format(**data) if isinstance(value, str) else str(value)]
-            )
+            command.extend([key, string(value, data)])
 
         command.extend(["-af", "fasta"])
         subprocess.run(command, check=True, stdout=subprocess.DEVNULL)

phylogenie/generators/factories.py

@@ -1,3 +1,4 @@
+import re
 from typing import Any
 
 import numpy as np
@@ -53,6 +54,14 @@ def scalar(x: cfg.Scalar, data: dict[str, Any]) -> pgt.Scalar:
     return x
 
 
+def string(s: Any, data: dict[str, Any]) -> str:
+    if not isinstance(s, str):
+        return str(s)
+    return re.sub(
+        r"\{([^{}]+)\}", lambda match: str(_eval_expression(match.group(1), data)), s
+    )  # Match content inside curly braces
+
+
 def many_scalars(x: cfg.ManyScalars, data: dict[str, Any]) -> pgt.ManyScalars:
     if isinstance(x, str):
         e = _eval_expression(x, data)

phylogenie/generators/trees.py

@@ -46,7 +46,7 @@ class ParameterizationType(str, Enum):
 class TreeDatasetGenerator(DatasetGenerator):
     data_type: Literal[DataType.TREES] = DataType.TREES
     min_tips: cfg.Integer = 1
-    max_tips: cfg.Integer = 2**32
+    max_tips: cfg.Integer | None = None
     max_time: cfg.Scalar = np.inf
     init_state: str | None = None
     sampling_probability_at_present: cfg.Scalar = 0.0
@@ -64,7 +64,7 @@ class TreeDatasetGenerator(DatasetGenerator):
         return simulate_tree(
             events=self._get_events(data),
             min_tips=integer(self.min_tips, data),
-            max_tips=integer(self.max_tips, data),
+            max_tips=None if self.max_tips is None else integer(self.max_tips, data),
             max_time=scalar(self.max_time, data),
             init_state=init_state,
             sampling_probability_at_present=scalar(
@@ -98,8 +98,8 @@ class CanonicalTreeDatasetGenerator(TreeDatasetGenerator):
         ParameterizationType.CANONICAL
     )
     states: list[str]
-    sampling_rates: cfg.SkylineVector
-    remove_after_sampling: bool
+    sampling_rates: cfg.SkylineVector = 0
+    remove_after_sampling: bool = False
     birth_rates: cfg.SkylineVector = 0
     death_rates: cfg.SkylineVector = 0
     migration_rates: cfg.SkylineMatrix = None
@@ -122,7 +122,7 @@ class CanonicalTreeDatasetGenerator(TreeDatasetGenerator):
 class FBDTreeDatasetGenerator(TreeDatasetGenerator):
     parameterization: Literal[ParameterizationType.FBD] = ParameterizationType.FBD
     states: list[str]
-    sampling_proportions: cfg.SkylineVector
+    sampling_proportions: cfg.SkylineVector = 0
     diversification: cfg.SkylineVector = 0
     turnover: cfg.SkylineVector = 0
     migration_rates: cfg.SkylineMatrix = None
@@ -218,7 +218,7 @@ class BDTreeDatasetGenerator(TreeDatasetGeneratorForEpidemiology):
     parameterization: Literal[ParameterizationType.BD] = ParameterizationType.BD
     reproduction_number: cfg.SkylineParameter
     infectious_period: cfg.SkylineParameter
-    sampling_proportion: cfg.SkylineParameter = 1
+    sampling_proportion: cfg.SkylineParameter
 
     def _get_base_events(self, data: dict[str, Any]) -> list[Event]:
         return get_BD_events(
@@ -233,7 +233,7 @@ class BDEITreeDatasetGenerator(TreeDatasetGeneratorForEpidemiology):
     reproduction_number: cfg.SkylineParameter
     infectious_period: cfg.SkylineParameter
     incubation_period: cfg.SkylineParameter
-    sampling_proportion: cfg.SkylineParameter = 1
+    sampling_proportion: cfg.SkylineParameter
 
