bump to version v1.4.0

CHANGELOG.md

@@ -2,7 +2,9 @@
 
 ## 1.4.0 
 
-
+- added:
+    - `bactopia-pubmlst-setup` to setup PubMLST REST API connections
+    - `bactopia-pubmlst-build` to build PubMLST databases compatible with `mlst` Bactopia Tool
 
 ## 1.3.0
 

README.md

@@ -9,14 +9,18 @@ A Python package for working with [Bactopia](https://bactopia.github.io/)
 
 There are many subcommands available in Bactopia. Here is a brief description of each command:
 
-| Command              | Description                                                                |
-|----------------------|----------------------------------------------------------------------------|
-| `bactopia-citations` | Print out tools and citations used throughout Bactopia                     |
-| `bactopia-datasets`  | Download optional datasets to supplement your analyses with Bactopia       |
-| `bactopia-download`  | Builds Bactopia environments for use with Nextflow.                        |
-| `bactopia-prepare`   | Create a 'file of filenames' (FOFN) of samples to be processed by Bactopia |
-| `bactopia-search`    | Query against ENA and SRA for public accessions to process with Bactopia   |
-| `bactopia-summary`   | Generate a summary table from the Bactopia results.                        |
+| Command                   | Description                                                                |
+|---------------------------|----------------------------------------------------------------------------|
+| `bactopia-citations`      | Print out tools and citations used throughout Bactopia                     |
+| `bactopia-datasets`       | Download optional datasets to supplement your analyses with Bactopia       |
+| `bactopia-download`       | Builds Bactopia environments for use with Nextflow.                        |
+| `bactopia-prepare`        | Create a 'file of filenames' (FOFN) of samples to be processed by Bactopia |
+| `bactopia-search`         | Query against ENA and SRA for public accessions to process with Bactopia   |
+| `bactopia-summary`        | Generate a summary table from the Bactopia results.                        |
+| `bactopia-atb-downloader` | Download assemblies from AllTheBacteria for use with Bactopia Tools        |
+| `bactopia-atb-formatter`  | Restructure All-the-Bacteria assemblies to allow usage with Bactopia Tools |
+| `bactopia-pubmlst-setup`  | One-time setup for interacting with the PubMLST API                        |
+| `bactopia-pubmlst-build`  | Build PubMLST databases for use with the 'mlst' Bactopia Tool.             |
 
 Below is the `--help` output for each subcommand.
 

bactopia/cli/pubmlst/build.py

@@ -13,6 +13,7 @@
     available_databases,
     build_blast_db,
     download_database,
+    print_citation,
 )
 
 # Set up Rich
@@ -21,7 +22,7 @@
 click.rich_click.USE_RICH_MARKUP = True
 click.rich_click.OPTION_GROUPS = {
     # Use underscores in parameters, since these are also passed to Nextflow
-    "bactopia-pubmlst-download": [
+    "bactopia-pubmlst-build": [
         {
             "name": "Required Options",
             "options": [
@@ -63,9 +64,7 @@
 @click.option(
     "--database",
     "-d",
-    default="pubmlst_yersinia_seqdef",
-    show_default=True,
-    help="The organism database to interact with for setup. (Use 'all' to download all databases.)",
+    help="A known organism database to download. (Use 'all' to download all databases.)",
 )
 @click.option(
     "--site",
@@ -78,7 +77,7 @@
 @click.option(
     "--token-dir",
     "-t",
-    default=".bactopia/pubmlst",
+    default=f"{Path.home()}/.bactopia",
     show_default=True,
     help="The directory where the token file is saved.",
 )
@@ -114,7 +113,7 @@ def pubmlst_download(
     verbose: bool,
     silent: bool,
 ):
-    """Download specific database files from PubMLST or Pasteur."""
+    """Build PubMLST databases for use with the 'mlst' Bactopia Tool."""
     # Setup logs
     logging.basicConfig(
         format="%(asctime)s:%(name)s:%(levelname)s - %(message)s",
@@ -184,9 +183,11 @@ def pubmlst_download(
     if not skip_blast:
         # Build MLST database
         build_blast_db(
-            f"{out_dir}/mlstdb",
+            out_dir,
         )
 
+    print_citation()
+
 
 def main():
     if len(sys.argv) == 1:

bactopia/cli/pubmlst/setup.py

@@ -9,7 +9,7 @@
 from rich.logging import RichHandler
 
 import bactopia
-from bactopia.databases.pubmlst.utils import setup_pubmlst
+from bactopia.databases.pubmlst.utils import print_citation, setup_pubmlst
 
