Merge pull request #47 from gbouras13/dev

v1.6.0

Author: gbouras13

.github/workflows/ci.yaml

@@ -36,9 +36,13 @@ jobs:
         shell: bash -l {0}
         run: |
           mamba install python=${{ matrix.python-version }}
-          just install
           python -m pip install --upgrade pip
-          pip3 install git+https://github.com/rrwick/Unicycler.git
+          pip install -e .
+          pip install black
+          pip install isort
+          pip install pytest
+          pip install pytest-cov
+          pip install git+https://github.com/rrwick/Unicycler.git
           unicycler --help
       - name: Check formatting
         shell: bash -l {0}

HISTORY.md

@@ -1,5 +1,11 @@
 # History
 
+1.6.0 (2024-03-01)
+------------------
+
+* Addes `--depth_filter`. This will filter out all contigs that have long- (and short-read for `plassembler run`) read copy numbers that are less than the specified depth filter. Defaults to 0.25x.
+* Adds `--unicycler_options`  and `--spades_options` which allows passing extra Unicycler options (#46)
+
 1.5.1 (2024-02-01)
 ------------------
 

justfile

@@ -12,9 +12,6 @@ check-fmt:
     poetry run black --check .
     poetry run isort --check .
 
-# lint code with flake8
-lint:
-    poetry run flake8 .
 
 # install latest version with poetry
 install:

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "plassembler"
-version = "1.5.1" # change VERSION too
+version = "1.6.0" # change VERSION too
 description = "Quickly and accurately assemble plasmids in hybrid sequenced bacterial isolates"
 authors = ["George Bouras <george.bouras@adelaide.edu.au>"]
 license = "MIT"
@@ -32,7 +32,6 @@ black = ">=22.3.0"
 isort = ">=5.10.1"
 pytest = ">=6.2.5"
 pytest-cov = ">=3.0.0"
-flake8 = ">=3.0.1"
 
 [[tool.poetry.source]]
 name = "pypi-test"

src/plassembler/__init__.py

@@ -268,6 +268,24 @@ def common_options(func):
             type=click.Path(),
             default="nothing",
         ),
+        click.option(
+            "--depth_filter",
+            help="Filters all contigs low than this fraction of the chromosome read depth. Will apply on both long- and short-read sets for plassembler run.",
+            type=float,
+            default=0.25,
+        ),
+        click.option(
+            "--unicycler_options",
+            help='Extra Unicycler options - must be encapsulated by quotation marks if multiple "--no_rotate --mode conservative" ',
+            type=str,
+            default=None,
+        ),
+        click.option(
+            "--spades_options",
+            help='Extra spades options for Unicycler - must be encapsulated by quotation marks "--tmp-dir /tmp" ',
+            type=str,
+            default=None,
+        ),
     ]
     for option in reversed(options):
         func = option(func)
@@ -374,6 +392,9 @@ def run(
     flye_assembly,
     flye_info,
     no_chromosome,
+    depth_filter,
+    unicycler_options,
+    spades_options,
     **kwargs,
 ):
     """Runs Plassembler"""
@@ -401,6 +422,9 @@ def run(
     logger.info(f"--flye_info is {flye_info}")
     logger.info(f"--no_chromosome is {no_chromosome}")
     logger.info(f"--use_raven is {use_raven}")
+    logger.info(f"--depth_filter is {depth_filter}")
+    logger.info(f"--unicycler_options is {unicycler_options}")
+    logger.info(f"--spades_options is {spades_options}")
     logdir = Path(f"{outdir}/logs")
 
     # check deps
@@ -605,7 +629,14 @@ def run(
             unicycler_dir: Path = Path(outdir) / "unicycler_output"
 
             run_unicycler(
-                threads, logdir, short_r1, short_r2, long_reads, unicycler_dir
+                threads,
+                logdir,
+                short_r1,
+                short_r2,
+                long_reads,
+                unicycler_dir,
+                unicycler_options,
+                spades_options,
             )
 