     def _get_base_events(self, data: dict[str, Any]) -> list[Event]:
         return get_BDEI_events(
@@ -250,7 +250,7 @@ class BDSSTreeDatasetGenerator(TreeDatasetGeneratorForEpidemiology):
     infectious_period: cfg.SkylineParameter
     superspreading_ratio: cfg.SkylineParameter
     superspreaders_proportion: cfg.SkylineParameter
-    sampling_proportion: cfg.SkylineParameter = 1
+    sampling_proportion: cfg.SkylineParameter
 
     def _get_base_events(self, data: dict[str, Any]) -> list[Event]:
         return get_BDSS_events(

phylogenie/treesimulator/events/core.py

@@ -64,8 +64,8 @@ class Sampling(Event):
 
 def get_canonical_events(
     states: list[str],
-    sampling_rates: SkylineVectorCoercible,
-    remove_after_sampling: bool,
+    sampling_rates: SkylineVectorCoercible = 0,
+    remove_after_sampling: bool = False,
     birth_rates: SkylineVectorCoercible = 0,
     death_rates: SkylineVectorCoercible = 0,
     migration_rates: SkylineMatrixCoercible | None = None,
@@ -100,7 +100,7 @@ def get_canonical_events(
 
 def get_FBD_events(
     states: list[str],
-    sampling_proportions: SkylineVectorCoercible = 1,
+    sampling_proportions: SkylineVectorCoercible = 0,
     diversification: SkylineVectorCoercible = 0,
     turnover: SkylineVectorCoercible = 0,
     migration_rates: SkylineMatrixCoercible | None = None,
@@ -134,7 +134,7 @@ def get_FBD_events(
 
 def get_epidemiological_events(
     states: list[str],
-    sampling_proportions: SkylineVectorCoercible = 1,
+    sampling_proportions: SkylineVectorCoercible,
     reproduction_numbers: SkylineVectorCoercible = 0,
     become_uninfectious_rates: SkylineVectorCoercible = 0,
     migration_rates: SkylineMatrixCoercible | None = None,
@@ -172,7 +172,7 @@ def get_epidemiological_events(
 def get_BD_events(
     reproduction_number: SkylineParameterLike,
     infectious_period: SkylineParameterLike,
-    sampling_proportion: SkylineParameterLike = 1,
+    sampling_proportion: SkylineParameterLike,
 ) -> list[Event]:
     return get_epidemiological_events(
         states=[INFECTIOUS_STATE],
@@ -186,7 +186,7 @@ def get_BDEI_events(
     reproduction_number: SkylineParameterLike,
     infectious_period: SkylineParameterLike,
     incubation_period: SkylineParameterLike,
-    sampling_proportion: SkylineParameterLike = 1,
+    sampling_proportion: SkylineParameterLike,
 ) -> list[Event]:
     return get_epidemiological_events(
         states=[EXPOSED_STATE, INFECTIOUS_STATE],
@@ -202,7 +202,7 @@ def get_BDSS_events(
     infectious_period: SkylineParameterLike,
     superspreading_ratio: SkylineParameterLike,
     superspreaders_proportion: SkylineParameterLike,
-    sampling_proportion: SkylineParameterLike = 1,
+    sampling_proportion: SkylineParameterLike,
 ) -> list[Event]:
     f_SS = superspreaders_proportion
     r_SS = superspreading_ratio

phylogenie/treesimulator/gillespie.py

@@ -11,20 +11,18 @@ from phylogenie.io import dump_newick
 from phylogenie.tree import Tree
 from phylogenie.treesimulator.model import Event, Model
 
-MAX_TIPS = 2**32
-
 
 def simulate_tree(
     events: Sequence[Event],
     min_tips: int = 1,
-    max_tips: int = MAX_TIPS,
+    max_tips: int | None = None,
     max_time: float = np.inf,
     init_state: str | None = None,
     sampling_probability_at_present: float = 0.0,
     seed: int | None = None,
     timeout: float = np.inf,
 ) -> Tree:
-    if max_time == np.inf and max_tips == MAX_TIPS:
+    if max_time == np.inf and max_tips is None:
         raise ValueError("Either max_time or max_tips must be specified.")
 
