Merge pull request #22 from metagenlab/dev

Release 0.7.7

.github/workflows/unit-tests.yml

@@ -29,7 +29,7 @@ jobs:
       matrix:
         python-version: ["3.12"]
       fail-fast: false
-    name: Linux Python ${{ matrix.python-version }}
+    name: Linux
     steps:
       - uses: actions/checkout@v4
         with:
@@ -63,7 +63,7 @@ jobs:
         python-version: ["3.12"]
         architecture: ["x64"]
       fail-fast: false
-    name: OSX Python ${{ matrix.python-version }}
+    name: OSX
     steps:
       - uses: actions/checkout@v4
         with:

Dockerfile

@@ -3,7 +3,15 @@ LABEL org.opencontainers.image.source=https://github.com/metagenlab/assembly_fin
 LABEL org.opencontainers.image.description="Snakemake-powered cli to download genomes with NCBI datasets"
 LABEL org.opencontainers.image.licenses=MIT
 ENV LANG=C.UTF-8 
-ENV SHELL=/bin/bash 
+ENV SHELL=/bin/bash
+
+USER root
+ENV APT_PKGS="curl rsync p7zip-full procps"
+RUN apt-get update \
+    && apt-get install -y --no-install-recommends ${APT_PKGS} \
+    && apt-get clean \
+    && rm -rf /var/lib/apt /var/lib/dpkg /var/lib/cache /var/lib/log
+USER $MAMBA_USER
 
 COPY --chown=$MAMBA_USER:$MAMBA_USER . /pkg
 RUN micromamba config set extract_threads 1 && \

assembly_finder/assembly_finder.VERSION

@@ -1 +1 @@
-0.7.6
+0.7.7

assembly_finder/workflow/rules/download.smk

@@ -185,6 +185,7 @@ rule filter_genome_summaries:
     params:
         rank=RANK,
         nrank=NRANK,
+        taxon=TAXON,
     script:
         os.path.join(dir.scripts, "select_assemblies.py")
 

assembly_finder/workflow/rules/functions.smk

@@ -21,6 +21,8 @@ def convert_query(wildcards):
             ("GCF" not in wildcards.query) and ("GCA" not in wildcards.query)
         ):
             return wildcards.query.replace("_", " ")
+        else:
+            return wildcards.query
 
 
 def get_limit(wildcards, nbs, dic):
@@ -44,18 +46,21 @@ KEY = ""
 if API_KEY:
     KEY += f"--api-key {API_KEY} "
 ARGS = ""
-if ANNOTATED:
-    ARGS += "--annotated "
-if ASM_LVL:
-    ARGS += f"--assembly-level {ASM_LVL} "
-if SOURCE:
-    ARGS += f"--assembly-source {SOURCE} "
-if ATYPICAL:
-    ARGS += "--exclude-atypical "
-if MAG:
-    ARGS += f"--mag {MAG} "
-if REFERENCE:
-    ARGS += "--reference "
+if not TAXON:
+    ARGS = ""
+else:
+    if ANNOTATED:
+        ARGS += "--annotated "
+    if ASM_LVL:
+        ARGS += f"--assembly-level {ASM_LVL} "
+    if SOURCE:
+        ARGS += f"--assembly-source {SOURCE} "
+    if ATYPICAL:
+        ARGS += "--exclude-atypical "
+    if MAG:
+        ARGS += f"--mag {MAG} "
+    if REFERENCE:
+        ARGS += "--reference "
 GZIP = ""
 if COMPRESSED:
     GZIP = "--gzip"

assembly_finder/workflow/scripts/select_assemblies.py

@@ -16,6 +16,7 @@ def read_json(file):
 # Params
 rank = snakemake.params.rank
 nrank = snakemake.params.nrank
+taxon = snakemake.params.taxon
 
 # format summary column names
 summary_df.columns = (
@@ -73,6 +74,9 @@ def read_json(file):
     "species",
 ]
 
+if taxon:
+    tax_cols.insert(0, "taxon")
+
 df[tax_cols].to_csv(snakemake.output.tax, sep="\t", index=None)
 df[summary_df.columns].to_csv(snakemake.output.gen, sep="\t", index=None)
 df[["accession"]].drop_duplicates().to_csv(

env.yaml

@@ -6,9 +6,6 @@ dependencies:
   - ncbi-datasets-cli =16.26.2
   - taxonkit =0.17.0
   - csvtk =0.30.0
-  - rsync =3.3.0
-  - curl =8.7.1
-  - p7zip =16.02
   - python =3.12
   - pip
   - snakemake-minimal =8.18.1

tests/test_assembly_finder.py

@@ -42,15 +42,31 @@ def test_cli():
     exec_command("assembly_finder -v")
 
 
-def test_taxons():
-    """download genomes from taxons"""
+def test_taxon_string():
+    """download genomes from string input"""
+    exec_command(
+        f"assembly_finder --threads {threads} -i bacteria -nb 1 --output {outdir}"
+    )
+    remove_directory(outdir)
+
+
+def test_accession_string():
+    """download genomes from string input"""
+    exec_command(
+        f"assembly_finder --threads {threads} -i GCF_000418345.1 --accession --output {outdir}"
+    )
+    remove_directory(outdir)
+
+
+def test_taxons_file():
+    """download genomes from taxons file"""
     input = os.path.join(test_data_path, "taxons.tsv")
     exec_command(f"assembly_finder --threads {threads} -i {input} --output {outdir}")
     remove_directory(outdir)
 
 
-def test_accessions():
-    """download genomes from accessions"""
+def test_accessions_file():
+    """download genomes from accessions file"""
     input = os.path.join(test_data_path, "accessions.txt")
     exec_command(
         f"assembly_finder --threads {threads} -i {input} --accession --output {outdir}"