Merge pull request #111 from faridrashidi/develop

release v0.0.21

.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.0.20
+current_version = 0.0.21
 commit = True
 message = [skip ci] {current_version} → {new_version}
 tag = False

.gitattributes

@@ -0,0 +1,13 @@
+*.py linguist-detectable=true
+*Makefile linguist-detectable=false
+*.md linguist-detectable=false
+*.h linguist-detectable=false
+*.cpp linguist-detectable=false
+*.rst linguist-detectable=false
+*.yaml linguist-detectable=false
+*.yml linguist-detectable=false
+*.cfg linguist-detectable=false
+*.txt linguist-detectable=false
+*.pyx linguist-detectable=false
+*.R linguist-detectable=false
+*.sh linguist-detectable=false

.pre-commit-config.yaml

@@ -7,7 +7,7 @@ default_stages:
 minimum_pre_commit_version: 2.9.3
 repos:
 -   repo: https://github.com/psf/black
-    rev: 21.11b1
+    rev: 22.3.0
     hooks:
     -   id: black
         additional_dependencies: [toml]
@@ -19,7 +19,7 @@ repos:
         additional_dependencies: [toml]
         args: [--order-by-type, --profile, 'black']
 -   repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v4.0.1
+    rev: v4.2.0
     hooks:
     -   id: detect-private-key
     -   id: check-merge-conflict
@@ -40,22 +40,22 @@ repos:
     -   id: check-toml
     -   id: requirements-txt-fixer
 -   repo: https://github.com/jumanjihouse/pre-commit-hooks
-    rev: 2.1.5
+    rev: 2.1.6
     hooks:
     -   id: script-must-have-extension
         name: Check executable files use .sh extension
         types: [shell, executable]
 -   repo: https://github.com/myint/rstcheck
-    rev: 3f92957478422df87bd730abde66f089cc1ee19b
+    rev: v5.0.0
     hooks:
     -   id: rstcheck
 -   repo: https://github.com/asottile/blacken-docs
-    rev: v1.12.0
+    rev: v1.12.1
     hooks:
     -   id: blacken-docs
         additional_dependencies: [black==20.8b1]
 -   repo: https://github.com/asottile/pyupgrade
-    rev: v2.29.1
+    rev: v2.32.0
     hooks:
     -   id: pyupgrade
         args: [--py3-plus, --py37-plus]
@@ -69,7 +69,7 @@ repos:
     -   id: rst-directive-colons
     -   id: rst-inline-touching-normal
 -   repo: https://github.com/PyCQA/doc8
-    rev: 0.10.1
+    rev: 0.11.1
     hooks:
     -   id: doc8
         args: [--max-line-length=88]

.readthedocs.yml

@@ -1,18 +1,21 @@
 version: 2
 
+build:
+    os: ubuntu-20.04
+    tools:
+        python: "3.9"
+    apt_packages:
+    -   pandoc
+
 sphinx:
     builder: html
     configuration: docs/source/conf.py
     fail_on_warning: false
 
-formats:
--   htmlzip
--   pdf
-
-build:
-    image: latest
-
 python:
-    version: 3.7
     install:
     -   requirements: docs/requirements_rtfd.txt
+
+formats:
+-   htmlzip
+-   pdf

README.rst

@@ -27,8 +27,7 @@ Please see our `preprint`_ on **bioRxiv** to learn more.
 
 Support
 -------
-Feel free to submit an `issue <https://github.com/faridrashidi/trisicell/issues/new/choose>`_
-or send us an `email <mailto:[email protected]>`_.
+Feel free to submit an `issue <https://github.com/faridrashidi/trisicell/issues/new/choose>`_.
 Your help to improve Trisicell is highly appreciated.
 
 Trisicell was developed in collaboration between the `Cancer Data Science Laboratory (CDSL) <https://ccr.cancer.gov/staff-directory/s-cenk-sahinalp>`_ and the `Laboratory of Cancer Biology and Genetics (LCBG) <https://ccr.cancer.gov/staff-directory/glenn-merlino>`_ at the `National Cancer Institute (NCI) <https://www.cancer.gov>`_.