     if max_time == np.inf and sampling_probability_at_present:
@@ -49,7 +47,11 @@ def simulate_tree(
         current_time = 0.0
         change_times = sorted(set(t for e in events for t in e.rate.change_times))
         next_change_time = change_times.pop(0) if change_times else np.inf
-        target_n_tips = rng.integers(min_tips, max_tips) if max_time == np.inf else None
+        if max_time == np.inf:
+            assert max_tips is not None
+            target_n_tips = rng.integers(min_tips, max_tips + 1)
+        else:
+            target_n_tips = None
 
         while current_time < max_time:
             if time.perf_counter() - start_clock > timeout:
@@ -66,7 +68,8 @@ def simulate_tree(
 
             if (
                 not any(rates)
-                or model.n_sampled > max_tips
+                or max_tips is not None
+                and model.n_sampled > max_tips
                 or target_n_tips is not None
                 and model.n_sampled >= target_n_tips
             ):
@@ -89,7 +92,9 @@ def simulate_tree(
             if rng.random() < sampling_probability_at_present:
                 model.sample(individual, current_time, True)
 
-        if min_tips <= model.n_sampled <= max_tips:
+        if min_tips <= model.n_sampled and (
+            max_tips is None or model.n_sampled <= max_tips
+        ):
             return model.get_sampled_tree()
 
 
@@ -125,7 +130,7 @@ def generate_trees(
 
     if os.path.exists(output_dir):
         raise FileExistsError(f"Output directory {output_dir} already exists")
-    os.mkdir(output_dir)
+    os.makedirs(output_dir)
 
     rng = default_rng(seed)
     jobs = joblib.Parallel(n_jobs=n_jobs, return_as="generator_unordered")(

{phylogenie-2.1.4.dist-info → phylogenie-2.1.6.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: phylogenie
-Version: 2.1.4
+Version: 2.1.6
 Summary: Generate phylogenetic datasets with minimal setup effort
 Author: Gabriele Marino
 Author-email: gabmarino.8601@gmail.com