 # Set up Rich
 stderr = rich.console.Console(stderr=True)
@@ -58,6 +58,7 @@
 @click.option(
     "--site",
     "-s",
+    default="pubmlst",
     show_default=True,
     type=click.Choice(["pubmlst", "pasteur"], case_sensitive=True),
     help="Only print citation matching a given name",
@@ -72,7 +73,7 @@
 @click.option(
     "--save-dir",
     "-sd",
-    default=".bactopia/pubmlst",
+    default=f"{Path.home()}/.bactopia",
     show_default=True,
     help="The directory to save the token",
 )
@@ -95,7 +96,14 @@ def pubmlst_setup(
         format="%(asctime)s:%(name)s:%(levelname)s - %(message)s",
         datefmt="%Y-%m-%d %H:%M:%S",
         handlers=[
-            RichHandler(rich_tracebacks=True, console=rich.console.Console(stderr=True))
+            RichHandler(
+                rich_tracebacks=True,
+                console=rich.console.Console(stderr=True),
+                show_time=True if verbose else False,
+                show_level=True if verbose else False,
+                show_path=True,
+                markup=True,
+            )
         ],
     )
     logging.getLogger().setLevel(
@@ -114,6 +122,7 @@ def pubmlst_setup(
         )
         sys.exit(1)
     setup_pubmlst(site, database, token_file, client_id, client_secret)
+    print_citation()
 
 
 def main():

bactopia/databases/pubmlst/utils.py

@@ -16,6 +16,21 @@
 from bactopia.utils import execute
 
 
+def print_citation():
+    """
+    Print the citation for PubMLST and Bactopia
+    """
+    logging.info(
+        "\n\n"
+        "If you make use of 'bactopia-pubmlst-(setup|build)' commands and 'PubMLST' database files, please cite the following:\n\n"
+        "'PubMLST'\n"
+        "Jolley, K. A., Bray, J. E., & Maiden, M. C. J. (2018). Open-access bacterial population genomics: BIGSdb software, the Pubmlst.org website and their applications. Wellcome Open Research, 3, 124. https://doi.org/10.12688/wellcomeopenres.14826.1\n\n"
+        "'Bactopia'\n"
+        "Petit III RA, Read TD Bactopia - a flexible pipeline for complete analysis of bacterial genomes. mSystems 5 (2020) https://doi.org/10.1128/mSystems.00190-20\n\n"
+        "adios!"
+    )
+
+
 def setup_pubmlst(
     site: str, database: str, token_file: str, client_id: str, client_secret: str
 ):
@@ -115,7 +130,7 @@ def check_session(site: str, token_file: str, database: str = "yersinia") -> dic
     tokens = parse_json(token_file)
 
     if not tokens["session_token"] or not tokens["session_secret"]:
-        logging.warning("Session token not found, will request a new one")
+        logging.debug("Session token not found, will request a new one")
         needs_update = True
     else:
         # Try querying the API to see if the session token is still valid
@@ -548,13 +563,14 @@ def download_database(
     logging.info(f"'{database}' complete")
 
 
-def build_blast_db(database_dir: str):
+def build_blast_db(out_dir: str):
     """
     Build a BLAST database from the loci FASTA files that is compatible with 'tseemann/mlst'
 
     Args:
-        database_dir (str): The directory containing the loci FASTA files
+        out_dir (str): The directory containing the loci FASTA files
     """
+    database_dir = f"{out_dir}/mlstdb"
     blast_db_dir = f"{database_dir}/blast"
     if not Path(blast_db_dir).exists():
         Path(blast_db_dir).mkdir(parents=True, exist_ok=True)
@@ -579,7 +595,7 @@ def build_blast_db(database_dir: str):
 
     # Build the BLAST database
     execute(
-        f"makeblastdb -hash_index -in mlst.fa -dbtype nucl -title 'PubMLST' -parse_seqids",
+        "makeblastdb -hash_index -in mlst.fa -dbtype nucl -title 'PubMLST' -parse_seqids",
         directory=blast_db_dir,
     )
 
@@ -589,4 +605,5 @@ def build_blast_db(database_dir: str):
 
     # Tar the BLAST database
     logging.info(f"Save BLAST database to '{database_dir}/mlst.tar.gz'")
-    execute(f"tar -czvf mlst.tar.gz mlstdb/", directory=database_dir)
+    execute("tar -czvf mlst.tar.gz mlstdb/", directory=out_dir)
+    shutil.move(f"{out_dir}/mlst.tar.gz", f"{database_dir}/mlst.tar.gz")

environment.yml

@@ -2,7 +2,8 @@ channels:
   - conda-forge
   - bioconda
 dependencies:
+  - blast
   - pigz
-  - poetry =1.3
-  - python >=3.8
+  - poetry
+  - python >=3.8,<3.12
   - wget