docs/source/citing.rst

@@ -23,6 +23,6 @@ BibTeX
       month         = mar,
       publisher     = {Cold Spring Harbor Laboratory},
       author        = {Farid Rashidi Mehrabadi and Kerrie L. Marie and Eva Perez-Guijarro and Salem Malikic and Erfan Sadeqi Azer and Howard H. Yang and Can Kizilkale and Charli Gruen and Wells Robinson and Huaitian Liu and Michael C. Kelly and Christina Marcelus and Sandra Burkett and Aydin Buluc and Funda Ergun and Maxwell P. Lee and Glenn Merlino and Chi-Ping Day and S. Cenk Sahinalp},
-      title         = {{Profiles of expressed mutations in single cells reveal subclonal expansion patterns and therapeutic impact of intratumor heterogeneity}}
+      title         = {{Profiles of expressed mutations in single cells reveal subclonal expansion patterns and therapeutic impact of intratumor heterogeneity}},
       journal       = {bioRxiv}
     }

docs/source/conf.py

@@ -11,7 +11,6 @@
 #
 import os
 import sys
-from datetime import datetime
 from pathlib import Path
 
 from pybtex.plugin import register_plugin
@@ -46,7 +45,7 @@
 title = (
     "Scalable intratumor heterogeneity inference and validation from single-cell data"
 )
-copyright = f"{datetime.now():%Y}, {author}"
+copyright = f"2021-2022, {author}"
 release = "master"
 version = f"master ({trisicell.__version__})"
 

docs/source/index.rst

@@ -9,7 +9,6 @@
    installation
    release_notes
    citing
-   references
 
 .. toctree::
    :caption: Docs

docs/source/release_notes.rst

@@ -7,6 +7,15 @@ Release Notes
 =============
 
 
+Version 0.0.21 :small:`April 22, 2022`
+-----------------------------------------
+
+This version includes:
+
+    - Fix some bugs.
+    - Improve the documentations.
+
+
 Version 0.0.20 :small:`November 22, 2021`
 -----------------------------------------
 

examples/reconstruction/compute_booster.py

@@ -8,6 +8,8 @@
 
 import trisicell as tsc
 
+# sphinx_gallery_thumbnail_path = "_static/thumbnails/trisicell-boost.png"
+
 # %%
 # First, we load a binary test single-cell genotype data.
 df_in = tsc.datasets.test()

setup.cfg

@@ -19,6 +19,7 @@ ignore =
     D107, # D107 Missing docstring in __init__
     W503, # W503 line break before binary operator
     D202, # D202 No blank lines allowed after function docstring
+    B020, # B020 Found for loop that reassigns the iterable it is iterating with each
     B902, # B902 blind except Exception: statement
     C400, # C400 Unnecessary generator - rewrite as a list comprehension
 [tool:pytest]

setup.py

@@ -12,7 +12,7 @@
     __author__ = ", ".join(["Farid Rashidi"])
     __maintainer__ = ", ".join(["Farid Rashidi"])
     __email__ = ", ".join(["[email protected]"])
-    __version__ = "0.0.20"
+    __version__ = "0.0.21"
 
 if platform == "linux" or platform == "linux2":
     os.environ["CC"] = "g++"

tests/test_commands.py

@@ -57,10 +57,9 @@ def test_scite_experiment(self):
                 tsc.ul.get_file("trisicell.datasets/test/test.tsv"),
                 "0.0000001",
                 "0.1",
-                "-r 3",
-                "-l 1000",
                 "-e",
-                "-h 0.005",
+                "-t 1",
+                "-s 1",
             ],
         )
         assert result.exit_code == 0

tests/test_logging.py

@@ -1,3 +1,5 @@
+import pytest
+
 import trisicell as tsc
 
 
@@ -9,4 +11,6 @@ def test_logging(self):
         tsc.logg.info("INFO")
         tsc.logg.warn("WARN")
         tsc.logg.info("TIME", time=True, color="red")
+        with pytest.raises(RuntimeError):
+            tsc.logg.error("ERROR")
         assert True

tests/test_pp.py

@@ -41,3 +41,10 @@ def test_readcount(self):
         tsc.pp.remove_cell_by_list(adata, ["C15_1"])
         tsc.pp.keep_cell_by_list(adata, ["C15_2", "C15_3"])
         assert adata.shape == (2, 267)
+
+    def test_local_cluster_cells_then_merge_muts_pseudo_bulk(self):
+        geno = tsc.datasets.example()
+        geno_merged, geno = tsc.pp.local_cluster_cells_then_merge_muts_pseudo_bulk(
+            geno, by="mut", n_clusters=11, min_n_cells=2, attr="group"
+        )
+        assert geno_merged.shape == (11, 452)

tests/test_tl_scores.py

@@ -11,6 +11,10 @@ def setup_method(self):
         self.grnd = tsc.io.read(f1)
         self.sol = tsc.io.read(f2)
 
+    def test_gs(self):
+        gs = tsc.tl.gs(self.grnd, self.sol)
+        assert np.abs(gs - 0.9895) < 0.0001
+
     def test_ad(self):
         ad = tsc.tl.ad(self.grnd, self.sol)
         assert np.abs(ad - 0.9778) < 0.0001

tests/test_tl_solvers.py

@@ -32,13 +32,13 @@ def test_phiscsb(self):
         df_out = tsc.tl.phiscsb(self.df_in, alpha=0.0000001, beta=0.1)
         assert tsc.ul.is_conflict_free_gusfield(df_out)
 