{phylogenie-2.1.4.dist-info → phylogenie-2.1.6.dist-info}/RECORD

@@ -1,10 +1,10 @@
 phylogenie/__init__.py,sha256=T2mRLsYtoLlWt8GlxrrUnfXJ9XVioq7hTvVq3uJpwQI,2215
 phylogenie/generators/__init__.py,sha256=zsOxy28-9j9alOQLIgrOAFfmM58NNHO_NEtW-KXQXAY,888
-phylogenie/generators/alisim.py,sha256=3mANgyQrlozhslV3_ryt-m4ItkRcKKRLufWf6SNBTnQ,2781
+phylogenie/generators/alisim.py,sha256=6IQEn8zNoVSLdHLRT1HBnrv1wfZzYaT-38SQNXE_MU4,2702
 phylogenie/generators/configs.py,sha256=AiiFS6rpH9BPwDKCkT4SVrRzfLFFrwRCJM4CRj0Srdk,1072
 phylogenie/generators/dataset.py,sha256=loVKC_1G7gzkPDN9W3GF-Rj9od8AeOJgIC0aJJa-4KA,2110
-phylogenie/generators/factories.py,sha256=jLwDuq0mrmDz2U5rZM19KJ2hSpamG3r6zb83YCc6snA,7619
-phylogenie/generators/trees.py,sha256=q03WPG82M4ucp-jyjoKEBy7TKMBzD3RkKn8hS0G0-i0,10463
+phylogenie/generators/factories.py,sha256=OKCNg9jwF2KahZCJKTdJVxhLTeLNBDMOqJs7XSVXTqY,7881
+phylogenie/generators/trees.py,sha256=KEDvyrfMjUPWwyp1xnhWpatNnjmDLPhSSQl6Msx19Ms,10508
 phylogenie/generators/typeguards.py,sha256=yj4VkhOaUXJ2OrY-6zhOeY9C4yKIQxjZtk2d-vIxttQ,828
 phylogenie/io.py,sha256=y7nQIvLgCvqELsXFKfm1GgKJO_saoQ-7zQpE3Kvajzc,3509
 phylogenie/main.py,sha256=vtvSpQxBNlYABoFQ25czl-l3fIr4QRo3svWVd-jcArw,1170
@@ -18,15 +18,15 @@ phylogenie/tree.py,sha256=KTx3m_tJPdeBqA5i0SA3dwCIobxsFcJwZFunMfDmVBY,2791
 phylogenie/treesimulator/__init__.py,sha256=XG_xwETKWgDmCihqNUFCcMHtFg4WvZu5qbqWn9Dndt8,879
 phylogenie/treesimulator/events/__init__.py,sha256=UGfvXOVJ_ZAkk_8sBPihjmxciiaEnXZEPFIY53sttWI,940
 phylogenie/treesimulator/events/contact_tracing.py,sha256=_nJ85yhgGkeruQgMHvGpDYoyhheBf8M4LgZWiWdi5dY,4801
-phylogenie/treesimulator/events/core.py,sha256=JokGmieAv2xEX7KsjBWZr05jHN1jB-XZbpxe9gwdbDA,7953
+phylogenie/treesimulator/events/core.py,sha256=RF7oHzAjkU675PnczaVc66d9gNrHBL-IhmVHtcy7MKE,7949
 phylogenie/treesimulator/events/mutations.py,sha256=xkXUIppbLIWZqKwVf-hi7d-_pS42TG2EPVfJA_grxBg,3443
-phylogenie/treesimulator/gillespie.py,sha256=naoxPyZixWVkd5f7B3KhEtOFiQI4NDIp_589NCLTHKM,4831
+phylogenie/treesimulator/gillespie.py,sha256=EfEbuMBQSk9izamBPaQ9rNKA4NvtrI3XVm891G2iMeM,5014
 phylogenie/treesimulator/model.py,sha256=0Im6cFTlpMlJrSP4pTTKtvLT9qrQWV8MSTesAsBxT8g,5422
 phylogenie/typeguards.py,sha256=JtqmbEWJZBRHbWgCvcl6nrWm3VcBfzRbklbTBYHItn0,1325
 phylogenie/typings.py,sha256=GknvAFXyiaWeeYJ8Lk5d6E2VHT-xW6ONEojYbtJYiB8,476
 phylogenie/utils.py,sha256=pCg9ob0RpLUHwM49x4knKxL4FNPr3-EU_6zMXsvxtAg,370
-phylogenie-2.1.4.dist-info/LICENSE.txt,sha256=NUrDqElK-eD3I0WqC004CJsy6cs0JgsAoebDv_42-pw,1071
-phylogenie-2.1.4.dist-info/METADATA,sha256=xp3v_zG0Cua0UQN6TlF7yS9NuchhdlKnDf8-O3e-l_8,5375
-phylogenie-2.1.4.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
-phylogenie-2.1.4.dist-info/entry_points.txt,sha256=Rt6_usN0FkBX1ZfiqCirjMN9FKOgFLG8rydcQ8kugeE,51
-phylogenie-2.1.4.dist-info/RECORD,,
+phylogenie-2.1.6.dist-info/LICENSE.txt,sha256=NUrDqElK-eD3I0WqC004CJsy6cs0JgsAoebDv_42-pw,1071
+phylogenie-2.1.6.dist-info/METADATA,sha256=dkSGsLeSFma9nDy-6J_gAHlOEs2NxjUeJvo4Ey4x5WM,5375
+phylogenie-2.1.6.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
+phylogenie-2.1.6.dist-info/entry_points.txt,sha256=Rt6_usN0FkBX1ZfiqCirjMN9FKOgFLG8rydcQ8kugeE,51
+phylogenie-2.1.6.dist-info/RECORD,,