             # check for successful unicycler completion
@@ -631,7 +662,7 @@ def run(
                 # processes output
                 plass.process_mash_tsv(database)
                 # combine depth and mash tsvs
-                plass.combine_depth_mash_tsvs(prefix)
+                plass.combine_depth_mash_tsvs(prefix, depth_filter)
 
                 # rename contigs and update copy bumber with plsdb
                 plass.finalise_contigs(prefix)
@@ -790,7 +821,14 @@ def run(
             unicycler_dir: Path = Path(outdir) / "unicycler_output"
 
             run_unicycler(
-                threads, logdir, short_r1, short_r2, long_reads, unicycler_dir
+                threads,
+                logdir,
+                short_r1,
+                short_r2,
+                long_reads,
+                unicycler_dir,
+                unicycler_options,
+                spades_options,
             )
 
             # check for successful unicycler completion
@@ -822,7 +860,7 @@ def run(
                 # processes output
                 plass.process_mash_tsv(database)
                 # combine depth and mash tsvs
-                plass.combine_depth_mash_tsvs(prefix)
+                plass.combine_depth_mash_tsvs(prefix, depth_filter)
 
                 # rename contigs and update copy bumber with plsdb
                 plass.finalise_contigs(prefix)
@@ -948,6 +986,7 @@ def assembled(
     logger.info(f"--skip_qc is {skip_qc}")
     logger.info(f"--pacbio_model is {pacbio_model}")
     logger.info(f"--no_copy_numbers is {no_copy_numbers}")
+    logger.info(f"--no_copy_numbers is {no_copy_numbers}")
     logdir = Path(f"{outdir}/logs")
 
     # check deps
@@ -1189,6 +1228,24 @@ def long_options(func):
             type=float,
             default=0.12,
         ),
+        click.option(
+            "--depth_filter",
+            help="Filters all contigs low than this fraction of the chromosome read depth. Will apply on both long- and short-read sets for plassembler run.",
+            type=float,
+            default=0.25,
+        ),
+        click.option(
+            "--unicycler_options",
+            help='Extra Unicycler options - must be encapsulated by quotation marks if multiple "--no_rotate --mode conservative" ',
+            type=str,
+            default=None,
+        ),
+        click.option(
+            "--spades_options",
+            help='Extra spades options for Unicycler - must be encapsulated by quotation marks "--tmp-dir /tmp" ',
+            type=str,
+            default=None,
+        ),
     ]
     for option in reversed(options):
         func = option(func)
@@ -1222,6 +1279,9 @@ def long(
     canu_flag,
     corrected_error_rate,
     no_chromosome,
+    depth_filter,
+    unicycler_options,
+    spades_options,
     **kwargs,
 ):
     """
@@ -1248,6 +1308,9 @@ def long(
     logger.info(f"--flye_info is {flye_info}")
     logger.info(f"--corrected_error_rate is {corrected_error_rate}")
     logger.info(f"--no_chromosome is {no_chromosome}")
+    logger.info(f"--depth_filter is {depth_filter}")
+    logger.info(f"--unicycler_options is {unicycler_options}")
+    logger.info(f"--spades_options is {spades_options}")
     logdir = Path(f"{outdir}/logs")
 
     # check deps
@@ -1509,7 +1572,13 @@ def long(
 
             unicycler_dir: Path = Path(outdir) / "unicycler_output"
             run_unicycler_long(
-                threads, logdir, corrected_fastqs, entropy_filtered_fastq, unicycler_dir
+                threads,
+                logdir,
+                corrected_fastqs,
+                entropy_filtered_fastq,
+                unicycler_dir,
+                unicycler_options,
+                spades_options,
             )
             remove_file(corrected_fastqs)
             assembled_fasta = os.path.join(outdir, "unicycler_output", "assembly.fasta")
@@ -1557,7 +1626,7 @@ def long(
             plass.process_mash_tsv(database)
 
             # combine depth and mash tsvs
-            plass.combine_depth_mash_tsvs(prefix)
+            plass.combine_depth_mash_tsvs(prefix, depth_filter)
 
             # rename contigs and update copy number with plsdb
             plass.finalise_contigs_long(prefix)

src/plassembler/utils/VERSION

@@ -1 +1 @@
-1.5.1
+1.6.0