-    def test_huntress_both(self):
-        df_out = tsc.tl.huntress(self.df_in, alpha=0.0000001, beta=0.1, kind="both")
-        assert tsc.ul.is_conflict_free_gusfield(df_out)
+    # def test_huntress_both(self):
+    #     df_out = tsc.tl.huntress(self.df_in, alpha=0.0000001, beta=0.1)
+    #     assert tsc.ul.is_conflict_free_gusfield(df_out)
 
-    def test_huntress_fn(self):
-        df_out = tsc.tl.huntress(self.df_in, alpha=0, beta=0, kind="fn")
-        assert tsc.ul.is_conflict_free_gusfield(df_out)
+    # def test_huntress_fn(self):
+    #     df_out = tsc.tl.huntress(self.df_in, alpha=0, beta=0.1)
+    #     assert tsc.ul.is_conflict_free_gusfield(df_out)
 
     @skip_rpy2
     def test_onconem(self):

trisicell/__init__.py

@@ -7,5 +7,5 @@
 __author__ = ", ".join(["Farid Rashidi"])
 __maintainer__ = ", ".join(["Farid Rashidi"])
 __email__ = ", ".join(["[email protected]"])
-__version__ = "0.0.20"
+__version__ = "0.0.21"
 __all__ = (datasets, io, logg, pl, pp, settings, tl, ul)

trisicell/commands/_bnb.py

@@ -35,7 +35,7 @@ def bnb(genotype_file, bounding, time_limit):
     A fast branch and bound algorithm for
     the perfect tumor phylogeny reconstruction problem :cite:`PhISCS-BnB`.
 
-    trisicell bnb input.SC 0.0001 0.1 -b real
+    trisicell bnb input.SC -b real
     """
 
     outfile = os.path.splitext(genotype_file)[0]

trisicell/commands/_gpps.py

@@ -0,0 +1,115 @@
+import os
+
+import click
+
+import trisicell as tsc
+
+
+@click.command(short_help="Run gpps.")
+@click.argument(
+    "genotype_file",
+    required=True,
+    type=click.Path(
+        exists=True, file_okay=True, dir_okay=False, readable=True, resolve_path=True
+    ),
+)
+@click.argument(
+    "alpha",
+    required=True,
+    type=float,
+)
+@click.argument(
+    "beta",
+    required=True,
+    type=float,
+)
+@click.option(
+    "--k_dollo",
+    "-k",
+    default=0,
+    type=int,
+    show_default=True,
+    help="k-Dollo.",
+)
+@click.option(
+    "--max_del",
+    "-d",
+    default=-1,
+    type=int,
+    show_default=True,
+    help="Maximum number of deletion allowed.",
+)
+@click.option(
+    "--neighbor_size",
+    "-s",
+    default=30,
+    type=int,
+    show_default=True,
+    help="Hill climbing neighborhood size.",
+)
+@click.option(
+    "--n_iters",
+    "-l",
+    default=100,
+    type=int,
+    show_default=True,
+    help="Hill climbing maximum iterations.",
+)
+@click.option(
+    "--time_limit",
+    "-t",
+    default=86400,
+    type=int,
+    show_default=True,
+    help="Time limit of the program (in second).",
+)
+@click.option(
+    "--n_threads",
+    "-p",
+    default=1,
+    type=int,
+    show_default=True,
+    help="Number of threads.",
+)
+def gpps(
+    genotype_file,
+    alpha,
+    beta,
+    k_dollo,
+    max_del,
+    neighbor_size,
+    n_iters,
+    time_limit,
+    n_threads,
+):
+    """gpps.
+
+    an ILP-based approach for inferring cancer progression with mutation losses from
+    single cell data :cite:`gpps`.
+
+    trisicell gpps input.SC 0.0001 0.1 -k 0 -s 30 -l 100 -t 86400 -p 1
+    """
+
+    outfile = os.path.splitext(genotype_file)[0]
+
+    tsc.settings.verbosity = "info"
+    tsc.settings.logfile = f"{outfile}.gpps.log"
+
+    df_in = tsc.io.read(genotype_file)
+
+    if alpha == 0:
+        alpha = 0.000000000001
+    df_out = tsc.tl.gpps(
+        df_in,
+        alpha=alpha,
+        beta=beta,
+        k_dollo=k_dollo,
+        max_del=max_del,
+        neighbor_size=neighbor_size,
+        n_iters=n_iters,
+        time_limit=time_limit,
+        n_threads=n_threads,
+    )
+    tsc.io.write(df_out, f"{outfile}.gpps.CFMatrix")
+
+    return None