Merge pull request #173 from rhysnewell/add-singlem-metapackage-downl…

…oading

Add SingleM metapackage downloading

README.md

@@ -141,6 +141,7 @@ These environment variables can also be configured manually, just set the follow
 ```
 export GTDBTK_DATA_PATH=/path/to/gtdb/gtdb_release207/db/ # https://gtdb.ecogenomic.org/downloads
 export EGGNOG_DATA_DIR=/path/to/eggnog-mapper/2.1.8/ # https://github.com/eggnogdb/eggnog-mapper/wiki/eggNOG-mapper-v2.1.5-to-v2.1.8#setup
+export SINGLEM_METAPACKAGE_PATH=/path/to/singlem_metapackage.smpkg/
 export CHECKM2DB=/path/to/checkm2db/
 export CONDA_ENV_PATH=/path/to/conda/envs/
 ```

aviary/aviary.py

@@ -266,11 +266,11 @@ def main():
 
     base_group.add_argument(
         '--download', '--download',
-        help='Downloads the required GTDB, EggNOG, & CheckM2 databases if required',
-        type=str2bool,
-        nargs='?',
-        const=True,
+        help='Downloads the requested GTDB, EggNOG, SingleM, & CheckM2 databases',
         dest='download',
+        default=[],
+        nargs="*",
+        choices=["gtdb", "eggnog", "singlem", "checkm2"]
     )
 
     base_group.add_argument(

aviary/modules/annotation/annotation.smk

@@ -40,9 +40,10 @@ if config['mag_extension'] == 'none':
 
 rule download_databases:
     input:
-        'logs/download_gtdb.log',
-        'logs/download_eggnog.log',
-        'logs/download_checkm2.log'
+        'logs/download_gtdb.log' if "gtdb" in config["download"] else [],
+        'logs/download_eggnog.log' if "eggnog" in config["download"] else [],
+        'logs/download_singlem.log' if "singlem" in config["download"] else [],
+        'logs/download_checkm2.log' if "checkm2" in config["download"] else [],
     threads: 1
     log:
         temp("logs/download.log")
@@ -111,6 +112,18 @@ rule download_gtdb:
         '  echo "[INFO] - Conda not found in PATH, please be sure to set the TARGET_DIR envrionment variable"; '
         'fi; '
 
+rule download_singlem_metapackage:
+    params:
+        metapackage_folder = os.path.expanduser(config['singlem_metapackage'])
+    conda:
+        "../../envs/singlem.yaml"
+    threads: 1
+    log:
+        'logs/download_singlem.log'
+    shell:
+        'singlem data --output-directory {params.metapackage_folder}_tmp 2> {log} && '
+        'mv {params.metapackage_folder}_tmp/*.smpkg.zb/payload_directory {params.metapackage_folder}'
+
 rule download_checkm2:
     params:
         checkm2_folder = os.path.expanduser(config['checkm2_db_folder'])

aviary/modules/processor.py

@@ -235,6 +235,8 @@ def __init__(self,
         except AttributeError:
             self.mag_directory = 'none'
 
+        self.download = args.download
+
         try:
             if args.gtdb_path is not None:
                 self.gtdbtk = args.gtdb_path
@@ -374,6 +376,7 @@ def make_config(self):
         conf["long_contig_size"] = self.long_contig_size
         conf["min_contig_size"] = int(self.min_contig_size)
         conf["min_bin_size"] = int(self.min_bin_size)
+        conf["download"] = self.download
         conf["gtdbtk_folder"] = self.gtdbtk
         conf["eggnog_folder"] = self.eggnog
         conf["singlem_metapackage"] = self.singlem

docs/examples.md

@@ -106,6 +106,9 @@ use the the `aviary configure` module to update the environment variables used b
 These environment variables can also be configured manually, just set the following variables in your `.bashrc` file:
 ```
 GTDBTK_DATA_PATH
+EGGNOG_DATA_DIR
+SINGLEM_METAPACKAGE_PATH
+CHECKM2DB
 CONDA_ENV_PATH
 ```
 

docs/faqs.md

@@ -66,6 +66,8 @@ the `activate.d/aviary.sh` or `.bashrc` files changing the specific paths:
 ```
 export GTDBTK_DATA_PATH=/path/to/gtdb/gtdb_release207/db/ # https://gtdb.ecogenomic.org/downloads
 export EGGNOG_DATA_DIR=/path/to/eggnog-mapper/2.1.7/ # https://github.com/eggnogdb/eggnog-mapper/wiki/eggNOG-mapper-v2.1.5-to-v2.1.7#setup
+export SINGLEM_METAPACKAGE_PATH=/path/to/singlem_metapackage.smpkg/
+export CHECKM2DB=/path/to/checkm2db/
 export CONDA_ENV_PATH=/path/to/conda/envs/
 ```
 

docs/installation.md

@@ -128,5 +128,7 @@ These environment variables can also be configured manually, just set the follow
 ```
 export GTDBTK_DATA_PATH=/path/to/gtdb/gtdb_release207/db/ # https://gtdb.ecogenomic.org/downloads
 export EGGNOG_DATA_DIR=/path/to/eggnog-mapper/2.1.7/ # https://github.com/eggnogdb/eggnog-mapper/wiki/eggNOG-mapper-v2.1.5-to-v2.1.7#setup
+export SINGLEM_METAPACKAGE_PATH=/path/to/singlem_metapackage.smpkg/
+export CHECKM2DB=/path/to/checkm2db/
 export CONDA_ENV_PATH=/path/to/conda/envs/
 ```